Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: Djole on November 07, 2017, 05:19:51 PM
-
Fast translate from: http://www.a1k.org/forum/showthread.php?t=63012
Hello,
Fast 68040 or 68060 CPUs are not to be found in quantities.
Therefore, it is difficult to build new 68060 cards.
The APOLLO team offers a number of very fast 68K maps -
but we think shared joy is a double joy.
We would therefore like to discuss the following suggestion:
To support hardware developers, we would like to make the following offer:
We offer hardware developers a
_FREE_ 68080 CORE that is
compatible with the 68040/68060
and is set to about 68040/68060 speed.
The core is FREE for hardware developers.
The core also contains the AMMX2 instruction set.
This means that software compatibility with vampire software such as new JPEG decoder, video player and some game ports remains.
The Core boots with over 100 Sysinfo MIPS and after some time down to 68040 @ 40/68060 @ 50 level.
The supported hardware cards are subject to some reasonable conditions. That means the FPGA has to be big enough to offer a complete, fully compatible core space. The cards must be intended for the general public. The offer is not valid for single pieces and Pershingraketen.
We are looking forward to your feedback.
Greetings
Gunnar
-
this started a quite interesting talk on A1k.
and.. there was HW Desogners more or less telling that "nope not interested"
Thing is.. that there are many like me that are purists. and 080 is not our cup of tea.
I got asked if I wanted to add this to the A3660. answer is simply: NO
2 reasons:
1. there would be just stupid to add 080 to a 040/60 board with all busmastering etc. better to do it from scratch
2. 080 is.. well nothing I want to support.. so not from me. :)
for those who like the 080.. it might be fun. yes. but the support from hw designers seems so far very low.
(yah! I might be wrong.. but I do not think so)
-
but people have issues with missing MMU and the (so far) lacking of FPU
-
It's just a sign of desperation after no-one approached them... him... about licensing the core.
-
We offer hardware developers a
_FREE_ 68080 CORE that is
compatible with the 68040/68060
That is not how the word "compatible" works - unless you plan to build a board with both 68080 CORE and 68040/68060. :hammer:
-
but people have issues with missing MMU and the (so far) lacking of FPU
I dont see the point of your first post and the need to express your lack of interest, you are a HW designer ? FPU is not lacking its in the works and shown working, faster than any moto FPU. MMU is also there it has been stated 100 times, its just not compatible with old moto MMU....
-
For me compatible means it acts like that cpu. and 080 is not compatible as it doesn't act like 040 OR the 060.. it acts like a mashup of all 68k cpus. but doesn't behave like any of them. And this is my biggest issue.
(and the fact it lacks both MMU and FPU so far)
-
HW Designer-lite maybe as I did "design" the A3660 PCB :)
-
and noncompatible MMU while it states it IS compatible with the 040/060?
Strange.. is it compoatible or not?
Anyway.. no MMU stuff means no debuggingtools meaning programmers avoid it meaning no software maning: no use of the bells and whistles..
FPGA is nice..and I would love it.. IF 060..
-
Are you trying to boost your post count or something ? 3 posts for one answer ? BTW new debugging tools are also in the works.... strange you dont mention FPU anymore ?
-
Well so far FPU is emulated via software.. and I find that VERY ironic (as vampirepeople say it is more compatible not needing the 040/060 lib)
and no.. it is not about "boosting postcount" or so.. not at all actually..
And well I posted on the A1k thread about my opinion.. and I am not alone of retro-computing.
-
Well so far FPU is emulated via software.. and I find that VERY ironic (as vampirepeople say it is more compatible not needing the 040/060 lib)
and no.. it is not about "boosting postcount" or so.. not at all actually..
And well I posted on the A1k thread about my opinion.. and I am not alone of retro-computing.
Check your facts again, Apollo core FPU in v2 Vampire is a mix of hw and sw, in v4 it should be full in hw an average user shouldnt worry about it anyway.... it should just work.
But anyway, good work on derailing another Vampire post.
-
so far the FPU that is "hwbased" is only for vampireteam.. not for the public.
and it still relies on software.. meaning it is EXACTLTY as the 040/060 library.
so if the 040/060 is nonocompatible due to this. vampire IS with that exact same logic noncompatible.
And well. as I told in the A1k thread: for me as a demoscener. the 080 is as interesting as yesterdays newspaper.. as I have no interest whatsoever of AMMX and stuff.. I want 060..
so I would welcome a 060 FPGA core any day.. we need it as 060s are gettong harder to get (saying that as I got my next batch of 21 rev6 060s yesterday :))
-
Trigger for me is:
"_FREE_ 68080 CORE that is
compatible with the 68040/68060 "
when it ISN'T!
It is maybe 68k-series compatible.. but it for sure is NOT 040/060 compatible.
as then it would have the FPU and MMU!.. it doesn't
it might be 68EC040 and 68EC060 compatible. but when I read 68040/060 I read the full deal and this is important.
-
Trigger for me is:
"_FREE_ 68080 CORE that is
compatible with the 68040/68060 "
when it ISN'T!
It is maybe 68k-series compatible.. but it for sure is NOT 040/060 compatible.
as then it would have the FPU and MMU!.. it doesn't
it might be 68EC040 and 68EC060 compatible. but when I read 68040/060 I read the full deal and this is important.
Its every time the same song.....
-
and I am not alone.. especially not in the demoscene..
if it was 060.. I would buy 10 vampires today.. and I would love to be involved in any hw project using that FPGA core.
080.. no. not at all. as it just simply doesn't do it for me.
And reading at the A1k thread etc I know I am far from alone. and this is something that the vampire/apolloteam must understand. we want our beta-max VCRs.. we do not want Bluray.. we use retromachines.
And what I understand.. Jens told in the A1k thread: not interested.
just of this reason.
it however was a nice reach-out from Gunnar. absolutley.. but he and the vampire team takes critics waaay to personal.
-
Chucky... dont you feel like you are repeating yourself ?? You said it all in the first post, no need to repeat 10 times you are not interested. You are not forced to use or buy it, you keep your betamax. You can buy your 020 and 030 cards from Jens, enjoy them. Should Vampire users invade any post about his HW and repeat the same boring story about missing features and how they are not interested ??
-
and it is reactions like this: taking it personal..
I just want the lie about being 68040/68060 compatible, it can fool people who thinks that they will get that. and when they try to run enforcer/mungwall etc.. it will fail..
It is called false marketing..
And belive me. there are always some vampireperson coming to discussionthreads about 040/060 etc that "no. the future is vampire" while sorry to say:it isn't..
the vampire is a neat way to boost your A600/500 and I recomend it to anyone wanting to use their old machines as a whdload machine.. it will be perfect for it..
for me as a demoscener: no..
for a replacement of 060: in most cases.. no as if you want to go highend. why choose a sorry to say: crippled cpu?
I read at the a1k forum and it is interesting to see how Gunnar just misses the whole point when people ask about MMU etc. telling that his MMU is better. and it might be. but with no software, noone cares. but programmers are used to have their debuggingtools why suddenly be without them? and wit no programmers.. who would use the new nifty functions?
and repeating myself? well yes. as there is always some vampireperson repeating that "vampire IS the thing everyone wants" but still completly ignores that there are many of us thinking the opposite.
I have told many times in different forums etc: if vampire was actually compatible (meaning ACTING like the real deal) I would buy 10.. (instead of that one I have. somewhere)
-
I have told many times in different forums etc: if vampire was actually compatible (meaning ACTING like the real deal) I would buy 10.. (instead of that one I have. somewhere)
Yeah, I don't have a problem with gunnar getting his wet dreams over designing 68080. If he also had a mode where it would run as a 68040 or 68060 with fpu and mmu.
If he can't do it, then just release the 68080 source and let someone else do it. As he rightly points out "but we think shared joy is a double joy."
Until that happens, it's kinda boring. Drop in replacement cpus for next cube or 68k macs ftw.
-
This thread is everything wrong with the Amiga scene.
Someone who hates something that somebody else has done that does't appeal to their purist ideals (despite the fact that the Moto 68k cores are very different, all have different FPU capabilities, and different MMU implementations), and they just can't stop pointing it out.
-
Andbuy 020/030 cards from jens? Nahh I did the reverseengineering of the 3640 and made the 060 for it as I hope the open community will do its magic..
and stuff IS happening. so I am pretty sure we soon will have open 040/060 solutions.. but YES getting the cpus might be an issue.
-
I don't get what the big hoopla about this news is, nobody will license it anyways, waaay too risky.
-
I see two major issues with this.
1 - Gunnar is perhaps concerned about other projects/open source CPUs, and by offering the core to other vendors he hopes to gain market share, however isn't he hurting the guys who make the vampire hardware/stand alone? Surely you can't have it both ways.
2 - Throttled back to something slower than they can do - so bit of a second class citizen ? no thanks.
-
and it is reactions like this: taking it personal..
I just want the lie about being 68040/68060 compatible, it can fool people who thinks that they will get that. and when they try to run enforcer/mungwall etc.. it will fail..
The people who are going to need/use Enforcer/Mungwall will know what features Apollo offers, belive me. 99% of the users dont even know know what it is, let alone use it.... And to be in your style I will repeat myself too, Apollo debugging tools are in the making.
It is called false marketing..
And belive me. there are always some vampireperson coming to discussionthreads about 040/060 etc that "no. the future is vampire" while sorry to say:it isn't..
Yeah those people should be banned like the people intentionally derailing every Vampire thread.... This thread isnt about what you like or dont like, its about the offer Gunnar has made for the HW designers. I dont see the reason you are posting here at all. Why should any1 care what you are interested in or not ? If you said, "I am making a new Amiga Turbocard but i dont want to use Apollo core", that I would understand.
the vampire is a neat way to boost your A600/500 and I recomend it to anyone wanting to use their old machines as a whdload machine.. it will be perfect for it..
Wrong, anyone with any knowledge would recommend a 020 or 030 card with some ram for a whdload machine....
for me as a demoscener: no..
for a replacement of 060: in most cases.. no as if you want to go highend. why choose a sorry to say: crippled cpu?
Performace wise, 080 is faster than any 060, integer AND floating point.... So obviously you are spreading misinformation here....
I read at the a1k forum and it is interesting to see how Gunnar just misses the whole point when people ask about MMU etc. telling that his MMU is better. and it might be. but with no software, noone cares. but programmers are used to have their debuggingtools why suddenly be without them? and wit no programmers.. who would use the new nifty functions?
Vampire has kickstarted many Amiga sw projects, not only usable on Apollo core and many more are in the pipeline....
and repeating myself? well yes. as there is always some vampireperson repeating that "vampire IS the thing everyone wants" but still completly ignores that there are many of us thinking the opposite.
Maybe you should set the example and stop. WE all got the point, you dont want a Vampire.
-
In time it would be lovely to able to replace the core on V2 cards with something more tinker friendly :)
-
and noncompatible MMU while it states it IS compatible with the 040/060?
Strange.. is it compoatible or not?
Anyway.. no MMU stuff means no debuggingtools meaning programmers avoid it meaning no software maning: no use of the bells and whistles..
FPGA is nice..and I would love it.. IF 060..
Most people understand 'compatibility' in a CPU to mean 'runs the same software'. The Apollo Core in Vampire appears to do that in the main. There is FPU support (however it is achieved), which is more than a lot of old Amigas ever had.
So it's the MMU, and one bit of software, Enforcer, that drives all the arguments. And if a developer uses that, then that is a problem, but plenty of developers do not use it. It's very niche software that most users won't need. But something to replace it would be nice.
Given the '060 doesn't behave like the '040, which doesn't behave like the '030, which doesn't behave like the 68000, we should stick with the latter?
You aren't going to get an FPGA '060 either, that would require a lot of reverse engineering, just to redo what has already been done.
I note that the vast majority of your posts here are anti-Vampire.
-
I see two major issues with this.
1 - Gunnar is perhaps concerned about other projects/open source CPUs, and by offering the core to other vendors he hopes to gain market share, however isn't he hurting the guys who make the vampire hardware/stand alone? Surely you can't have it both ways.
2 - Throttled back to something slower than they can do - so bit of a second class citizen ? no thanks.
Well its clear he wants to gain market share. Maybe the current Vampire hardware makers are getting a better deal ?
Other (open) projects we still need to see beat Apollo performance and compatibility....
You can pay for first class, like in a plane :) So But a free core in the 060 speed class is not bad, what other core can match that ??
-
The people who are going to need/use Enforcer/Mungwall will know what features Apollo offers, belive me. 99% of the users dont even know know what it is, let alone use it.... And to be in your style I will repeat myself too, Apollo debugging tools are in the making.
issue is..people do not want to use different tools for Apollo and different for others.. they want to use the same tool.
Yeah those people should be banned like the people intentionally derailing every Vampire thread.... This thread isnt about what you like or dont like, its about the offer Gunnar has made for the HW designers. I dont see the reason you are posting here at all. Why should any1 care what you are interested in or not ? If you said, "I am making a new Amiga Turbocard but i dont want to use Apollo core", that I would understand.
Banning is the wrong way. throwing out people is "managment by fear" and is totally worthless. it is the worst way of dealing with people with a different opinion.
Wrong, anyone with any knowledge would recommend a 020 or 030 card with some ram for a whdload machine....
well yes. but you get a nice package with the vampire.. you get memory, you get RTG (and supposly AGA later) instead of trying to get that separate for those old machines. that's why. and is also the biggest reason why most get vampire.. you smack in one card and get it all.. (even if many never even use the RTG!
Performace wise, 080 is faster than any 060, integer AND floating point.... So obviously you are spreading misinformation here....
no debuggingtools. and a halfworking FPU. == crippled for me as a demoscener.
Maybe you should set the example and stop. WE all got the point, you dont want a Vampire.
And I want the vampirepeople using the word "Compatible" more carefully..
AND listen to people with other opinions. and NOT block people.. it is the wrong way.. as people say: I want MMU that works as it is on the old machines. and there are a lot of them: TAKE THE HINT! we do not care if that solution is 100000 times better..
same with the FPU! "not needed" (or even "not wanted") until someone makes a fpu-emulator.. then suddenly it is needed..
listen to critics: take input.. and do not take it personal. and then suddenly they will get a awesome products that have stuff for everyone.
-
Wrong, anyone with any knowledge would recommend a 020 or 030 card with some ram for a whdload machine...
Noted, let's remember this one :)
Performace wise, 080 is faster than any 060, integer AND floating point.... So obviously you are spreading misinformation here....
Not the core that is available, and we have been told over and over again that it's what's available *NOW* that matters.
Vampire has kickstarted many Amiga sw projects, not only usable on Apollo core
Such as?
(Please answer FEMU!)
Maybe you should set the example and stop. WE all got the point, you dont want a Vampire.
He has one, I have two, we are your "users who don't give a rat's ass about WHDLoad (or Quake for that matter)" :)
-
You can pay for first class, like in a plane :) So But a free core in the 060 speed class is not bad, what other core can match that ??
You'll see..
-
@Chucky
If they listened to every chap out here they would never got this far. They have their plan and they work on it, accept it or dont. You keep on babbling about what you want or dont want, their success has shown what the most people want. You already have what you want, your old 060.... And you are constantly derailing this thread, none of your replies are on topic, that should be sanctioned, if you feel the need to force your opinion on Apollo core, there are other threads out there you could find that are more related.
-
Clarification question... the developer/base card core is "free", and at what speed? 040/50 like?
Can the purchaser of the card then buy upgraded cores for features/speed later?
This might let a card manufacture a low end and higher end option.
Locked card at one price, unlocked card at another.
And BTW, what is the moderation like around A.org these days? Seems the few topics I drop in on are often already on their way to spinning off-topic. But my visits are few.
Plaz
-
Reading the translated pages now. Seems the answer to my questions are both "yes"
Plaz
-
I guess Gunnar feels "threatened" by all the cheap, open and available lowend cards that Terrible Fire and others have been making lately, they are "stealing his market share" so to speak. So now he wants in on that too. I recall him ranting along the lines of "only idiots make 030 designs" and using less flattering words to describe these efforts and the people involved. Well, but why go through the trouble of Apollo Core with licensing and contracts and unclear customer support, for users who 99% (his numbers) of the time just want to load old games with WHDLoad? :)
-
You'll see..
Mike, with all the respect, i think you said that before to me.... That was a while ago. A lot of people will be very impressed if you show us a better softcore then Apollo.
-
If they listened to every chap out here they would never got this far. They have their plan and they work on it, accept it or dont. You keep on babbling about what you want or dont want, their success has shown what the most people want. You already have what you want, your old 060.... And you are constantly derailing this thread, none of your replies are on topic, that should be sanctioned, if you feel the need to force your opinion on Apollo core, there are other threads out there you could find that are more related.
NONE of topic? this is very much ON topic. and YES me and MANY OTHERS want our old 060s.. if I want a new machine. I take a PC.. I can get a laptop for the same price as a vampire that runs browsers 1000 times better, runs videos etc. and actually runs UAE pretty damn decent.
this topic is to use a 080 core as a 060 CPU-replacement due to the shortage. When I try to say that sorry bit it isn't many of us does not think it is good enough. (even if 080 is pretty awesome.. but hey again new? that laptop?)
IF it was 060 compatible then it would be a AWESOME idea that I would love to hop on etc.
I want to preserve my nostalgic machine. and so is the most who wants the 060 and highend-machines. (not all! no!)
and this IS a reachout of highend users, then input from highendusers should be taken.
but as always. when saying "nah. this would be awesome" it is always taken way too damn personal..
at Amiga32 I told parts of the team: make it 060 and AGA 100% and the rest as RTG/AHI I would LOVE it! it would be the perfect solution for me and my Amiga-needs. (and many with me)
videos, browising .. well that I use my PC, phone etc. Quake and doom is not interesting for me.
But vampire-people takes things way way WAY too personal..
-
@Kolla
You where in the frontline of FPU fighters, mentioning it in every needed and unneeded topic, ranting about it, demanding....now you have some other life goal ?
-
Mike, with all the respect, i think you said that before to me.... That was a while ago. A lot of people will be very impressed if you show us a better softcore then Apollo.
I've been busy making new hardware as well, it's coming.
-
This will be good received in the atari side of the force. I hear some people that where looking to license the core.
Owning a vampire i don't see al that compatibility issues that where mentioned on the thread.
Games from whdload run fine and most of the amiga sofware i have installed works, nothing different from my old (sold a long time ago) amiga 4000/060 and quite faster.
Yes the fpu emulation for now (running gold 2.5) is quite slow quake crawls for example.
But there no so much stuff that need an fpu, so theres no hurry for me to see this fixed.
-
For me. incompabilities started from the beginning.. I have a 3.1 and 3.9 template with a lot of software installed.. working on 68000-68060 and ALL amigas.. but on Vampire.. it all failed. had to do a some tweaks to get it boot and be stable. And I brought it to the datastorm demoparty.. we checked it out. and "ok this is for checking sysinfo and doom" meehh lets have some beers instead..
and there it is. since that I haven't tried my Vampire.. oh yes I did try it on my Amy.. did not boot. but this is most likly not the vampires fault but some soldererror on the amy.
However.. it was not "compatible" for me. it did not behave like any amiga... so that's why I think it is not compatible.
and for me as demoscener: fpu is important for highend-demos.
-
@Chucky
It sounds like you dont know how to set it up and/or use it.... But this is not a Vampire help thread.
@topic
We will see if any other retro scene will take this offer and produce something useful.
-
And there are ONE more issue here and it is an important one:
You rely on ONE person.. as trhe core is closed..
if a bus runs over gunnar or so. you are in a situation where you are screwed...
and at work being a consultant: NEVER EVER put your stuff into something being dependent of one person.
-
@Chucky
It sounds like you dont know how to set it up and/or use it.... But this is not a Vampire help thread.
if it was COMPATIBLE.. it should run!.. I solved it. but it did not run directlty! I needed to do changes.. so NOT COMPATIBLE! this system works on any machine with ANY motorola cpu.
(ok 3.9 boots but will have issues on 68000 and 68010 due to some libs are 020+)
-
@Chucky
I will start reporting all your off topic posts in this thread, so you know. You constantly try to derail the thread for some reason.
Your other 060 cards are open source ? Or even available ??? Or ever been available for a600 and a500 ? You are making yourself ridiculous...
-
HOW can you get this as OFF TOPIC!???
ok I have disagreements with the 080.. but OFF TOPIC?? HOW is it off topic? please explain?
and yes.. my A3660 project is fully open. yes.. "%&$#?@!%&$#?@!%&$#?@!%&$#?@!ty" card.. absolutley.. but still totaslly open.. just download the files. do the pcb.. do whatever you like with it
-
So I can fit your card in a500 or a600 ? You are comparing apples and oranges.... Or you are just a bit slow ?
-
hehe did not like the Sh.. word.. :)
but I get it. typical Vampire thing: if someone say somethingh that you doesn't approve with: MAKE IT SILENT!
actually that scares me.. that is a horrible way of taking input.. that looks like dicatorship.. scary.. REALLY scary..
how to even rely on that? put your design on that? I would not dare.. sorry..that is absolutley not how to do it.. you will not get a good product if yo only have "yes-sayers"
-
I know my A3660 project is a bad card. really bad.. but it is my part of giving to the community.. it sure needs improvements for sure.
but I WANT comments.. I know what is bad and what is good. thing is.. if you want improvements and can do it youself: feel free (but do NOT close!)
and if you want people to use your techs: make no-sayers silent but shuttingh them down is scary.. it is wrong.. this means: if I want to use your core. but find an issue.. what support will I get? I might get shutdown? maybe after putting long time of effort.. maybe even lots of money doing prototypes etc?
dangerous.. please show that this is wrong. take the input. do not shut it down..
-
What input do you have on 060 CPUs ? People get spoiled very fast. They think they should get custom made core and cards just by asking and repeating it on forums. Anyway I am done here, this is starting to get very funny.
-
if a bus runs over gunnar or so. you are in a situation where you are screwed...
One of the companies I work for had their entire online advertising campaign managed by one person. A $20 mil company and only one person had all the passwords. Well, imagine how much fun it is to try to recover the passwords for all of your AdWords accounts when the only person who knew them has died? Hint, it isn't. :mad:
So yeah, this kind of thing happens. :(
-
the input. well it is a STATIC product. we all know how i works.. it have worked for 20+ years for us.. we have software that runs on it.. and it is the last of the CPUs on the Amiga. the cpu where MY heart stands at: the demoscene (and well the 68000 of even more retro)..
and for me it is the goal.. it is what I want and need. and what the most of my fellow amiga-demosceners want.. and EXACT that product..
no ammx, no 64 bit, no hyperthreading... but yes it IS right sourcing the 060s is getting harder. we NEED a fpga replacement. but sorry. many does not need quake and doom or videos. we need a fpga replacement that still runs like our beloved 060.
and this is what I try to tell you: I am not alone with this..
the 080 is a core under development.. currentlty for us without any fpu and without any usable MMU.. so if you silent no-sayers that want stuff.. HOW would I trust the development? WHY put in money in developing it if I might get silent what I have an opinion? I might end in an unfinished product as I got shutdown maybe? maybe spending lots of money in prototyping?
this is a big tip for you: if you seek partners (as thsi thread actually IS about) do not shut down people with opinions or ideas.. that is bad marketing.
-
Many companies I work with have put their production into systems where only one or 2 persons know about the system: cancer. and BAM! no support.. no nothing.. just a dead product..
thats the hard fact of life.. noone wants it.. but it can happen. and it have impact.
-
They have their plan and they work on it, accept it or dont.
If they were honest about it then this thread would be dead.
-
This thread is dead. The Vampire is still a work in progress, is it not?
-
It amazes me that people can devote so much time and energy to be negative about a product that claim they have no interest in ("yesterdays newspaper"). If it is not for you, fine state that if you feel the need and move on, don't hang around posting continuous negativity like a bad smell. I'm sure you have better things to be getting on with.
For the record I'm not in favour of the proposal as I see it as a backwards step. If hardware devs want to licence the full Apollo core and provide a full SAGA capable expansion then great. I don't like the idea of cut down and throttled Apollo 68080 boards (at least not for the Amiga).
-
It amazes me that people can devote so much time and energy to be negative about a product that claim they have no interest in ("yesterdays newspaper"). If it is not for you, fine state that if you feel the need and move on, don't hang around posting continuous negativity like a bad smell. I'm sure you have better things to be getting on with.
For the record I'm not in favour of the proposal as I see it as a backwards step. If hardware devs want to licence the full Apollo core and provide a full SAGA capable expansion then great. I don't like the idea of cut down and throttled Apollo 68080 boards (at least not for the Amiga).
Your post is a bit of a mixed message, I hope the apollo gestapo are kinder to you.
-
Hey Chucky -
Having fun in this thread? ;)
I think your A3660 board is a great idea for upgrading A3640s. I may look into that some time next year (I have an 060 and an A3640 kicking around).
As for the Vampire/Apollo core... This isn't your thing now. Sure, I get that. Let's see how the FPU issues resolve in V2.7 and also the debugging tools that will take advantage of the internal MPU.
I think that in the next few months there should be better answers to the issues you are highlighting. In addition there will be more end-user experience out there.
Cheers!
-
For me. incompabilities started from the beginning.. ... And I brought it to the datastorm demoparty.. we checked it out. and "ok this is for checking sysinfo and doom" meehh lets have some beers instead..
and there it is. since that I haven't tried my Vampire..
So you are basing your opinion on old information. Did it ever occur to you that the core is a work-in-progress and that many of your compatibility problems may have been already solved? I do understand that a work-in-progress may not be an attractive alternative for many NOW (especially for people who'd invest time and money into developing some new hardware based on this work-in-progress). Yes, so just wait some more but don't dismiss this prematurely and based on past experience. Let me add another point: it is fine that you and your demoscener pals all have 060 accelerators. Congratulations on forming part of the Amiga elite. However, many people don't have one and there are not enough 060s left to satisfy all the market demand for fast Amiga accelerators. The only option for them is to buy a new 020 or 030 based accelerator (usually without an FPU and MMU), spend a ridiculous amount of money for an aging accelerator card or buy an 080-based accelerator. Anyone can make their choice for themselves. Any new option should be welcomed and even if it is only some ARM core dedicated to running UAE-jit 68k emulation.
-
well I still see people having that same issues that I had. so those incompabilities is not solved yet. and I constantly hear that it is "MORE COMPATIBLE" when it doesn't behave like any 68k. so how can it be more compatible?
for me Compatible = Behaves like THAT product..
Anyway. I will someday finish my blog-post of this..
Amnyway. you are right. 060 cards is getting insanly expensive.. and I am doing whatever I can to solve that. but the thing is. the vampirecards are soon approaching that very same pricetag... yes professional assembly costs way more than doing it by hand (damn I am impressed how Majsta and kipper had the energy doing it!) but I got the feeling they soon get to that EXACT pricetag. except that the difference is: vampire and 080 does not depend on old hard-to-get (read: expensive) technlogy..
Anyway my point is still: IF they want people to use their core. DO listen to more people. I see often in different forums (here in Sweden and IRC etc etc) that people want FPU and MMU or they are not interested.
but when people say that. they take it personal like if you call them stupid.. WE DO NOT! damn the Vampire and 080 is actually damn awesome.. too sad that there are many wanting "just" the old toaster, but we want it preserved as tech fail, cpus are getting harder to get etc. no MMU as a good example is giving coders the finger. and all new stuff they want to impleent means: new software needed. do NOT give the coders the finger. you need them. listen to input. do not block the people with requests... I am very often to demoparties etc. I talk to coders etc and I still have to fine ONE democoder remotly interested in doing something. as it breaks the bonderys of why they even care about the amiga. they just stopped even caring about comenting it as if you do. things like this thread happens.
the GESTURE of the core "for free" is nice. very nice. but in this thread (and others I have read about this subject) you get an impression that this is more or less one-man-mission of how to do it and everyone else is wrong nd they just want to silence anyone with an opinion.
At a1k I notice comments about FPU and MMU. there IS demand. and so far we users.. do NOT have FPU and Gunnar really showed that he is NOT interested in implementing a MMU (or.-. MPU!) compatible to older software as his is SOO much better. and it might be right.. still people asks for it. WANTING to do new software. still "no. my stuff is much better"
-
ANYWAY! I will stop here..
I will do my part. I got mail etc asking if I am interested of this deal for my 3660 project and. well (tipping me of it.. ) and. NO.
if you can solder: you can do a 060 for your bigbox Amiga for aprox 190eur. not the fastest. actuallyt the slowest.. :) but it works- and stuff is happening.
I know that I need to add memory (and a ROM disabling that damn FPU at boot), remove old components and replace with CLPD etc. and I am NOT negative as ths "competes" with the 3660. actually I would glad if I just could sit down and relax if someone else did it, but thinking like that: nothing happens.
This thread just proved what I suspected. but I do wish them good luck.
-
At a1k I notice comments about FPU and MMU. there IS demand. and so far we users.. do NOT have FPU and Gunnar really showed that he is NOT interested in implementing a MMU (or.-. MPU!) compatible to older software as his is SOO much better. and it might be right.. still people asks for it. WANTING to do new software. still "no. my stuff is much better"
With regards to FPU; Upgrade your core to 2.7 WHEN it is released, and then pass judgement on the FPU implimentation. If you as a coder (or organizer) have spesific compiler/development issues with regards to the FPU internals, then listing spesific issues would be great.
With regards to MMU; the roadmap forward for MMU implimentation isnt clear to be honest. I remember when there where a big discussion regarding lack of FPU several months back. Gunnar eventually got tired of all the bickering from us (myself included), and just said "No FPU for you!" etc. Obviously he just kicked sands in our eyes cause we had the patience of children, not really accepting that WORK IN PROGRESS takes time. I see on ADA forum that the democoders say "Ive been working on this on and off for x months" etc. Designing a core probably takes time too, just saying.
So just cause MMU as Enforcer requires it is available or offered in Betatesting demostration, doesnt mean it wont surface eventually, just like the FPU did.
So I defintly would like to hear your detailed mapping of whats needed to be tweaked for the demoscene with regards to the FPU, ONCE Gold 2.7 with SoftFPU (with currently unknown level of hardFPU for V2).
-
Speaking as a demosceener:
WHY do we do demos on our amigas? well we love to show off what we can do with the limits. you know. poor chipmembus, bitmapmode etc etc.
we love to show off what we can do theree. if suddnely the chipmem is a 100+MHz bus etc. nah. then I can use a PC and do shadercoding. there is no more "fun" into that.
same with the CPU.. a demoplatform is 2 different machines todaty mainly:
1. OCS machine, meaning 7MHz 68000 with 512k chip and 512k fakefast (not even fastmem)
2. AGA machine (14MHz chipmem, 32bit) 2MB Chip, and 50MHz 68060 with FPU. 50MHz and aprox 64MB of RAM.
so it must deliver a exact version of ANY of this to be accepted. more speed can be fun. but the FPU must perform exactly MMU is more used to figure out bugs when developing utils.. not so much for demos.
but. vampire and 080 needs utils to be used. so that's why the MMU is important for the very few doing the actual coding.
so in short: demoscerns WANT their restrictions, we do not care about AMMX and stuff..
for utils using all the new stuff to get new software (like browswers) programmers needs debuggingtools they are used to have.
-
well I still see people having that same issues that I had. so those incompabilities is not solved yet.
If they really appear on the newest core release, then you should report those issues. And, if you want to make a good job while you are at it, try whether the same problem appears if you run the suspect code on WinUAE with 040-jit setting. This is because most reported incompatibility are just coding bugs where the software cannot deal with a CPU that is way faster than the programmer expected. The highest-ranking probability of such coding flaws is where coders start a blitjob without checking whether the preceding one has been finished already because they assume that whatever the CPU is doing in parallel will keep the CPU busy long enough for the preceding blitjob do end in time. and I constantly hear that it is "MORE COMPATIBLE" when it doesn't behave like any 68k. so how can it be more compatible?
for me Compatible = Behaves like THAT product..
So either the Apollo-Team is stupid or lying. Hey, wait, there is a third option: you are applying a definition of the word "compatible" that obviously cannot be the same one as used by the Apollo-Team. Obviously it would be impossible to behave more like something than the very thing itself, right? So if you keep pondering this for a second it should become apparent that it means that it does not behave like an 060 because it does not bork when it sees e.g. a 64bit multiplication. It does better, it just executes the instruction. Amnyway. you are right. 060 cards is getting insanly expensive.. and I am doing whatever I can to solve that. but the thing is. the vampirecards are soon approaching that very same pricetag...
How is 300EUR approaching anything? Yes, other versions with more capable hardware might cost more. Do you think this "trend" deducted from two distinct points will continue if alternative 080-based accelerators will hit the market? Anyway my point is still: IF they want people to use their core. DO listen to more people.
There is no Vampire that has not been sold. And, frankly put, the input to be expected is none that we couldn't come up with ourselves: make it compatible. Sure, but this takes work and work takes time. It is as simple as that. In the a1k thread Gunnar pointed out that he can do the compatible MMU now but this would mean changing priorities. If somebody believes that more money can be made with an accelerator that has a compatible MMU, this somebody would have to fund the development. Right now the AGA reimplementation seems to be much more important because there are many times as many people wanting a stand-alone Amiga or an AGA upgrade for their OCS/ECS hardware than people that cannot do without the legacy-compatible MMU. I am very often to demoparties etc. I talk to coders etc and I still have to fine ONE democoder remotly interested in doing something. as it breaks the bonderys of why they even care about the amiga.
Why would anyone want to cater a market with a new CPU that doesn't want to have any restrictions removed? I understand that some people find it interesting to code for an unexpanded A500. At the upper end of this pretty much the same people find it interesting to code for an A1200 with an 060/50. Well, so be it. Even availability of 75 MHz 060 didn't change the latter. This clearly isn't the market for a new CPU.
-
Hey Chucky -
Well in fact the demo platforms have 3 configurations:
1. OCS machine (7MHz), etc...
2. AGA machine (14MHz), etc...
3. AGA 060 machine (50MHz), etc...
And now the Vampire opens up the possibility of a fourth level:
4. SAGA 080 machine (100MHz) etc...
I would like to think that some (definitely not all) demo coders would be interested in taking their demos to the next level. Compared to multi-GHz PCs this is still a very restricted platform and does pose its own challenges and also opens up new possibilities.
But let's just see how things evolve over the next few months after V2.7 is out.
Cheers!
-
Oh I get what you are saying from that point of view :)
Im a big fan of C64 demos, and its amazing what they can do with that limited hardware.
Or more recently Altair - Zener Drive https://www.youtube.com/watch?v=Ei3IqD_zSkk
But take Kioea by Mad Wizards. A while back it was uploaded by ShK using the WIP core in July. It worked, but was laggy/buggy, and could make you cry watching it butcher that amazing demo; https://www.youtube.com/watch?v=q7Ye2c5K8GI
Now fast forward to October/November and it has improved massivly;
https://www.youtube.com/watch?v=dtCWfmnYOsk
Its really nice and smooth!
Then take the latest masterpiece from Haujobb and Ghostown - Beam Riders; https://www.youtube.com/watch?v=uJsZXgSaELE
It requires 1260, and it will just flip its middlefinger to my Blizzard 1230 III.
I could run in WinUAE or buy 1260 (or can I without selling my liver?).
Enter Vampire.
As for the future development for the demoscene vs Vampire; You could still just develop within the constraints of 1260 code, not utilizing the extra features of the 68080, but get a bigger audience that aint watching 99% of the demos in Youtube (Of which Im one sadly).
So there doesnt need to be a "either or" situation. You can still limit yourself.
But again, I do understand the charm of platform constraints and challenging yourself.
-
well yes. mainly it is ocs or expanded A1200. long time since I saw anything "stock A1200" (too sad.. I like that platform)
but no. Vampire 080 will not be a demosceneplatform when I talk to democoders: they are simply not interested.. as then they could do shaders on a PC instead.. there is no sport in it. and it is what it is all about: it must be HARD to do.
there might be SOME 080 demo. as there was some PPC ones. but all will be as wild-entrys I guess.
actually more and more are dumping the 060 awell. it seems that OCS is the platform they more aim for. they want TIGHTER limits.
-
well yes. mainly it is ocs or expanded A1200. long time since I saw anything "stock A1200" (too sad.. I like that platform)
but no. Vampire 080 will not be a demosceneplatform when I talk to democoders: they are simply not interested.. as then they could do shaders on a PC instead.. there is no sport in it. and it is what it is all about: it must be HARD to do.
there might be SOME 080 demo. as there was some PPC ones. but all will be as wild-entrys I guess.
actually more and more are dumping the 060 awell. it seems that OCS is the platform they more aim for. they want TIGHTER limits.
Wanting to limit themselves, and see how far they can push any given hardware is a valid argument for sure.
Like some that do Textmode, DOS demos with software rendering etc on PC.
-
As for the future development for the demoscene vs Vampire; You could still just develop within the constraints of 1260 code, not utilizing the extra features of the 68080, but get a bigger audience that aint watching 99% of the demos in Youtube (Of which Im one sadly).
The problem being that you would still need a real 060-card to make sure the demo actually works correctly on the target platform. This is pretty much why I dislike the added crap, as I'd rather see something replicating the 060 as close as possible.
-
Im sure you are not intrested to spend too much time trying to educate a code illiterate guy like me on this topic, but;
Does/might the added crap throw off timings and functionalities of current 680x0, or is it cause its an unknown variable as to how a routine will actually run?
-
Hi Guys... regarding OS 3.9 not working with Vampire... or not being compatible... Installing OS3.9 was the first thing I installed on Vampire and I got it working, and I am by no means an expert.
-
Does/might the added crap throw off timings and functionalities of current 680x0, or is it cause its an unknown variable as to how a routine will actually run?
There's always the risk of something not working as it should. I mean, even if I do most of my development using WinUAE, and pretty much know it emulates everything pretty good, I still do a lot of checking on actual hardware just to make sure.
But my main gripe is maybe more with the perceived focus of this whole project, where added features seem to be more important than true compatibility (with the 68060).
-
Hi Guys... regarding OS 3.9 not working with Vampire... or not being compatible... Installing OS3.9 was the first thing I installed on Vampire and I got it working, and I am by no means an expert.
try that will all boingbags etc...
my template works on ALL amigas 68000 (ok except the few libraries requiring 020.. but it BOOTS if you have enough memory) to 68060 EC or LC or RC. it works. and the 3.9 libs that are 020 (2 I think?) is the only CPU optimised software in this template.. (I have then anotyher dir to install cpu dependent versions if wanted...
but this template does NOT work on vampires..
-
There's always the risk of something not working as it should. I mean, even if I do most of my development using WinUAE, and pretty much know it emulates everything pretty good, I still do a lot of checking on actual hardware just to make sure.
But my main gripe is maybe more with the perceived focus of this whole project, where added features seem to be more important than true compatibility (with the 68060).
Fair enuff.
But the compability seems to be improving, based on the evolvement of how it runs Kioea for example. Hence Work In Progress :)
Seems to me that both compability and new "crappy" features is being tweaked as time passes. Ofcourse less "crap" would have increased the pace for the evolving of the compability, but it seems to me that they have found a good balance, espesially after Jari Eskelinen decided to teach himself Assembly thru working on FEMU.
And since Jari decided to put effort into FPU related code, the Apollo Team diverted their attention to assist him, which includes HardFPU for v2, and it looks like full HardFPU on v4.
-
Hi Guys... regarding OS 3.9 not working with Vampire... or not being compatible... Installing OS3.9 was the first thing I installed on Vampire and I got it working, and I am by no means an expert.
One thing I have noticed with the Vampire scene is that there is a big backlash by collectors of old hardware, they get really defensive about their 68060's and old gear. I used to have an 060, ppc card. The vampire trounces it, and a Vampire accelerator offers better value for dollar.
There are enough non haters to produce a Terrible Fire style 080 card that would be a nice addition to the range of Amiga classic stuff out there...
For those sitting on the fence the FPU stuff is a work in progress but most anything I ever ran on my 060 capable of running integer works with Vampire.
By releasing a free version I reckon a 68080 will become a standard. No one else is really producing, lots of talk and multi-decade projects with slow progress. Perhaps this development will inspire growth.
-
noone producing?? I see a lot of stuff happning with hardware..
-
But sorry. NO I guess the 080 will be as standard as the PPC..
some few datatypes etc. some small programs. and MAYBE a few demos. but that's it
-
but this template does NOT work on vampires..
You are doing it again. I take the freedom to correct your statement: your template does not work on your Vampire having an outdated core.
-
But sorry. NO I guess the 080 will be as standard as the PPC..
some few datatypes etc. some small programs. and MAYBE a few demos. but that's it
You do know that you can run 68k software on the Vampire? Unlike the ppc.
-
You do know that you can run 68k software on the Vampire? Unlike the ppc.
All system friendly software that follow CBM guidelines run also on PPC with the built in 68k emulator,'it run even faster!
For the rest you've an Amiga emulator.
Kamelito
-
If they really appear on the newest core release, then you should report those issues. And, if you want to make a good job while you are at it, try whether the same problem appears if you run the suspect code on WinUAE with 040-jit setting.
i guess you meant "040-no jit setting". thats how the binary can be properly tested. recompilation on the fly may hide issues. other than that whether you use 040 or 060 setting on uae to my knowledge, its something close to 040 thats being emulated anyway.
-
All system friendly software that follow CBM guidelines run also on PPC with the built in 68k emulator,'it run even faster!
For the rest you've an Amiga emulator.
Kamelito
Oh sorry i must have misunderstood chucky.
From his post about a few datatypes and smal programs i thought he was talking about PowerUP or WarpOS on classic AmigaOS.. (Not MorphOS or AmigaOS 4.x)
And with that point of view i cant understand how it compares. My PPC did stay mostly unused on OS3.9 but my Vampire is used for everything and all my 68k programs.
But from your point of view yes they are the same only the Vampire is quicker on 68k code...even faster than native PPC code in many cases :)
-
Does/might the added crap throw off timings and functionalities of current 680x0, or is it cause its an unknown variable as to how a routine will actually run?
The timing is off on purpose, which is forgivable, but you could end up writing code that only runs on the 68080 but fails on the 68060 because of race conditions (and vice versa). This can happen on real cpus as well, but the hostility from 68080 supporters indicates that a lot of people don't care if software only runs on 68080 and not 68060. So don't expect anything to work.
Any software taking advantage of the "added crap" can't run on current 680x0 in any case, again the hostility from 68080 supporters indicates they don't care.
Meanwhile anyone writing software that uses 680x0 features that aren't supported by 68080 gets hostility if they request it.
You can't put the 68080 into a nextcube or 68k mac. That doesn't seem to be something that apollo cares about, it's not really 68040/68060 compatible (next and mac require 68040 and won't work with 68060). They would be even better demos of their core, if only they'd support the FPU and MMU (it has been suggested that peoples requests for it are the reason why it hasn't been worked on, which is kinda weird).
-
There are hostile/agressive people on any side of any discussion, but that doesnt equate that any given side is overall of that demeanor.
Stereotyping any group based on the loud minority isnt really productive.
-
@Chucky
Good, grief, would you please shut your pie hole! You act as if someone is forcing you to buy a Vampire. If you don't like it and it doesn't suite your needs then don't buy it. Do us all a favor and buy an 060 card and drive off into the sunset.
@psxphill
It was never part of the Apollo/Vampire's design goals to support a Nextcube or a Mac so why do you even bring up such a ridiculous point? Next you'll be b@tching that it doesn't support your toaster oven.....
-
You are doing it again. I take the freedom to correct your statement: your template does not work on your Vampire having an outdated core.
If I sell a machine. I prep it with that card sp the buyer can poweron and use.. I have got many reports telling "ohjh I have issues after I use my vampire" so even new cores fails.
might not fail as much. but still fail.
-
Oh sorry i must have misunderstood chucky.
From his post about a few datatypes and smal programs i thought he was talking about PowerUP or WarpOS on classic AmigaOS.. (Not MorphOS or AmigaOS 4.x)
And with that point of view i cant understand how it compares. My PPC did stay mostly unused on OS3.9 but my Vampire is used for everything and all my 68k programs.
But from your point of view yes they are the same only the Vampire is quicker on 68k code...even faster than native PPC code in many cases :)
that is my point.. the bells and whistles of the 080 will be unused as the PPC was unused on most CSPPCs/BPPCs. (I just mine for datatypes and archivers.. not that much) so the 080 addons as they hype so much will never really be noticed.. if even used. as you need coders actually DOING the software...
-
@Chucky
Good, grief, would you please shut your pie hole! You act as if someone is forcing you to buy a Vampire. If you don't like it and it doesn't suite your needs then don't buy it. Do us all a favor and buy an 060 card and drive off into the
buy.. I got several. I also got stock of 060 cpus..
guess what I am doing instead? Developing a new.. opensourced 060 solution..
(takingh next steps from the A3660 I released 2 weeks ago already)
this is that I need to tell that the080 is NOT the future.. we are a lot that doesn't like it. and actually interesting enough. those I know who like the 080 is the noncoders. they who often does NOT like 080 are the coders..
so WHO will do the coding if they can't get the programmers on their side?
especially from a team that does NOT listen instead kick and ban?
-
So for this thread.. so far I am one of those who would BE a target of the offer.. as I could skip doing the A3660 Rev 2 and do he A3680 instead..
(or call it something different.. but you get the point)
but.. I simply do not dare putting all work into the hands of one person, AND the fact that .. retrocomputing for me is very very static.. and I try to tell what could make me change my mind..
luckly for me. I have still a quite good source of real 060 cpus so I will not have the issue. but yes. they are getting harder to source.. (even if I still buys them in packs of 10)
-
that is my point.. the bells and whistles of the 080 will be unused as the PPC was unused on most CSPPCs/BPPCs. (I just mine for datatypes and archivers.. not that much) so the 080 addons as they hype so much will never really be noticed.. if even used. as you need coders actually DOING the software...
Hmm still dont get your point. In that case wouldnt a 060 be the same? not much software is tuned for 060 either. Mostly a couple of demos and some datatypes..
You still argue in a way that i think you think the Apollo core is only AMMX and 64bit instructions.. and when not using that it it pointless???
If it runs 040 code faster then a real 68040 or 060 code faster then a 68060 is that so bad?
It has every instruction from the 68k motorola family included in the core. (Except TAS2 and MMU)
-
Hmm still dont get your point. In that case wouldnt a 060 be the same? not much software is tuned for 060 either. Mostly a couple of demos and some datatypes..
You still argue in a way that i think you think the Apollo core is only AMMX and 64bit instructions.. and when not using that it it pointless???
If it runs 040 code faster then a real 68040 or 060 code faster then a 68060 is that so bad?
It has every instruction from the 68k motorola family included in the core. (Except TAS2 and MMU)
"couple demos" there is a LOT of 060 demos.. and for me that is what I use my Amiga for.... (except coding on DiagROM :) )
the 080 still lacks FPU!.. when I bought my Vampire I read a promise about AGA and FPU.. what I understand. we will get AGA and a PARTIAL FPU. as there is no room for the fpu.. (but room for crap.. ie stuff that never will be used. instead of that real fpu....)
and no mmu.. and programmers needs mmu..
so.. a crippled cpu. that behaves more lika a buggy emulator (NO! I know FPGA is not emulation.. it BEHAVES like it!)
what do YOU guys use your beloved Amigas for? be HONEST now..
I would guess: games, demos and mostly just.. fiddeling with it..
production use? web? nah!
-
I bought an 060 card some years back just to watch demos, as thats my main interest with "retro" machines.
But guess what?
Most of the AGA/060 demos in recent years seem to be targeting overclocked 060 or WinUAE.
I watch the stream from Revision every year, then I run the demos on my Cyberstorm 060 and most of them run like crap compared to the stream.
Later I found out the specs of the compo AGA Amiga, which is overclocked!
So for me, the A1200 version of the Vampire (with full FPU) is great, since I'll be able to enjoy all these demos running smoothly, at half (or less) the price of an 060 card with latest mask CPU that can be overclocked.
-
most is aimed for 50MHz..
but I happen to know that the revision compomachine is clocked to 66MHz
my A4000 is clocked at 72.. my demoparty macine is 50MHz (also compomachine on the Edison demoparty here in Sweden)
I have done 100MHz overclocks to some CS MK2 to people as they want.
anyway. 50 is usually the maingoal for most demos anyway.
(elude always runs slower as they refuse to do the needed cheats.. so they "need" the 100MHz one :))
ofcorse fastermachine runs the stuff .. faster
-
"couple demos" there is a LOT of 060 demos.. and for me that is what I use my Amiga for.... (except coding on DiagROM :) )
060 demos is not that interesting for me so i guess i dont keep track of how many there are.
the 080 still lacks FPU!.. when I bought my Vampire I read a promise about AGA and FPU.. what I understand. we will get AGA and a PARTIAL FPU. as there is no room for the fpu.. (but room for crap.. ie stuff that never will be used. instead of that real fpu....)
Then someone lied to you or fooled you. Sorry to hear. For me when i bought it i was clear on that i bought what was available right then. Not something that could be available in the future.
As you are an FPGA expert and know the size of the different components in the core can you please enlighten us about what the "crap" is and how big that is?
SAGA?
AMMX?
64-Bit Support?
Instruction Bonding?
Instruction Fusing?
Big Instruction-Cache?
Big Data-Cache?
and no mmu.. and programmers needs mmu..
Im guessing DiagRom would have been impossible for you to do without an MMU?
so.. a crippled cpu. that behaves more lika a buggy emulator (NO! I know FPGA is not emulation.. it BEHAVES like it!)
what do YOU guys use your beloved Amigas for? be HONEST now..
I would guess: games, demos and mostly just.. fiddeling with it..
production use? web? nah!
Playing some games, listening to mods, looking at demos, Painting in DPaint, Configuring my WB to look like i want it, showing it of to friends and family and my daughter.
-
So I can fit your card in a500 or a600 ? You are comparing apples and oranges.... Or you are just a bit slow ?
Lovely excuse this.
So were all the vampirized A500 and A600 systems you speak of running OS3.9 before their owners downloaded and installed ApolloOS? I suppose all A500 and A600 owners has 020/030 boards already?
-
As you are an FPGA expert and know the size of the different components in the core can you please enlighten us about what the "crap" is and how big that is?
SAGA?
AMMX?
64-Bit Support?
Instruction Bonding?
Instruction Fusing?
Big Instruction-Cache?
Big Data-Cache?
no I am no FPGA expert.. but 64bit support etc must take space.. and if they can fit parts of the FPU. skipping all that bells and whistles must take space that the rest should fit..
Im guessing DiagRom would have been impossible for you to do without an MMU?
Diagrom is a VERY VERY VERY special thing. as it runs without any OS whatsoever.. as.. it is REMOVED.. there IS no kickstart.. there IS no possability to run that stuff.. without the debugger in UAE it would be more or less impossible to do surtain things..
Playing some games, listening to mods, looking at demos, Painting in DPaint, Configuring my WB to look like i want it, showing it of to friends and family and my daughter.
yes.. so.. simply: will not use AMMX etc etc you either...
as dpaint etc does not use it.. and as sources most likly are gone: will never use it.
so putting alot of effort AND FPGA space for.. "nothing"
-
You are doing it again. I take the freedom to correct your statement: your template does not work on your Vampire having an outdated core.
I use "gold 2.5", is that outdated too?
-
But sorry. NO I guess the 080 will be as standard as the PPC..
some few datatypes etc. some small programs. and MAYBE a few demos. but that's it
Datatypes supporting AMMX should already have been here by now, one would think.
And that is the biggest problem - independent development with Apollo Core is tricky, to get any knowledge you essentially have to sign a social contract with "the team", something most developers are not keen on doing, having seen how many have tried alreadt and gotten the burn.
MorphOS to a degree suffered from the same for a period, but at least documentation, toolkits, and a vast number of *software developers* (not hardware engineers) were available.
-
...yes.. so.. simply: will not use AMMX etc etc you either...
as dpaint etc does not use it.. and as sources most likly are gone: will never use it.
so putting alot of effort AND FPGA space for.. "nothing"
Electronic Arts released the source code for version 1 so they probably do have the other source code archived.
DPaint is not developed anymore AFAIK but the DPaint inspired GrafX2 was very recently ported to Amiga 68k and that uses SDL. I believe SDL has or is being optimised for the 68080.
-
Then someone lied to you or fooled you.
And that might very well have been the official wiki faq. I pointed out several times that it was incorrect, and it was changed several times too, but always leaving the impression that *some* FPU was right around the corner, even at times when Gunnar expressed on IRC that he was fed up and rather saw no FPU core on V2 at all.
-
Electronic Arts released the source code for version 1 so they probably do have the other source code archived.
DPaint is not developed anymore AFAIK but the DPaint inspired GrafX2 was very recently ported to Amiga 68k and that uses SDL. I believe SDL has or is being optimised for the 68080.
Or they just randomly found that source and. heck we release this..
btw what C compiler supports the apollo-stuff?
-
And that might very well have been the official wiki faq. I pointed out several times that it was incorrect, and it was changed several times too, but always leaving the impression that *some* FPU was right around the corner, even at times when Gunnar expressed on IRC that he was fed up and rather saw no FPU core on V2 at all.
yes an as it changed.. well. I lost trust. but OK gunnar now say: you buy it as it IS.. so true..
as it IS now is: no fpu, no MMU (and NO AGA!) (YES I was at A32 I saw it working. but it is not a release.. so "as it is" for us.....
-
Anyway I lovethis:
- "Vampire is most compatble"
-- "Ehh this does not work"
- "you are doing it wrong!!!" (when it shuld be: "oh damn. we have an issue here"
- "It is most compatible"
so even if you present incompabilities... there is deniers...
and my point is: with every change: you introduce NEW ways of being incompatible..
-
And that might very well have been the official wiki faq. I pointed out several times that it was incorrect, and it was changed several times too, but always leaving the impression that *some* FPU was right around the corner, even at times when Gunnar expressed on IRC that he was fed up and rather saw no FPU core on V2 at all.
Well, Im sure you know that Gunnar and the team got a roadmap of sorts they follow, but IF someone steps up and offer real assistance for a certain aspect, they will help out/tweak their roadmap.
Jari decided to do the SoftFPU, which set him apart from most of us; instead of just talking/posting on forums, he decided to DO something/code.
This spurred the team to help him, effectively Jari convinced them to change their roadmap cause he put REAL effort into enhancing the Apollo/Vampire expirience.
If YOU had done the same, Im sure they would reconsider/tweak their focus, but we that offer nothing but words cant expect them to upend their direction.
As for Chucky; You realise that the SoftFPU isnt a real problem IF Gunnar/team finds enough space on the V2 for the performance critical aspects to put on HardFPU. Gunnar did say to effect "dont count on HardFPU on v2, currently its internal build. Hopefully it pans out" etc. Im paraphrasing tho. The performance increase of Kioea is evidence of the improvements.
Why you seems to consciously ignore this aspect is puzzeling to me. Britelites comments makes sense to me. He is concerned with the POTENTIAL issues running routines on a real 060 vs 080. Thats a fair point, but since its Work In Progress those potential compability issues might be adressed.
-
Anyway I lovethis:
- "Vampire is most compatble"
-- "Ehh this does not work"
- "you are doing it wrong!!!" (when it shuld be: "oh damn. we have an issue here"
- "It is most compatible"
so even if you present incompabilities... there is deniers...
and my point is: with every change: you introduce NEW ways of being incompatible..
then it is. People decide if they want to spend their money on this "incompatible" platform and up to now most people are very happy with it (with few exceptions of course). Everybody who is not happy with it can sell it again with no losses, I assume that need is above supply at the moment. Keeping it and constantly moaning is strange to me, even more so if you do not own it and not interested to buy it but still moaning. Guys, search for another constructive hobby please. It is boring...
-
Jari decided to do the SoftFPU, which set him apart from most of us; instead of just talking/posting on forums, he decided to DO something/code.
And alb42? Did he not code?
I don't contribute code to closed source projects in general, and certainly not to a project that has the intention of hijacking this platform hardware wise.
-
Ofcourse he code.
But towards the core?
I would have to re-read the thread a while back, but wasnt it FPU that he asked for, which is what Jari has contributed to.
It did finally arrive, but wether or not its to the precision he needed, he would have to answer.
Im fuzzy on the timeline, but I belive Jari entered the scene after alb42 expressed misgivings about how he was treated. I might be wrong. Im sure someone will take the time to check :)
And ofcourse; the SoftFPU+HardFPU combo is currently for the betatesters/Apollo Team only, so alb42 wouldnt have been able to fully utilize his programs potential.
As for hijacking the platform. Really? Are we now hoping that NO new efforts are going to have any measure of success, so we can complain about the lack of development for decades to come...? With the exceptions of those that already have their 060 cards with good masks for overclocking...? So we can spend 100s (1000s?) of dollars on mediator boards plus addons to get a somewhat modern usability.
As I said to Britelite; if you dont want to utilize all the added features of the Apollo/Vampire, ignore them, and code for 1260 in mind. Ofcourse Britelites counter to that was uncertainty of compability, and thats a fair argument, which can be taken care of.
-
I am keeping mines thank you very much, who knows, maybe there will be alternative cores in the future, maybe even the Apollo Core "shapens up" over time. Not like I don't have loads of other cards around that I am also not selling just because.
-
As for hijacking the platform. Really?
Eh, yes, really, the goal was/is to make Apollo Core the de-facto CPU for Amiga.
Are we now hoping that NO new efforts are going to have any measure of success, so we can complain about the lack of development for decades to come...?
Nope, we are hoping for something that is more compatible (on so many levels, not just hardware and software, but also socially) and more developer friendly.
-
Eh, yes, really, the goal was/is to make Apollo Core the de-facto CPU for Amiga.
Nope, we are hoping for something that is more compatible (on so many levels, not just hardware and software, but also socially) and more developer friendly.
You are a Lord now I guess ;)
You are using "we" obviously if meaning I ;)
There are different groups in the community obviously but the biggest group seems to be happy with vampire/apollo. Finally customers decide even if you dislike that
-
De-factor standard? You saying that like its the only option. Ofcourse if you said that in Saudi Arabia regarding a decreet from the Saudi family, Im sure that the implication of discrimination your are putting forward would be valid.
But given that people are free to choose based on what the hardware/software delivers, the quality of Vampire/Apollo will decide wether or not it becomes what "most amigans" use.
By using hijacking, you are giving the platform a negative attribute that doesnt hold water.
Ofcourse, anyone that puts immense amounts of hours into a product hopes it succeed, which means alot of "amigans". But noone is being flogged publicly for not purchasing said platform.
Socially? Compared to what?
On the amigahardware scene its hard to find more accessable developers than the Apollo Team. I do love TF/PlasmaB thread/youtubes ofcourse, so Im not putting down any other developers efforts.
Does Gunnar have unlimited patience with people? No, and I dont always agree with his social interactions. But I havent put in the crazy amount of efforts into the Apollo Core, so its extremely easy for me to sit on my high horse and judge him when he looses his temper.
If I were in his shoes I would probably have lost it much quicker.
Ofcourse, if the current version of the core isnt accomodating what you need from it; fine, I can see that. But since its currently a moving target/work in progress; passing final judgement like some do based on the current specs is unfair.
-
You are using "we" obviously if meaning I ;)
I was responding to the patronizing "we" that Niding was using, obviously.
-
I was responding to the patronizing "we" that Niding was using, obviously.
??
-
??
https://en.wikipedia.org/wiki/We#The_patronizing_.22we.22 :)
-
All I know is I was excited about the release of it, I helped Magista get a parts order so he could complete V2 500+ orders faster and bought my own to test. It got less compatible not more compatible and when I had problems with it I was brutally attacked on the forums and on the facebook page. They admins in the facebook page even kicked me out after I made a post asking for help in reflashing it. Once I did get it working again I got rid of it. I love helping out, but dont need drama like that and wish them all the best luck.
Anything negative is shut down and they hide the complete truth. That is my issue with this team and the project. The hardware is amazing, but those running the show are not so amazing. They take things way to personal.
-
Jari decided to do the SoftFPU, which set him apart from most of us; instead of just talking/posting on forums, he decided to DO something/code.
This spurred the team to help him, effectively Jari convinced them to change their roadmap cause he put REAL effort into enhancing the Apollo/Vampire expirience.
I wouldn't code SoftFPU because I think it's a backward step, no offence for the work put into it but you shouldn't need any software patches/shims/etc if the CPU is as compatible as they claim.
If gunnar opened the 68080 source to let other people work on the HardFPU then that would be a different matter. He would fine other people willing to help out. He has put himself in an untenable position because he can't deliver what is needed, but he wants to keep ultimate control over it.
Bad management with a promising product but they don't know what to do with it, seems a perfect fit for Amiga. The curse strikes again.
-
What i read on the apollo core is that the hard fpu don't fit on the vampire FPGA RTG + AGA + CPU seem to have filled up.
The full FPU will be only for the new vampire 4.
From my point of view i have the vampire 2+ i if could choose between SAGA or the hard FPU i could prefer to have the AGA support.
-
maybe if they remove the cpu stuff that is "crap" you know.. AMMX, 64bit,. hyperthreadin bnlahblahneverbeingusedanway..
Shortly: first you make the stuff you want it to be compatible with. and THEN you add the "goodies"
-
@psxphill
V2 has limited space for FPU. It might be able to fit a full HardFPU, or it might not. Which is why Soft and HardFPU combo is intresting to provide owners of this version of Vampire this functionality.
V4 can supposedly fit the full HardFPU.
So how can Jaris work be a waste of time, when it gives the team a tool to compensate for the limited FPGA space.
@Acill
Who treated you badly on the facebook page? Some of the "Vampire loyalists" or the team itself? If you read my comment on Vampire forum and IRC, I have misgivings with both sides of the aisle with regards to bent/broken social antennas.
If its "loyalists", then you shouldnt really attribute the drama they drum up to the Team itself. Imho.
@Chucky
You already know they have a "roadmap", and only change that if someone (like Jari) joins in with their own efforts to contribute directly to the development. If you want them to divert from the "crap" maybe you should divert their attention by contributing something worthwhile, dragging their attention from their original roadmap? ;)
-
All I know is I was excited about the release of it, I helped Magista get a parts order so he could complete V2 500+ orders faster and bought my own to test. It got less compatible not more compatible and when I had problems with it I was brutally attacked on the forums and on the facebook page. They admins in the facebook page even kicked me out after I made a post asking for help in reflashing it. Once I did get it working again I got rid of it. I love helping out, but dont need drama like that and wish them all the best luck.
Anything negative is shut down and they hide the complete truth. That is my issue with this team and the project. The hardware is amazing, but those running the show are not so amazing. They take things way to personal.
It is propably a question of character partly but I assume if you invest 10 years in your dream to do something modern to benefit everyone and then you are constantly critisized and urged to make things different by people that never contributed anything you can become thin skinned propably.
-
@psxphill
V2 has limited space for FPU. It might be able to fit a full HardFPU, or it might not. Which is why Soft and HardFPU combo is intresting to provide owners of this version of Vampire this functionality.
isn't it very VERY ironic that you need to add software to get functionality. from the team that say that the 68040/68060 is crap as you need 68040/68060 library to add functionality....
You already know they have a "roadmap", and only change that if someone (like Jari) joins in with their own efforts to contribute directly to the development. If you want them to divert from the "crap" maybe you should divert their attention by contributing something worthwhile, dragging their attention from their original roadmap? ;)
Contribute.. well people asked for FPU.. I read it in forums. result was threatening of ban etc as NOONE NEEDED THAT!.
so with that attitude.. you know that they are not interested of input.
same here.. I have issues with lack of mmu etc. gets reported as OFF TOPIC... meaning: NOT INTERESTED!
so. I contribute with a cheapass (bad) 060 cpucard for bigbox.. (currentlty working on Rev2.. will not be as cheapasscrap)
and I AM interested of input.. AND help.. doing amigastuff is not a one-man show.. maybe one man does it. but there is always a lot of input to take.
-
I wouldn't code SoftFPU because I think it's a backward step, no offence for the work put into it but you shouldn't need any software patches/shims/etc if the CPU is as compatible as they claim.
If gunnar opened the 68080 source to let other people work on the HardFPU then that would be a different matter. He would fine other people willing to help out. He has put himself in an untenable position because he can't deliver what is needed, but he wants to keep ultimate control over it.
Bad management with a promising product but they don't know what to do with it, seems a perfect fit for Amiga. The curse strikes again.
No Gunnar explained it... the FPGA used up to now is too small to include everything including the full FPU. So now there is the softFPU and some FPU commands might transferred to hardware later but that depends what space is available.
-
isn't it very VERY ironic that you need to add software to get functionality. from the team that say that the 68040/68060 is crap as you need 68040/68060 library to add functionality....
Does it matter how a functionality is added, as long as it works? If it causes problems, like Britelite is concerned about, then fine. Then its a valid point to raise. But given how the performance has increased in relativly short time, the refusal to accept the functionality is soon approaching pendatic at this point.
Contribute.. well people asked for FPU.. I read it in forums. result was threatening of ban etc as NOONE NEEDED THAT!.
so with that attitude.. you know that they are not interested of input.
You know, Jari never demanded or argued about anything when it came to the FPU. He just started coding, and asked for help. Instead of rolling with the "omg, FPU isnt included. USELESS!!". How is that constructive!?
I denfintly would become hostile after a while of such attitude if people offer nothing but insults over the hard work.
same here.. I have issues with lack of mmu etc. gets reported as OFF TOPIC... meaning: NOT INTERESTED!
Have you considered doing like Jari? Just start coding, and then ask for help?
Posting negatively about the lack of features (some which has been adressed in 2.7) for pages on forums does not count as contributions. And again, it would defintly piss me off to after months of it.
so. I contribute with a cheapass (bad) 060 cpucard for bigbox.. (currentlty working on Rev2.. will not be as cheapasscrap)
and I AM interested of input.. AND help.. doing amigastuff is not a one-man show.. maybe one man does it. but there is always a lot of input to take.
Good luck with your efforts, and I actually mean that :)
-
Does it matter how a functionality is added, as long as it works? If it causes problems, like Britelite is concerned about, then fine. Then its a valid point to raise. But given how the performance has increased in relativly short time, the refusal to accept the functionality is soon approaching pendatic at this point.
well they use the argument that the 080 is so much better as 040/060 lib is not needed.. while to add stuff peopoe ask for, they are doing the same (except here it is not a library.. but still software)
Have you considered doing like Jari? Just start coding, and then ask for help?
Posting negatively about the lack of features (some which has been adressed in 2.7) for pages on forums does not count as contributions. And again, it would defintly piss me off to after months of it.
the lack of MMU cannot be done in software.. that is the big issue here.
Good luck with your efforts, and I actually mean that :)
thanx.
-
well they use the argument that the 080 is so much better as 040/060 lib is not needed.. while to add stuff peopoe ask for, they are doing the same (except here it is not a library.. but still software)
Well, given the fact that they have limited space on the V2, the capability of the 080 doesnt come into question, merely the space on the FPGA.
You could argue they have wasted space on new functionality, but thats a fundemental design choice going forward, in addition to an increasing degree of compability.
What remains in that regard will be seen at release of 2.7. And finally 3.0.
If that still doesnt suite you, fine, your own design, Jens, PlasmaB, MikeJ etc could cover those needs.
-
Putting in new stuff that would REQUIRE software to be rewritten/recompiled but not stuff that oldf software already support?
and also. ironic enough. not putting in stuff that HELPS developing new software. it is like.. like. releasing a new cpuarchitecture for the amiga and not release programmingsoftware with it.. (PPC anyone?? :))
but I guess gunnar had this wet dream of MMX on the Amiga etc. so he completley forgot.. the rest of the worls.. yes. his design..
but then. WHY THE HYPE? especially for stuff that will be "crap" even if they might be good?
well. yes. this IS repeating..
-
AMMX is kind of cool, you can watch movies on the amiga i have a few converted and look great.
-
Putting in new stuff that would REQUIRE software to be rewritten/recompiled but not stuff that oldf software already support?
and also. ironic enough. not putting in stuff that HELPS developing new software. it is like.. like. releasing a new cpuarchitecture for the amiga and not release programmingsoftware with it.. (PPC anyone?? :))
but I guess gunnar had this wet dream of MMX on the Amiga etc. so he completley forgot.. the rest of the worls.. yes. his design..
but then. WHY THE HYPE? especially for stuff that will be "crap" even if they might be good?
well. yes. this IS repeating..
You dont need to recompile old software unless you want to utilize the full potential of the Vampire for that particular program.
And for the most part, Vampire will accelerate programs, and give the user more memory than has been the mainstream standard beyond relativly few that got into the higher end accelerators on the classic.
Again; you can just use the Vampire as a 1260 accelerator and ignore the "fluff".
-
you NEED to recompile to use AMMX etc.. so if noone needs to recompile that noone needs to implement AMMX!
and no.. I cannot use the vampire as a 68060 accelerator..
it lacks the FPU
it lacks the MMU!
it is a 68EC060. I THOW cpus like that in the bin.. I have some that I use when I do 040-060 adapters. so I did not do a %&$#?@!%&$#?@!%&$#?@!%&$#?@!up sending 5V to the cpu frying it.. (and as it is EC. it is by design broken anyway so)
-
Ok.
Its been repeatedly pointed to that FPU will be implimented in 2.7.
Your refusal to accept any input tells me you are a waste of time to discuss with.
Im checking out of this thead.
-
PARTS of the FPU! is not the cpu.
aybe if EXACT the same parts as 68060 have..
and.. the vampirepeople have told that you get the vampire for what there is. there IS no fpu now.. as the version you say.. is closed for team only and we never know if it of some odd reason gets scrapped..
-
Ok.
Its been repeatedly pointed to that FPU will be implimented in 2.7.
It was repeatedly pointed out that we don't need and won't be getting a compatible FPU. Then there was a u turn, but the last I heard they were only going to support 64 bit floats, not 80 bit. Again you don't need 80 bit and won't be getting 80 bit etc etc, maybe there has been another u turn but nothing has been released.
Apparently we still don't need and won't be getting a compatible MMU.
Problem is gunnar spends fpga gates on features he wants to do so there aren't enough for the things that we want. To defend this he has essentially created his own cult.
He wants to license it, but he pisses too many people off with his attitude & then plays the victim. So we'll see how that goes.
-
Problem is gunnar spends fpga gates on features he wants to do so there aren't enough for the things that we want.
Well this is his project, he want to make a modern implementation of a 68000 processor and he is using the Vampire and this shareware core as a development vehicle. I suppose AMD shouldn't have introduced 64-bit processors either by the poor logic that seems to be thrown around in this thread. The bottom line is the Vampire is out there, the Apollo core is being lead by a Processor designer that wants to do new things, yet still strives for full Amiga compatibility. It's work in progress and it aims to be better in the long run.
So bitch and moan all you want, although I wish you people wouldn't, it's already here. get over it, it's apart of the Amiga community now.
-
thing is that he is aprox 20 years too late..
so comparing to AMD and 64bit is poor.. as it was (is) in a time when it is developed... the amiga, 68k series etc died in the 90s.. we who are stupid enoufh to use those machines use it for nostalgic reasons. with 20+ year old software (with most lost sources) that will never be able to use all development anyway.
-
thing is that he is aprox 20 years too late..
so comparing to AMD and 64bit is poor.. as it was (is) in a time when it is developed... the amiga, 68k series etc died in the 90s.. we who are stupid enoufh to use those machines use it for nostalgic reasons. with 20+ year old software (with most lost sources) that will never be able to use all development anyway.
And yet its still here...
Can you point me to a forum about your projects, so I can relentlessly criticize and dismiss them with little knowledge of what is actually going on like you do here?
-
Difference heer is that Gunnar more or less say that his product is thebest, no flaws. it is the ULTIMATE one..
however. it might be pretty awesome. but no it is not the ultimate one and when people asks about stuff we had for 20+ years.. the %&$#?@!%&$#?@!%&$#?@!%&$#?@!storm starts.
it is all like a religious sect about a very VERY hyped product. DO NOT SAY NO TO THE LEADER.
What i have the issue is like the thread:
* Compatible with 68040/68060 while it isn't. it sure looks like it.. but not compatible.
* it is closed..
* They are basically searching for partners. but any partner should just accept and do not ASK. (as it might be off topic.. etc)
* you say: there is NO MMU! answer is: YES IT IS! but.. closed.. (so.. no MMU!)
and they let vampire V1 users in the cold when v2 was out.. V2 was released you will get 080, the most compatible cpu, AGA and fpu.
then, suddenly. no fpu. until someone made a sofwareeulation. now parts of it will be there. but not all. as there is no space suddenly. (as it apparenlty is filled with stuff noone asked for)
all my stuff is open and free, just add it yourself. you do not need to ask. maybe say what you want to add so we can agree of a method of implementation.. this is closed stuff. then you really need to take the input
for my stuff. well. do your changes, do a pullrequest and if it looks ok.. I will accept it. if you do not accept my deny, do a fork.. done. problem solved
but my stuff sure ISN'T the greatest.. code is messy (as donemostly while drinking beers) but can be changed by more or less anyone
As told many times: if this was a 68060 compatible solution (and also not a one man show). damn I would close my eaglewindow on my left screen and redo all with that now..
remeber.. this ALL started as I just told that they should add MMU! it ALL created this %&$#?@!%&$#?@!%&$#?@!%&$#?@!storm. DO NOT ASK FOR STUFF!
-
DO NOT SAY NO TO THE LEADER.
Ah yes, first rule of the Apollo Cult...
* Compatible with 68040/68060 while it isn't. it sure looks like it.. but not compatible.
Okay, you have any proof of said incompatibilities out side the tired FPU or MMU argument? It's striving to be fully compatible. the FPGA nature of the core allows updates to said core.
* it is closed..
Just like most processors, supported instruction sets and such are openly documented.
* you say: there is NO MMU! answer is: YES IT IS! but.. closed.. (so.. no MMU!)
I'm sure it could be documented in the future when the core is more solidified.. (see my turn to be dismissive)
and they let vampire V1 users in the cold when v2 was out.. V2 was released you will get 080, the most compatible cpu, AGA and fpu.
then, suddenly. no fpu. until someone made a sofwareeulation. now parts of it will be there. but not all. as there is no space suddenly. (as it apparenlty is filled with stuff noone asked for)
The original vampire was never intended to be a commercial item, just happened to become one after the fact. And the FPU argument again... how many times are you going to revisit this thing. FPGA can get updates, gold 2.7 HW+SW hybrid solution is quicker than anything else out there. Then you will try and say it's not out yet...
all my stuff is open and free, just add it yourself. you do not need to ask. maybe say what you want to add so we can agree of a method of implementation.. this is closed stuff. then you really need to take the input
for my stuff. well. do your changes, do a pullrequest and if it looks ok.. I will accept it. if you do not accept my deny, do a fork.. done. problem solved
but my stuff sure ISN'T the greatest.. code is messy (as donemostly while drinking beers) but can be changed by more or less anyone
As told many times: if this was a 68060 compatible solution (and also not a one man show). damn I would close my eaglewindow on my left screen and redo all with that now..
remeber.. this ALL started as I just told that they should add MMU! it ALL created this %&$#?@!%&$#?@!%&$#?@!%&$#?@!storm. DO NOT ASK FOR STUFF!
Okay....
Is this an FPGA solution, are you doing processor design? or is it just some accelerator using the limited stock of existing processors that will likely be very scarce to source in the near future?
-
Ah yes, first rule of the Apollo Cult...
yeah! THAT you notice.. for sure!
Okay, you have any proof of said incompatibilities out side the tired FPU or MMU argument? It's striving to be fully compatible. the FPGA nature of the core allows updates to said core.
How about read in the threads here.. a 3.9BB4 setup that works on *ALL* amigas indpendent of cpus fails on vampire.. after working with it for long I just ditched it.. actually found my Vampire today looking through my scrap-box..
Is this an FPGA solution, are you doing processor design? or is it just some accelerator using the limited stock of existing processors that will likely be very scarce to source in the near future?
no I am not. but if it would be open. there would be some kind of community. always someone actually adding.. and there have been many request of JUST FPU and MMU they have been denied. AND also users asking for it. threatend of blocking or even blocked on the forum.
it is a horrible dictatorship that destroys this product that could be the solution for the cpu-shortage there is.
Actually.. I do not know so many advanced users using vampires/080. I know a lot of people using it in thir A600 for whdload.. they haven't even connected the HDMI cable. (same on A500.. as it adds a way of getting HD pretty nicly)
and oh! people run sysinfo and doom/quake
but even if I know a lot of coders (ok mainly democoders!) NONE of them use a 080.. and I met people at A32. walking around vampirepeople saying to them "yeah it is nice" etc. but when being alone from then. "but.. no MMU why should I do programming wityout the tools I had for 20 years)
It is the CONSTANT attack of people asking for stuff that is so damn scary! this thread is a perfect example of it. YOU SHOULD NOT SAY ANYTHING BAD! YOU SHOULD NOT ASK! YOU OBEY!
and now searching partners with that attitude?
damn so sad that all nice stuff we have on the amiga fails of.. attitude or managment. (and as it always closed and the amiga HAD it with closed stuff)
and then dies of it in the end...
-
I don't know much about that mmu and fpu stuff, but what i see for sure is that they are no other options for a plain Amiga 500 to get that amount of upgrades and features for the price point the vampire 2 offers.
For example when i purchased it theres no networking.
But now i got the sdnet support on the core 2.5, the module for cost 5 USD that kind of upgrades are quite nice and even unsuspected for me.
I don't have see much products that offer this kind of flexibility that you can just do a download an upgrade your hardware.
Thanks to the vampire i could get back on the Amiga, apart from using an emulator that was i was doing before.
-
YES you are right.. and THAT is why it is so damn sad that they do not listen..
this could be the best thuing there is. hardware is totally AWESOME..
problem is the strange path they are taking, totally forgotten every damn issue we had whe there was the SLIGHTEST change in cpu or chipset register.
I mean. if it was ANY motorola cpu done perfect AGA done exact and RTG and AHI .. it would be so damn freaking awesome.
but trying to do a "new amiga" introducing new ways of software to fail. is just so sad.. Sorry but noone will use their amiga as their main platform again.
AMMX etc etc will not be used as i would guess 90% of all software we use, the sources have been lost forever and cannot be re-written or recompiled.
or they need to be rewritten, but reember WITHOUT using the debuggingtools people are used to......)
new screenmodes and stuff in chipset later. NO softtware will support it. NONE. NADA! but with RTG we have software already that supports it.
not available in RTG API!. well create a new API for it. let more peole add stuff that way. (Warp3D etc etc)
16 bit audio. same thing there.. NO software supports it.. AHI is the solution.
so if they just listen it would be an awesome product..
so it is so sad. it could be so, so awesome.. the vampire that is.
and this offer of licesing out for HW developers? well. they just say MMU and %&$#?@!%&$#?@!%&$#?@!%&$#?@!storm starts here AND on a1k.. it is not accapted to ask about things. HOW can you then start a cooperation when there are tension the first 10 minutes over a legit question!.
-
(even IF apparently asking for a MMU in a cpu. in a thread of a cpu-replacement is .. eh.. OFF TOPIC apparently)
-
I'm sure they are going to fix most of the issues, this things take time.
Waited almost a year to have the vampire on my hands after preorder.
So patience is the name of the game.
-
well.. not if they ban you asking for a function.. etc..
takes times. OH YES! I know it does.. but then (shortend):
1. oh. FPU would be nice
2. NO! NOONE USES IT!
3. Eh. I do?
4. BAN!
is the wrong way, instead of:
1. Oh FPU would be nice.
2. Does anyone use it?
3. yeah! I do!
4. well. if there is enough requests maybe..
so it is the attitude.. it is so sad. Gunnar might be a very intelligent person (and doing this. he must be) but.. social and contact with its users.. well doesn't work)
I met Majsta etc at Amiga32.. we had a short talk. and he was nice to talk to. and I think he understood my point. it is the persons around, not listening that does the damage.
it can do so much damage being a big ego and not take input. you do not need to implement all. but you do nto need to say that your product is the bestm if noone can even .. eh. use it?
-
I have always wondered why people in Amiga scene want to comment things that they are not interested? Previous example was Amiga OS4, every thread got tens of post from MorphOS people who are not interested Amiga OS4.
Seems that Vampire is now in that position.
I'm not interested Vampire, develobment took wrong course. For me that is end of story. I usually just don't read Vampire threads.
What is wrong with you people????
Only thing wich I have against Vampire team, is that they named their product to Apollo. Now it is very annoying to look anything (real) Apollo accelerator related from internet.
***************************************************************
There should be limit how many times one person can comment to one thread.
***************************************************************
Commenting 100 posts of nonsense is just pointless and ruins interesting news threads.
-
YES you are right.. and THAT is why it is so damn sad that they do not listen..
It's not that they don't listen, it's just the goal of this project has a long term goals, and it may be different than what everyone else wants. But full compatibility is part of that long term goal.
problem is the strange path they are taking, totally forgotten every damn issue we had whe there was the SLIGHTEST change in cpu or chipset register.
It's why it's a living project, the end user tests things, reports back issues and hopefully they work out the problem... literally there are people testing the hell out of the beta cores.
I mean. if it was ANY motorola cpu done perfect AGA done exact and RTG and AHI .. it would be so damn freaking awesome.
That is essentially the goal of the Gold 3 core update...
but trying to do a "new amiga" introducing new ways of software to fail. is just so sad.. Sorry but noone will use their amiga as their main platform again.
AMMX etc etc will not be used as i would guess 90% of all software we use, the sources have been lost forever and cannot be re-written or recompiled.
or they need to be rewritten, but reember WITHOUT using the debuggingtools people are used to......)
Honestly I was about to get a Furia when I saw information on the Vampire... the concept of writing code for AMMX and breaking ground a new 68000 architecture with 64-bit instructions had me intrigued.
new screenmodes and stuff in chipset later. NO softtware will support it. NONE. NADA! but with RTG we have software already that supports it.
not available in RTG API!. well create a new API for it. let more peole add stuff that way. (Warp3D etc etc)
SAGA uses RTG... it works just like a Picasso device... RTG software works fine without any tweaking on the Vampire.
16 bit audio. same thing there.. NO software supports it.. AHI is the solution.
It's fully Paula compatible, and likely will have an AHI support, the team consists of people with driver writing experience.. not sure where the argument is.
so if they just listen it would be an awesome product..
so it is so sad. it could be so, so awesome.. the vampire that is.
and this offer of licesing out for HW developers? well. they just say MMU and %&$#?@!%&$#?@!%&$#?@!%&$#?@!storm starts here AND on a1k.. it is not accapted to ask about things. HOW can you then start a cooperation when there are tension the first 10 minutes over a legit question!.
Well it IS an awesome product, only people I hear complain about it are people that don't own one. I suppose I never experienced the Fast 68060 or PPC accelerated era of Amiga, I just played games from floppy drives. The Vampire gives me something I never had, so it's awesome for me. As for the licensing issue, it's Gunnar's project, he hopes to have an ASIC created eventually that could reach more moderns speeds, I can see why he wants to protective of it, but still want to give something to the community, even if that community has people that feel threatened by it... so...
-
That is essentially the goal of the Gold 3 core update...
no. they apparently changes stuff.. so not FULL AGA done exact!
done exact means ALL restrictions mustbe there, AND all bugs! as it might be used in.. well. demos etc.
Honestly I was about to get a Furia when I saw information on the Vampire... the concept of writing code for AMMX and breaking ground a new 68000 architecture with 64-bit instructions had me intrigued.
I got my Vampire as I guessed it sould be the killerthing.. and I got a nice offer so yeah I take it now. I will get the fpu etc later..
this however seems not to fully happen.. I guessed I could make an A500 an awesome AGA demomachine later.. but. .no.
and it should be so compatible.. it wasn't. it ended in yes.. it runs sysinfo nicly.. but.. nah. this doesn't even feel amiga. (and this comes from me using WINUAE mostly when coding)
SAGA uses RTG... it works just like a Picasso device... RTG software works fine without any tweaking on the Vampire.
well yes! but they also call the AGA addons "SAGA" just to confuse even more.. and that will just introduce more issues.
It's fully Paula compatible, and likely will have an AHI support, the team consists of people with driver writing experience.. not sure where the argument is.
no it isn't as they apparently added stuff. and. apparntly via chipset.. sigh! WHYYYY!..
Well it IS an awesome product, only people I hear complain about it are people that don't own one. I suppose I never experienced the Fast 68060 or PPC accelerated era of Amiga, I just played games from floppy drives. The Vampire gives me something I never had, so it's awesome for me. As for the licensing issue, it's Gunnar's project, he hopes to have an ASIC created eventually that could reach more moderns speeds, I can see why he wants to protective of it, but still want to give something to the community, even if that community has people that feel threatened by it... so...
I own an V500 so I have tested it. and.. it is not in my box of "well crap pcbs" haven't even bothered to remove the CF adapter on it :)
if this was in the beginning of 2000 it would be aswesome.. now it is too late for the new cpustuff etc.
Would ANYONE accept that people did changes like this on the C64? not if there is not "go64" mode faster cpu for crunching etc maybe. (butthat is done on a PC today..)
Anyway. it is time to stop.. vampire fanatics will never understand my point, I understand theirs.. but their hope will fail as the team apparntlty does not want the programmers to use their platform not suppling the possability of using their debugtools.
only difference with the PPC I got in my Amigas (both of them) is that the 080 doesn'tneed to switch to a different cpu and flushing caches etc but the new stuff will be there. unused and sad.
so number of software supporing the 080 stuff I guess will be aprox the same as the PPC. (and that is very low. damn I should sell my CSPPC and BPPC :)) while consuming space making you need to softwareemulate other stuff you MIGHT use..
same as 68040/68060 library. but THAT is bad.. FPU emulator is not 68040/68060.library then it is ok... weird..
his project. sure.. yes! true.. it IS! if he want to do bananas shooting out the screen every 10th minutes. he can do that. but if someone thinks that is a bad idea maybe not be angry of it.. especially when asking for cooperation..
-
@Kolla
You where in the frontline of FPU fighters, mentioning it in every needed and unneeded topic, ranting about it, demanding....now you have some other life goal ?
No, I'd still like a better GPU solution than the half assed one that the V2 has.
And when the V4 is actually available, THEN you can tell me what it features, because right now that's meaningless.
In the meanwhile, fanatics like you are the people than ought to be banned.
-
I met Majsta etc at Amiga32.. we had a short talk. and he was nice to talk to. and I think he understood my point. it is the persons around, not listening that does the damage.
LOL. Majsta wasnt at Amiga32.
-
LOL. Majsta wasnt at Amiga32.
Maybe it was an FPGA-not-quite-100%-compatible version of Majsta?
;)
Plaz
-
no. they apparently changes stuff.. so not FULL AGA done exact!
done exact means ALL restrictions mustbe there, AND all bugs! as it might be used in.. well. demos etc.
Gold 3 core is still in internal debugging likely release is 1Q 2018, IE it's still being worked on and getting tested thoroughly. It's not a one man operation, several people are helping develope it. Why do I get a feeling you still don't understand that an FPGA is programmable?
I got my Vampire as I guessed it sould be the killerthing.. and I got a nice offer so yeah I take it now. I will get the fpu etc later..
this however seems not to fully happen.. I guessed I could make an A500 an awesome AGA demomachine later.. but. .no.
and it should be so compatible.. it wasn't. it ended in yes.. it runs sysinfo nicly.. but.. nah. this doesn't even feel amiga. (and this comes from me using WINUAE mostly when coding)
ugg not sure exactly where you are going with this, don't know which things didn't work. I haven't seen an OCS/ECS title that didn't run with WHDLoad, all my floppy games work as intended too. Most things that use RTG work fine too... It's likely just your setup.
well yes! but they also call the AGA addons "SAGA" just to confuse even more.. and that will just introduce more issues.
So programmer/engineers are bad at naming things, nothing new.
no it isn't as they apparently added stuff. and. apparntly via chipset.. sigh! WHYYYY!..
You can add things without breaking compatibility, also you can just use the real paula if you want to...
I own an V500 so I have tested it. and.. it is not in my box of "well crap pcbs" haven't even bothered to remove the CF adapter on it :)
I need one for my A2000 willing to sell it?
if this was in the beginning of 2000 it would be aswesome.. now it is too late for the new cpustuff etc.
Well myself and others are having fun with it, not too interested in Emulation or PPC offerings so it's right up my ally.
Would ANYONE accept that people did changes like this on the C64?
sure why not, PIMP it out.
Anyway. it is time to stop.. vampire fanatics will never understand my point, I understand theirs.. but their hope will fail as the team apparntlty does not want the programmers to use their platform not suppling the possability of using their debugtools.
Just need to get new debug tools, software just needs to catch up, which is why this offering is trying to bring in a bigger base of people.
.
so number of software supporing the 080 stuff I guess will be aprox the same as the PPC. (and that is very low. damn I should sell my CSPPC and BPPC :)) while consuming space making you need to softwareemulate other stuff you MIGHT use..
Well legacy software works, it works very well. the inherit nature of 68080 processor design being able to run more instructions in a clock cycle alone boosts current software even if clocked at legacy speeds. As for dedicated software. Optimized libraries could do wonders and it's not as if their team have people that can do that type of work, just they're off doing other aspects of the project. But then I suppose al this will induce another false "Incompatibilities" rant on your behalf.
his project. sure.. yes! true.. it IS! if he want to do bananas shooting out the screen every 10th minutes. he can do that. but if someone thinks that is a bad idea maybe not be angry of it.. especially when asking for cooperation..
He is temperamental for sure at times, but he is faced by a wall of people making negative comments based on falsehoods spread by people being dismissive or purposely attacking the project. I can see how he may be upset from time to time or mistake a legitimate inquiry as another trolling attempt that happens often on IRC.
-
LOL. Majsta wasnt at Amiga32.
@Chucky So yeah talking to Gunnar on IRC, it wasn't Majsta you where talking to. You where complaining about Gunnar directly to his own face...
-
@Chucky So yeah talking to Gunnar on IRC, it wasn't Majsta you where talking to. You where complaining about Gunnar directly to his own face...
:roflmao::roflmao::roflmao:
It would have been worth it to go just to see this.
-
tl;dr
Since everyone here seems to be gentlemen, I'll just come out and say it, myself:
DUDE, SHUT THE FRAK UP ALREADY!
We get it!
You don't like the Vampire nor the '80 softcore.
It's not what *YOU* want.
All I'm seeing you do is b!tch and complain about how "crappy" it is, ad nauseam.
You're as pathetic as McEwen, BBRV, Hyperion, H&P, et al.
I'm *SORRY* that your accelerator got bumped off the Top News on Amiga.org. I'm *SORRY* that the Vampire's FPGA softcore is more popular than your board. I'm *SORRY* that Vampire users "don't get you", when in reality, they just see you as a whiny crybaby, complaining for the sake of trolling.
GET OVER IT!
Let me spell it out for you: There is NOTHING in the world that makes EVERYONE 100% happy.
Nothing.....
You need to spend more time on your own threads, for your OWN board, than derailing another thread, for the sake of being toxic.
Speaking of your board. Let's get a few things out of the way:
ZOMG! You ripped off Commodore's original '40 design! That's theft!
ZOMG! You turned it into an '60 board, which means it's not "pure!"
ZOMG! You added ram to it, which it didn't have before and that means it's not original hardware. It's FAKE!!!!
You see how ridiculous you sound?
*rolls eyes*
You know, Chucky, it's people like you, that made people like me leave the Amiga scene. I was offered a free Vampire, in exchange for testing trackers, by making mods. It never happened, but it wouldn't matter, 'cos my heart isn't in it as I am so SICK and TIRED of the incessant b!tching, moaning, complaining and whining of scenesters that act like the Amiga is the second coming or the frakking Kiwsatz Haderach.
I was one of the ones, in the early 2000s that wanted a switch to x86. I wanted hardware, I could buy off the shelf, on my own and be done with it. I didn't get that and when the first of the over priced PPC boards started coming out, I voiced my opinion, but I wasn't as petulant as you're coming across. I didn't CONSTANTLY gripe and piss and moan about what I didn't get, what I didn't have, what it should have been.
I moved on.
That's something YOU need to do. Move on.
And do us all a favor. Go set up a play date with Atheist already!
That's MY 2p.
I'm going to go back to lurking now. Don't bother responding and to everyone else, especially if we've spoken on a personal level, my apologies for this, but this nutbag needed to be told off.
PS - Where the frak is the moderation around here? Has it REALLY gotten this bad? Geez.
-M
-
I am just wandering what would people say if this offer included exactly the same core we know, but was only advertised as being a fast 68020 feature reduced compatible VampireV2 core? Would any developer be interested in it anyway?
Or is it that Gunnar is so unfriendly/difficult to work with/to deal with, that no one wants to bother with his offer?
I have seen all known Amiga hardware developers reject his offer. So either the core sucks or Gunnar sucks or something else sucks. Which one is it?
-
ok I mixed up people. AGAIN!.. this is my handicap.. so. sure. put me in a odd situation. used to it. move along.. nothing to see here... me and faces %&$#?@!%&$#?@!%&$#?@!%&$#?@!ing never works.. thats it. laugh about it? please do. it is on me. used to it...
Anyway.. interesting how a simple thing as pointing out that things like a MMU that we are used to can fork to this..
and.. I see that many HW developers reject this. and I know things JUST like the MMU is one for the main reasons. and THIS thread is a good reason why. but yeah.. better maybe be quiet.. fool more people into the hype. (if they DO know that they are buying something that can do more new incompabilities. FINE! that's what I knew when I back in the days bought a new CPU card.. I was always early with them all (except 060 due to price))
but now. 20+ years later.. it is about preserving our old memories.
we STILL need that debuggingtools to get developers on the platform. and without the developers, platform is dead like the PPC.
demosceners will have no interest as they like to show off what they can do within the limits there is.. remove them and shadertoy.com is suddenly more intersting.
Anyway.. moderator: please. remove this thread completley. put in the first post.. I will stay out..
Some day. when I get time. I will post my blogpost with more detailed info about pros (YES! there are!) and cons..
but.. repeat Iknow. HW designers declined this generous offer.. think about why...
-
Hi Chucky,
If you refer as "majsta" to "the guy to who you gave ShK's Fusion Forty", it was actually me.
You told me that you will write an article on your blog to explain why you don't like Vampire and after will stop bashing it. I found it pretty smart coming from you because it means you understand what you are doing.
Reading all this, it's pretty obvious you still haven't written it :-)
-
ok sorry for mixing up!. it is me and faces. just never f... works. but still I am doing that same %&$#?@!%&$#?@!%&$#?@!%&$#?@!ing thing EVERY FUC. TIME.. :)
it is halfwritten.
so! serious: please anyone. give me ANY arguement. (in some kind of private way) for pros and cons of what YOU think..
I will do a calm (not forumwise. :) ) argument.
and as I told: I LOVE the hwardware. it is the implementation that I have issues with. and the bad handling of the feedback.
anyway!.. I am out.
-
Chucky - I suggest you wait until the 2.7 core is out and some user feedback happens.
-
good idea.. and maybe prep my vamp with it..
(but if it requires me to register. well this will not happen unless taking a copy from someone)
-
I am just wandering what would people say if this offer included exactly the same core we know, but was only advertised as being a fast 68020 feature reduced compatible VampireV2 core? Would any developer be interested in it anyway?
Or is it that Gunnar is so unfriendly/difficult to work with/to deal with, that no one wants to bother with his offer?
I have seen all known Amiga hardware developers reject his offer. So either the core sucks or Gunnar sucks or something else sucks. Which one is it?
I do not understand what people are thinking... Gunnar and the others in the team have a plan with priorities behind what in which order and how it will be implemented. That would only change if someone offers money or when there are serious developers who help them implementing something f.e. by doing new software showing the new features. When they asked about that a long silence. Now they offer what they have for free because they think it might be interesting outside there is only moaning and people claiming this or that is needed otherwise it is not useful, even though that would mean lots of work and delays for the other projects/developments. It is simply a take or leave it offer, if the few hardware developers think it is not useful they are still free to buy limited original processors. No problem...
-
All of this back and forth about pros and cons seems counterproductive.
I'm in the camp of people who aren't interested in FPGA accelerators until they can do at least what a real m68k can already do, and that means an FPU and an MMU. I haven't added anything to these threads because nobody is talking about a compatible FPU and a compatible MMU in these products, but the naysayers and the apologists going back and forth are just making this all too annoying.
What I would genuinely like to see would be real, technical information about the technical roadblocks for a proper compatible implementation.
As far as I'm concerned, there are so few differences between the m68040 and m68060 that a core could and should be a superset of both. An option can be set at boot time to determine if the exception stack and status register should match one or the other. Options to enable / disable superscalar and branch folding in the PCR and CACR are a superset of the '040, and 64 bit instructions from the '040 can be implemented instead of trapped without breaking anything.
The FPU could easily be a superset of m68881/m68882/m68040/m68060 since anything emulated by the M68060SP can just be made in to directly working instructions.
With regards to an MMU, there are many differences between the m68040 and m68060 MMUs, so one would have to be chosen, probably based on whether there's enough support for the '060 MMU model. The access error stack frame is more detailed in the '060 MMU, the '060 has the PLPA instruction instead of PTEST to translate logical to physical addresses, the cache modes for the TTRs are different, and so on.
I wish I knew more about FPGA development because I think it'd be interesting to try to implement an MMU. It'd certainly be challenging!
-
All of this back and forth about pros and cons seems counterproductive.
I'm in the camp of people who aren't interested in FPGA accelerators until they can do at least what a real m68k can already do, and that means an FPU and an MMU. I haven't added anything to these threads because nobody is talking about a compatible FPU and a compatible MMU in these products, but the naysayers and the apologists going back and forth are just making this all too annoying.
What I would genuinely like to see would be real, technical information about the technical roadblocks for a proper compatible implementation.
As far as I'm concerned, there are so few differences between the m68040 and m68060 that a core could and should be a superset of both. An option can be set at boot time to determine if the exception stack and status register should match one or the other. Options to enable / disable superscalar and branch folding in the PCR and CACR are a superset of the '040, and 64 bit instructions from the '040 can be implemented instead of trapped without breaking anything.
The FPU could easily be a superset of m68881/m68882/m68040/m68060 since anything emulated by the M68060SP can just be made in to directly working instructions.
With regards to an MMU, there are many differences between the m68040 and m68060 MMUs, so one would have to be chosen, probably based on whether there's enough support for the '060 MMU model. The access error stack frame is more detailed in the '060 MMU, the '060 has the PLPA instruction instead of PTEST to translate logical to physical addresses, the cache modes for the TTRs are different, and so on.
I wish I knew more about FPGA development because I think it'd be interesting to try to implement an MMU. It'd certainly be challenging!
they never ruled something out for future but it is a question of priorities. They need a advanced AGA compatible chipset for standalone so this is priority, the same is true for all new features of the V4 generation, f.e. new drivers are needed. FPU will be in hardware where possible or as software emulation depending on FPGA size. All is left for naysayers is compatible MMU. It has no high priority because not much software uses it, only development software. Normal user related software not needs it at all. They do the project for average users, not the few developers who need MMU. As I understand it they also will offer development software using the implemented MMU replacing the old software, perhaps that already ends the discussions. People should simply wait. It is a project by a small group of people doing it in their spare time. They do it because they want to do it. If people want them to do different things they should offer money or give good very good reasons for it. I have not seen anything like that, just the typical whining. If someone bought a vampire in past and is unhappy with it sell it, I am pretty sure that there is no problem to sell it. If these people keeps the hardware than make a proposal what you would like to have, developers will read it and perhaps do it later or not because of good reasons but the agressive bitching only makes bad impressions.
-
I have always wondered why people in Amiga scene want to comment things that they are not interested? Previous example was Amiga OS4, every thread got tens of post from MorphOS people who are not interested Amiga OS4.
Seems that Vampire is now in that position.
I'm not interested Vampire, develobment took wrong course. For me that is end of story. I usually just don't read Vampire threads.
What is wrong with you people????
Only thing wich I have against Vampire team, is that they named their product to Apollo. Now it is very annoying to look anything (real) Apollo accelerator related from internet.
***************************************************************
There should be limit how many times one person can comment to one thread.
***************************************************************
Commenting 100 posts of nonsense is just pointless and ruins interesting news threads.
This particular kind of nonsense really pisses me off.
OK, I have MorphOS systems, I'm buying an X5000 in the future, BUT I own an A2000.
So why don't you censor yourself, or just STFU, instead of critisising me for having an opinion about something I'd like to buy?
The compatibility issues ARE important. And since high end accelerators for the A2000 are not common, this device is about the only option.
On one side you talk about future goals, and then you say we need to judge the device only on its current feature.
You keep telling us how great everything is, and what a wonderful person Gunnar is, then you attack anyone who dares challenge your Rose colored glasses viewpoint.
Apparently none of you have experience in actual commerce, because your lack of social skills would doom you to failure if you tried to use these tactics on an average consumers.
Your eccentric approach is turning people off and your losing potential supporters.
It's immature and offensive.
-
Could be cool if somebody can post a list of incompatible software?
From my experience i don't see any particular issue on the software i have run.
Even old games that if run out of whdload could crash on my old 68060 run nice and dandy on the vamp.
Also you don't even need the setpatches and the 68060 libs that if not installed could crash the os even before getting to the desktop.
That for me look a lot more compatible than a plain 68060.
-
@Chucky:
If you were to implement the integer CPU, an FPU, an MMU, and the chipset, in which order would you tackle those?
I think it is reasonable to do the integer CPU first because else you cannot execute any code. The FPU should be implemented before the MMU because there is far more code using the FPU than the MMU. So the order Integer, FPU, MMU makes a lot of sense. FPU is in the making so MMU will have to wait. This is where we are now and it makes a lot of sense.
What about AGA and SAGA? Obviously there is some confusion about what SAGA is. SAGA is an enhanced AGA with an additional chunky playfield. The chunky playfield is controlled via registers in the Amiga register address range and can be controlled by the copper. Since AmigaOS doesn't know about chunky playfields, there is a P96 driver for the chunky playfield which makes this chipset feature appear as standard RTG even though it can/could do much more (e.g. copper scrolling and copper palette shading).
The most notable advantage of SAGA is that the screen DMA can read directly from the local Vampire RAM which is faster than previous fastmem but functionally chipmem. All previous RTG software was slowed down by outdated hardware buses with something like 16 MB/s maximum throughput. The Vampire does hundreds of MB per second.
So far we have only seen the chunky playfield part of SAGA, i.e. the part that seems to be simple RTG but is implemented in an Amiga-way, in public releases. The Gold3 previews already have the AGA chipset, i.e. the planar part of SAGA added but there are still some bugs. You could call this the AGA in SAGA. Simply put SAGA = AGA + RTG on one and the same screen.
I think it wouldn't make any sense to make an effort to purposely slow down the planar part of SAGA to AGA speed. It is not even required for (my broader understanding of) compatibility. I can understand that democoders want to work around a known set of limitations and that for this reason the Vampire/Apollo project doesn't make sense for them. Well, so be it, it is not as if the demosceners were such a big part of the Amiga usergroup. Their wishes just don't comply with the goals of the project and this is not going to Change no matter how much "input" and "listening" there might be.
Now let's return to the CPU: you are very emotional about AMMX, hyperthreading and 64 bit extensions which you consider crap that is not going to get used and that take FPGA real estate that could be used for stuff that does get used like FPU and MMU. I know that you do not agree to Gunnar's vision but perhaps this explanation will make you see that there is an inner logic to the sequence in which the individual items on the TODO list are tackled.
Gunnar wants to improve the 68k ISA and extend it to what it might look like if Motorola hadn't dropped the ball over 20 years ago. This is what drives him, this is why it is happening at all.
You want something different. Everybody understood that and the project is not going to change its goals for you and those who think alike no matter how much you voice your opinion and personal preferences. You are free to conclude that Apollo/Vampire does not cater your very well defined needs and that's all there really is.
Anyway, if one wants to do a modern extension of the 68k architecture, it immediately follows that the 8 address, 8 data and 8 FPU registers just aren't enough. There is a need for 64 bit datatypes handled in the integer part while the need for 16 bit datatypes is almost zero in today's computing. More registers and especially more FPU registers are a must.
Whether you agree with this vision or not, if you want to end up with such a CPU, you can either build a 32bit CPU, throw it away and start with another one that is 64bit OR you build a 64bit CPU right from the start which means to have those additional registers and to make them wider right from the start. Adding a few AMMX instructions that actually work on those wide registers then is far less work than building and testing an FPU.
And AMMX actually does get used not only in RiVA but more importantly in the P96 driver. So even the totally-anti-AMMX user will have some benefit from the presence of AMMX even if it remains completely hidden from the user.
The same goes for the FPU itself: if you implement a 68k FPU with just 8 registers and everything the way it used to be, then all this work will go to the trashcan if you do an FPU afterwards that can access 24 registers. It is simply easier to first implement the _superset_ and then the _subset_ based on the finished superset.
The AMMX parts and additional registers are the foundation of a house that will have an FPU and MMU as a roof. You simply can't start with the roof and then build the foundation.
Another thing you have addressed multiple times is how femu is not better than the 68060.library approach. This is wrong for a simple reason: femu will be included in the core and thus not visible to outside code, most of all it cannot be overwritten by accident. More importantly femu will be present from power-on and does not need to get loaded from disk or wherever. This means that thanks to femu the FPU looks to any code exactly like an 882 FPU right after power-on/reset.
When you compare the Apollo FPU in its current state to the 040/060 FPU, the Apollo FPU is just as well a software/hardware mix as the 040/060 FPUs are. While the 040/060 FPUs have a fixed distinction between hardware instructions and emulated instructions, the Apollo FPU is flexible and can be handtuned for the FPGA at hand. For a smaller footprint you can implement less instructions in hardware, for a larger FPGA you can implement all instructions in hardware. I think that is actually a very interesting approach.
It is unfortunate that the FPGA in the V2 is too small for the full-featured Apollo Core. However, please understand that this isn't a concern to Gunnar. The Vampire isn't his baby, it is a card that licenses his core and is now too small to hold all of it. He doesn't develop the core for the Vampire, he develops the core as an end in itself.
I hope this clears up a few things...
-
It is unfortunate that the FPGA in the V2 is too small for the full-featured Apollo Core. However, please understand that this isn't a concern to Gunnar.
So i have been ripped off twice again and sits on hardware for a fortune that will not be fully functional for me just like the V600? ;(
I knew i should have stayed away from it when i bought the second card but i was even more stupid when i bought the third one.
-
So i have been ripped off twice again and sits on hardware for a fortune that will not be fully functional for me just like the V600? ;(
I knew i should have stayed away from it when i bought the second card but i was even more stupid when i bought the third one.
I don't think you are entitled to get the full-featured Apollo Core for free just because you bought the V2. You got what you paid for and you even got free updates you didn't pay for. The next update will give you an FPU that is faster than any 882 alongside the integer unit that is faster than any 060. I can't see what's the rip-off here.
-
I don't think you are entitled to get the full-featured Apollo Core for free just because you bought the V2. You got what you paid for and you even got free updates you didn't pay for. The next update will give you an FPU that is faster than any 882 alongside the integer unit that is faster than any 060. I can't see what's the rip-off here.
He moans that his Vampires are not big enough to include all features in hardware
-
So i have been ripped off twice again and sits on hardware for a fortune that will not be fully functional for me just like the V600? ;(
I knew i should have stayed away from it when i bought the second card but i was even more stupid when i bought the third one.
to each his own. if you couldnt wait, its what you are being left with. contrary to you and probably some others customers constantly banging their doors, intead of hoarding stuff and trying to get hold of any new release you might have remained supportive for apollo project from the distance while waiting for it to unfold and ripe to reflect your expectations. i do.
certainly, developing new features and polishing old ones doesnt render delivered product any more unusable than it was at the time of release. i think it was clear up front, that the team will utilize sales to upgrade their proposition eventually. demanding that the upbringing needs to quit at the moment of your aquisition, in order to preserve its value is unrealistic and rather asocial. live with your choices.
-
@Chucky:
If you were to implement the integer CPU, an FPU, an MMU, and the chipset, in which order would you tackle those?
First CPU and MMU - the MMU is the uttermost important, now you can already attract an audience and developers far beyond the Amiga realms.
Then FPU.
Make sure this runs really well, and gets tested and used with a range of operating systems and hardware. Not just Amiga.
Amiga chipset is already implemented several times, so I would initially just use what exists, Minimig AGA. Other platforms use their chipsets, or implement them on FPGA as well.
Then have the "Amiga team" extend and improve AGA core, and the "CPU core" team add extra features to CPU, FPU, MMU, whatever, _in communication and agreement_ with core developers of all the involved 68k platforms.
Well - this is how I would have done it.
-
He moans that his Vampires are not big enough to include all features in hardware
But it is.
-
He moans that his Vampires are not big enough to include all features in hardware
No, i dont care about features i will not use. I moan about the slow development.
I hate it when people beg to open source stuff but here it might be kind of suitable.
Maybe someone with the interest and knowledge to build an 68k cpu could complete it and make the V2 fully functional and gunnar can fullfill his dream and play with this ammx stuff etc. Everyones happy :)
-
The next update will give you an FPU that is faster than any 882 alongside the integer unit that is faster than any 060. I can't see what's the rip-off here.
Let's see what the faq says these days....
Do Vampires have a FPU ?
A: The implemented FPU is a 68060 compatible FPU. FPU is currently disabled and is a WIP.
Various teases and hints have said "HW 040/060 FPU", but I guess this is not for 2.7 then.
-
Let's see what the faq says these days....
Various teases and hints have said "HW 040/060 FPU", but I guess this is not for 2.7 then.
It is for the V4 vampire, which will likely release before Gold 3 hits.
The V2 obviously can't fully implement it.
-
It is for the V4 vampire, which will likely release before Gold 3 hits.
The V2 obviously can't fully implement it.
The FAQ entry was there long before V4 was announced, and it does not say anything about this being V4 specific, it says "Vampires", as in plural, which certainly suggests more than just V4.
-
All the problems with the Apollo Core project have their base in one thing:
BAD COMMUNICATION!!!
And there is no "language barrier" to hide behind.
-
So - V2 - what is the reality?
The teaser videos showing V2 with "hardware 040/060 FPU" running Quake etc, is that not the kind of FPU that supposedly does not fit into the V2 FPGA?
-
So - V2 - what is the reality?
I think it was invented just to annoy you...
The teaser videos showing V2 with "hardware 040/060 FPU" running Quake etc, is that not the kind of FPU that supposedly does not fit into the V2 FPGA?
It doesn't fit, there is a small subset of FPU functions on Hardware the rest is being done by an 68080 optimized version FEMU made by Jari. That of course is the V2. The V4 won't need FEMU as it will have the full fledged FPU in the core.
All the problems with the Apollo Core project have their base in one thing:
BAD COMMUNICATION!!!
And there is no "language barrier" to hide behind.
Of course, they're a small team focused on getting things done and updating pages and wikis takes time out of development work.
-
It doesn't fit, there is a small subset of FPU functions on Hardware the rest is being done by an 68080 optimized version FEMU made by Jari.
So "hardware 040/060 FPU" is not true, and it really relies on FEMU.
FEMU is by default in the provided Kickstart? Can it be disabled and enabled by the user?
What does this mean for AROS and EmuTOS? Would they require their own emulation implementations?
-
First CPU and MMU - the MMU is the uttermost important, now you can already attract an audience and developers far beyond the Amiga realms.
Then FPU.
Make sure this runs really well, and gets tested and used with a range of operating systems and hardware. Not just Amiga.
Amiga chipset is already implemented several times, so I would initially just use what exists, Minimig AGA. Other platforms use their chipsets, or implement them on FPGA as well.
Then have the "Amiga team" extend and improve AGA core, and the "CPU core" team add extra features to CPU, FPU, MMU, whatever, _in communication and agreement_ with core developers of all the involved 68k platforms.
Well - this is how I would have done it.
it might be that MMU attracts users outside amiga community but you miss the point that the apollo/vampire was designed for amiga users (or former amiga users) primarly
-
I have thought about getting a vampire and finally did for the 600, as there is nothing comparable available. For the bigger miggies the 2 and 4000 i think keeping the tekmagic and the cs will be better for my needs. Now, winuae has been emulating fpus without noticeable problems. Including the integrated ones ie 68040 and 68060. Can the routines used for this not be implemented in the vampire core?
I am not an electronic engineer but correct me if I am totally off here
-
Of course, they're a small team focused on getting things done and updating pages and wikis takes time out of development work.
I get the impression that there are several people in the team who are more or less twiddling their thumbs or "fighting for the cause" on various forums, while the experts are busy. It should be a priority to keep the hyperbole down, but I understand it is tough when chief architect is Mr Hyperbole himself, who apparently at whim announce new features on the forum, causing confusion, expectations, hyperbole, eye rolling... in one word - ATTENTION.
-
FEMU is by default in the provided Kickstart? Can it be disabled and enabled by the user?
What does this mean for AROS and EmuTOS? Would they require their own emulation implementations?
You didn't read or understand what I wrote above. Femu will become part of the core and be present from power-on/reset. This is one important advantage over 060.library and does not depend on the OS.
-
it might be that MMU attracts users outside amiga community but you miss the point that the apollo/vampire was designed for amiga users (or former amiga users) primarly
Really?
It took literally YEARS before even the word "Amiga" came up on the apollo-core site.
-
You didn't read or understand what I wrote above. Femu will become part of the core and be present from power-on/reset. This is one important advantage over 060.library and does not depend on the OS.
So it is an emulator, a piece of software running out of the reach for the operating systems, hosted by the Apollo Core.
What other pieces of software are running out of reach for the operating system?
-
I moan about the slow development.
you must be joking. slow development? compared to what? your favourite os4? in my book the development here is furious..
I hate it when people beg to open source stuff but here it might be kind of suitable
yeah. put everything you gathered on forums upside down and there, we have a sensible post. congrats.
-
@kolla
Do you realize of course that the 68000 was “microcoded”, right ? And the OS has no exposure to that. Do you realize that Intel periodically still pushes microcode updates to some of his cpus, and the OS does not directly access it?
I am not siding with the Apollo team, but I frankly miss the Amiga community of old. When we called ICs by name, knew what we were doing, complained zero and produced a lot.
Amiga OS needs apps and users, not useless quarrels
@everyone
Repeat after me: FPGA is as hardware as your Gayle and Gary chips, just not requiring a foundry to overlay connection maps via lithography masks.
And repeat again: either you code your HDL into NAND gates or MUX/LUT blocks, it remains equivalent, except the additional complexity
So it is an emulator, a piece of software running out of the reach for the operating systems, hosted by the Apollo Core.
What other pieces of software are running out of reach for the operating system?
-
Who cares what is hidden in there, if make the CPU faster bring it on.
Look at the intel management engine ?, you are running full blown minix os inside your chip set and nobody even cared or realized it. And this on "real chips" like you guys love.
-
Look at the intel management engine ?, you are running full blown minix os inside your chip set and nobody even cared or realized it. And this on "real chips" like you guys love.
Ah, you followed the links I gave on the Amiga Facegroup page earlier today perhaps? Good :)
I do not love Intel chips, and I do not love proprietary management systems. At all.
-
@Tellurium
Yes, I am aware of that - it is fairly well described and documented, and what isn't, is to be picked apart by various communities. Is that what we want in Amiga land too?
Is it wrong to raise questions about what takes place in the core? All experience says that provoking questioning is the only way to get solid information.
-
Asking and not trusting is not a good pattern. Whatever the Apollo crew would say, could ever change your mind?
If I were in a similar position as you are now, I’d get AsmOne with its interactive mode and type small snippets to Unit test if output corresponds to what I expect. AMMX or whatever they call the extension cannot activate if you don’t ping it, so in case you don’t need it, it would be just like empty space.
It’s a general purpose execution unit in a cheap and dirty low end fpga. I Do not expect it to have an AI aware Bitcoin wallet sniffer within. And if in doubt, i’d Keep the wallet away from an Amiga and/or not buy it.
@Tellurium
Yes, I am aware of that - it is fairly well described and documented, and what isn't, is to be picked apart by various communities. Is that what we want in Amiga land too?
Is it wrong to raise questions about what takes place in the core? All experience says that provoking questioning is the only way to get solid information.
-
Yes, kolla, right, you're not "fighting for the cause".
You've been trolling about Vampire/Apollo since you got some trouble with the guys.
Get over it.
-
Who cares what is hidden in there, if make the CPU faster bring it on.
If it means it can't run MacOS or Nextstep?
If your answer is that has nothing to do with Amiga and you therefore don't care, then you're part of the problem.
Asking and not trusting is not a good pattern. Whatever the Apollo crew would say, could ever change your mind?
If they stopped deliberately misleading people it would be a start. If they actually delivered on 100% compatibility then I'd support them. Nothing gunnar has ever said has indicated he is interested in 100% compatibility. He's interested in his own agenda and needs enough people behind him to get him there. If it does what you want then I guess you'd be happy to be bribed.
I'm sure gunnar knows how controversial his 100% compatibility claims are, it's similar to the tactics used in the UK/US/Europe with the rise of fascism (Trump/brexit/etc).
Look at the intel management engine ?, you are running full blown minix os inside your chip set and nobody even cared or realized it. And this on "real chips" like you guys love.
You're obviously wrong, or the researchers wouldn't have taken the time to investigate it. Of course I didn't realise the secret minix os was running inside my computer, because it was secret. Even the guy who wrote minix didn't know for sure. I couldn't care about it because I didn't know. I did care that IME existed, but without credible evidence that it was exploitable then it's difficult to do anything about it.
People care now.
-
What's with all the merely curious people who've made accounts just to... not answer any questions nor bring any more information to the table?
-
What's with all the merely curious people who've made accounts just to... not answer any questions nor bring any more information to the table?
No doubt they want to 'fight for the cause'. :hammer:
-
Joined here 2011, no idea what you're referring to.
-
Asking and not trusting is not a good pattern.
Actually it is. Do you prefer trusting, no questions asked?
Whatever the Apollo crew would say, could ever change your mind?
Change my mind about what?
If I were in a similar position as you are now, I’d get AsmOne with its interactive mode and type small snippets to Unit test if output corresponds to what I expect. AMMX or whatever they call the extension cannot activate if you don’t ping it, so in case you don’t need it, it would be just like empty space.
So you are telling me to only test the part of the core that is designed to be compatible, and to stay away from the rest?
It’s a general purpose execution unit in a cheap and dirty low end fpga. I Do not expect it to have an AI aware Bitcoin wallet sniffer within.
I think it is rather funny that you bring up Bitcoin in this, considering that crypto currency was such a re-occuring topic on the IRC channel. Anyways, that's totally beside the point. I asked if there are other pieces of software running behind the scenes, that should be easy to answer.
-
I asked if there are other pieces of software running behind the scenes, that should be easy to answer.
Yes, there is WinUAE running inside the core! Now you found our secret! And that we are violating the GPL on top of it all!
-
Yes, there is WinUAE running inside the core! Now you found our secret! And that we are violating the GPL on top of it all!
Wouldn't be any weirder than Intel running Minix3 in its hardware. :lol:
-
Wouldn't be any weirder than Intel running Minix3 in its hardware. :lol:
I thought AROS would be an interesting choice for the IME :)
-
Sanity has left this place since a long time I see.
Been Tellurium since 2009. My previous account was deleted in the great purge of dormant accounts. Not posting due to the state of decay of this once glorious community. I came to buy an Amiga a bit late, in 1992, it was the beginning of the AGA, the beginning of the end already. I have no idea about you, but I am 40 and really too old for this insanity.
These tribal wars you fight make me just as sad as I feel reading the counter-trolling and the other troll wars.
I do not have a vamp and do not need one. But I may test one in the future and if it runs Amiga OS, the way I want it, then it is gonna be fine for me.
I am closing here. I do not care of these childish quarrels in fashion on this forum. I guess I can spend another decade lurking, for what is worth. All this is so terribly sad.
-
@Tellurium,
+1, feel that way often here.
-
@Tellurium
I enjoyed your posts in this thread.
The level of nitpicking, regardless of target, be it AOS 3.x, AOS 4.x,Morphos, AROS, 68k, FPGA, PPC etc is quite selfdestructive.
Its almost like Life of Brian quarrel between The People's Front of Judea vs Judean's Peoples Front.
https://www.youtube.com/watch?v=WboggjN_G-4
-
@Niding
Better moderation would be nice. But not gonna happen.
I too lurk mainly only reading posts specifically by users that are producing sw/hw and offering info on their wares. Everything else I treat as static and filter, which seems to be 95% of the content here any more. Too many years of the same people, same gripes/grudges and a weird need to repeat themselves endlessly. re-re-re-re-referring to posts of parallel arguments years past. Unless you pick a side of the current recycled argument your average posts are completely ignored. Just weird.
Someone offered free use of a 68k fpga core. Use it or don't. 10 pages of wah wah... spare us.
Anyway, to much time on the subject already. Just for a moment I felt like griping too. :)
Plaz
-
As happy user of ppc amiga I see two groups of people here.
Unsatisfied users of 68k who get that they will have better results by simple overclocking 68060 to 100 MHz,
and powerpc haters who still after so many years hope that natami/apollo/vampire will kick ass this bad ppc which they so much hate.
From ppc amiga user point of view use there is nothing wrong in use "virtual machine" ("amiga os in back seat") or 68040.library or FEMU.
It just prove that Amiga NG was RIGTH way to develop new amiga.
But it not fit in natami/apollo/vampire propaganda and 68k users may fill used to support worse (slower) 68k NG.
I also understand that 68k users may not like Intel MMX on Amiga, because they may use MMX on pc.
(Or not because everyone on pc switch from 20 years old MMX to shaders many years ago).
Ours 68k brothers keep the faith!
Look into the bright side of life!
You can always switch from assembler to C and be happy users of Amiga PPC.
-
Well Im a happy Vampire user and I dont hate the PowerPC based Amigas. In fact I think theyre a necessity in keeping the Amiga spirit alive. PowerPC architecture is certainly the likely direction that Commodore would have headed, if they hadnt gone bust, since Motorola had ceased further development of the 68k processors.
How I would have reacted back then to PPC Amigas, I dont know. Being younger and more willing to accept change perhaps I would have embraced them. I hope A-Eon find success as I do Apollo team and other hardware manufacturers.
-
Well Im a happy Vampire user and I dont hate the PowerPC based Amigas. In fact I think theyre a necessity in keeping the Amiga spirit alive. PowerPC architecture is certainly the likely direction that Commodore would have headed, if they hadnt gone bust, since Motorola had ceased further development of the 68k processors.
How I would have reacted back then to PPC Amigas, I dont know. Being younger and more willing to accept change perhaps I would have embraced them. I hope A-Eon find success as I do Apollo team and other hardware manufacturers.
As I don't see too much overlap, I embrace both legacy and NG hardware myself.
No matter what they do with the Apollo core, it isn't going to outperform a PPC, but it could be a very useful replacement for the 68K.
-
I've never heard of any PPC development happening at Commodore. CBM was focused on the Hombre chipset using PA-RISC and an AGA/M68K SoC. The SoC would allow cost reduced low end classic Amigas and could be included with a PA-RISC machine for backwards compatibility. If Commodore had survived these architectures would probably have seen them through to a transition to x86.
-
No matter what they do with the Apollo core, it isn't going to outperform a PPC, but it could be a very useful replacement for the 68K.
At least as an FPGA.
And in time, the PPC will almost certainly end up as a replacement candidate as well.
-
I've never heard of any PPC development happening at Commodore.
That is correct. PowerPC only came to the Amiga after commodore went bankrupt. The first PowerPC chip shipped in an IBM workstation in October 1993, Apple launched the first PowerMac in March 1994, Commodore went bankrupt in April 1994.
Amiga Technologies said they were porting AmigaOS to PowerPC in 1995 along with new machines with Phase 5 providing the cards for existing machines, ESCOM died in 1996 and Phase 5 had to come up with their own software support, which later turned into the PowerUP vs WarpOS war.
WarpOS became standard even though there were really compelling technical reasons why it shouldn't, the same reason Trump gets elected and people love Apollo
CBM was focused on the Hombre chipset using PA-RISC and an AGA/M68K SoC. The SoC would allow cost reduced low end classic Amigas and could be included with a PA-RISC machine for backwards compatibility.
Hombrew was a clean break from the Amiga, it didn't use AGA or 68K at all.
A couple of documents turned up earlier this year "Hombre: Beyond Amiga (October 8th 1993)"
http://eab.abime.net/showthread.php?t=87342
It would be nice to see the HDL for Hombre, but it's likely unfinished and contains copyrighted PA-RISC as well. PlayStation was similarly designed in HDL mostly from an LSI library, then cut down. From what I can tell the PlayStation was a better design, I've always felt it was the spiritual successor to the Amiga.
If Commodore had survived these architectures would probably have seen them through to a transition to x86.
And then they'd have died trying to compete with Dell/HP/Compaq/etc.
It's kinda weird that nintendo, sony & microsoft transitioned to PowerPC years later and then Sony & Microsoft have followed apple to x86 as well.
Even the later slim ps2 had a power pc hidden in them, running a software emulation of the IOP (IO Processor, which is also used for running PS1 games).
-
this thread is everything wrong with the amiga scene.
Someone who hates something that somebody else has done that does't appeal to their purist ideals (despite the fact that the moto 68k cores are very different, all have different fpu capabilities, and different mmu implementations), and they just can't stop pointing it out.
this
-
This thread is everything wrong with the Amiga scene.
Someone who hates something that somebody else has done that does't appeal to their purist ideals (despite the fact that the Moto 68k cores are very different, all have different FPU capabilities, and different MMU implementations), and they just can't stop pointing it out.
That appears to have been missing from the original post.
"We offer hardware developers a
_FREE_ 68080 CORE that is
compatible with the 68040/68060
and is set to about 68040/68060 speed. "
What is wrong with the Amiga scene is that people are still enabling Gunnar despite the lies he has told, it's like you want to be lied to. If the cargo cult stopped, then Gunnar would have to start telling the truth or disappear into obscurity. Maybe people would still buy into it if he told the truth, I guess we'll never know because the enablers won't risk it.
Saying something is compatible means that it's compatible, but it's not compatible. If you have to rely on an alternative definition of compatible then it's a lie.
this
You appear to have missed the point of this thread.
-
That appears to have been missing from the original post.
"We offer hardware developers a
_FREE_ 68080 CORE that is
compatible with the 68040/68060
and is set to about 68040/68060 speed. "
What is wrong with the Amiga scene is that people are still enabling Gunnar despite the lies he has told, it's like you want to be lied to. If the cargo cult stopped, then Gunnar would have to start telling the truth or disappear into obscurity. Maybe people would still buy into it if he told the truth, I guess we'll never know because the enablers won't risk it.
Saying something is compatible means that it's compatible, but it's not compatible. If you have to rely on an alternative definition of compatible then it's a lie.
You appear to have missed the point of this thread.
I think you are playing the semantics game. Did the release notes say 100% compatible with all features available er No. Does the whole world and his dog know that there is not a like for like MMU currently available.
So in your world you are waiting for the mythical phase 5 products to be release and skip merrily to 060 heaven. Back in the real world finding a supply of 060 processors is impossible, it would be easy to find Lord Lucan.
If a professional hardware developer wish to build an 080 accelerator board which gives the 060 a run for its money and create a 68EC060 like board at a good price then this is a good thing.
And for the record a do own a Vampire for my 600 and this is the only reason a got back into using an Amiga after 15 years+.
-
When the trolls are going to say what is the incompatibility ?.
Because they speak and speak about it but never list what software don't run on the vamp.
-
I think you are playing the semantics game. Did the release notes say 100% compatible with all features available er No.
I'm not playing the semantics game, you are playing the semantics game.
If they say compatible with no qualification of compatibility then yes you can safely expect it's 100% compatible.
They know it's not 100% compatible but are prepared to mislead, otherwise why wouldn't they just admit it? Are you saying they are too stupid to know?
Does the whole world and his dog know that there is not a like for like MMU currently available.
I'm not entirely sure what you mean by that, Gunnar would simply just have to write the HDL to create a compatible MMU. He refuses.
So in your world you are waiting for the mythical phase 5 products to be release and skip merrily to 060 heaven. Back in the real world finding a supply of 060 processors is impossible, it would be easy to find Lord Lucan.
I'm waiting for Gunnar to add an 040 or 060 compatible MMU and FPU. Or for someone else to do an equivalent board that supports it. Unfortunately Gunnar is doing his best to stop either of those things happening.
When the trolls are going to say what is the incompatibility ?.
Because they speak and speak about it but never list what software don't run on the vamp.
That isn't true at all, the only trolls are speaking in favour of Apollo.
If you're serious then lets start with an easy one, NetBSD http://wiki.netbsd.org/ports/amiga/
Then how about http://aminet.net/package/util/libs/MMULib
Quite a useful one here http://aminet.net/package/dev/debug/enforcer (Enforcer 37.73, supports 68060). Apparently 68080 is compatible with 68060 and this version of Enforcer supports 68060 so of course it will work (wrong).
I don't know if you really believe what you're saying, or whether you're just trying to be a troll yourself.
-
Enforcer, mungwall...
etc..
-
A whole range of plugins for AdPro, SView, ImageFx etc, various ixemul versions and everything that depend on them, like sox, lame, ffmpeg etc.
-
Btw - anyone who think I waste any hate on something as mundane as this nonsense, is wrong :)
-
I'm waiting for Gunnar to add an 040 or 060 compatible MMU and FPU.
Nah, he cannot do that, can you imagine the outrage?! :laughing:
I am glad that he again is trying to reach 68k communities outside the Amiga realm, since that will challenge the apollo core on the problematic issues.
-
I'm not entirely sure what you mean by that, Gunnar would simply just have to write the HDL to create a compatible MMU. He refuses.
simply write hdl.. know what? he said its down the priority list, to write hardware mmu compatinility wrapper. if you think its immediately important, you can contibute to funding it. other than that i dontt think he owes you some feature or other. you can simply stay away from his offer, rather than trying to scare him away and disapoint the majortity of those, fond of such an exceptional option.
I'm waiting for Gunnar to add an 040 or 060 compatible MMU and FPU. Or for someone else to do an equivalent board that supports it. Unfortunately Gunnar is doing his best to stop either of those things happening.
yeah gunnar tries to prevent implementation of an fpu.. and whats that:
http://apollo-core.com/knowledge.php?b=2¬e=10876&z=VbGYA8
O: ??!
That isn't true at all, the only trolls are speaking in favour of Apollo.
If you're serious then lets start with an easy one, NetBSD http://wiki.netbsd.org/ports/amiga/
Then how about http://aminet.net/package/util/libs/MMULib
Quite a useful one here http://aminet.net/package/dev/debug/enforcer (Enforcer 37.73, supports 68060). Apparently 68080 is compatible with 68060 and this version of Enforcer supports 68060 so of course it will work (wrong).
I don't know if you really believe what you're saying, or whether you're just trying to be a troll yourself.
hard time trying to find an mmu dependant amiga software? i agree mutools would be nice to have, but it starts to be boring that you guys always refer to it and netbsd as support of your rants, coul go find another example to entertain us a bit more;)
-
A whole range of plugins for AdPro, SView, ImageFx etc, various ixemul versions and everything that depend on them, like sox, lame, ffmpeg etc.
i didnt recall ixemul itself depends on fpu, it rather depends on what math lib you are linking your program against, before even linking against ixemul stub.
-
simply write hdl.. know what? he said its down the priority list, to write hardware mmu compatinility wrapper.
Well that is an improvement, he used to go into melt down when anyone suggested they needed a compatible MMU. I guess we're having an effect, good to know.
if you think its immediately important, you can contibute to funding it.
How?
other than that i dontt think he owes you some feature or other.
He owes it to anyone who buys one due to misleading advertising.
you can simply stay away from his offer, rather than trying to scare him away and disapoint the majortity of those, fond of such an exceptional option.
If he stopped lying then I wouldn't have commented. Why don't you convince him to stop lying? It would be much more productive.
yeah gunnar tries to prevent implementation of an fpu.. and whats that:
http://apollo-core.com/knowledge.php?b=2¬e=10876&z=VbGYA8
O: ??!
What I said was he is trying to make sure nobody else produces an alternative, by lying to get people to buy into his product.
The last I heard even the compatible FPU wasn't designed to give the correct results due to difference in precision. If the complaints have made him change his mind then that is a bonus, you might want to check though because he might just not mentioning it anymore.
hard time trying to find an mmu dependant amiga software?
No, I listed a few plus there are more if I wanted to spend time on it & FWIW the press release specifies nothing about amiga software either. It says "compatible with the 68040/68060" so Next Step would be another piece of software.
Maybe he should have said "compatible with software that I decide is worthwhile running".
i agree mutools would be nice to have, but it starts to be boring that you guys always refer to it and netbsd as support of your rants, coul go find another example to entertain us a bit more;)
Right, so you admit that you only do this to troll. Good to know. This is not a rant, it's a response to the whinny apollo trolls rants.
-
I'm not playing the semantics game, you are playing the semantics game.
If they say compatible with no qualification of compatibility then yes you can safely expect it's 100% compatible.
They know it's not 100% compatible but are prepared to mislead, otherwise why wouldn't they just admit it? Are you saying they are too stupid to know?
I'm not entirely sure what you mean by that, Gunnar would simply just have to write the HDL to create a compatible MMU. He refuses.
I'm waiting for Gunnar to add an 040 or 060 compatible MMU and FPU. Or for someone else to do an equivalent board that supports it. Unfortunately Gunnar is doing his best to stop either of those things happening.
That isn't true at all, the only trolls are speaking in favour of Apollo.
If you're serious then lets start with an easy one, NetBSD http://wiki.netbsd.org/ports/amiga/
Then how about http://aminet.net/package/util/libs/MMULib
Quite a useful one here http://aminet.net/package/dev/debug/enforcer (Enforcer 37.73, supports 68060). Apparently 68080 is compatible with 68060 and this version of Enforcer supports 68060 so of course it will work (wrong).
I don't know if you really believe what you're saying, or whether you're just trying to be a troll yourself.
So that's it 3 programs ?
They should change it to 99.5% compatibility?
-
except that you do not count compability in number of programs that work on it.. but functions used and why the program breaks.
and in this case the reason is: MMU is missing (or compatible one) and as they claim 68040/68060 compabilty.. those programs (functions) should work.
-
[SARCASM ON]
Amiga users everywhere lay awake at night wondering how they will get NetBSD running on their systems....all the hand-wringing and pacing the floor.....oh the humanity! And not being able to run NextStep on an Apollo core is sheer blasphemy! Now if the Apollo team refuses to support BeOS, then I'm done! It's back to 1990 for me!
[/SARCASM OFF]
The attacks presented here against Gunnar and the Vampire are so over-the-top-ridiculous that it leads one to believe that those who are launching them suffer from mental illnesses along with a bad case of drama-queen syndrome. To call their constant rants "obsessions" would be an understatement. They act as if they're being forced to buy a Vampire or the Apollo core. What's more likely is that they've been offered a financial incentive to bash the Vampire and the Apollo core.....either that, or they really are just crazy. They certainly are playing the parts of drama queens perfectly.
-
Enforcer, mungwall...
etc..
You've stated this since page 1. How many more pages do you need? I'd love to start deleting posts again :-)
-
You've stated this since page 1. How many more pages do you need? I'd love to start deleting posts again :-)
Well still people asks for examples? isn't it strange? TECHNICALLY I never mentioned the software by name.. only by function.
-
I don't care about MMU tools for AmigaOS, since I am not developing on Amiga. Classic Amiga is useless as developing platform anyways. All development these days is done with cross compilers and debugging is done using WinUAE. Right?
-
well as WinUAE does not have 080 emulation..... :)
but yeah! I do 98% of my coding in uae.. true..:)
-
At least as an FPGA.
And in time, the PPC will almost certainly end up as a replacement candidate as well.
I can't see that being fast enough to interest me, but it could be done. It's already done under UAE.
For now, we have ( or will soon have) the PPC hardware we need, and my primary NG OS is eventually moving to X64, so I won't really need PPC FPGA emulation.
Besides, Gunnar doesn't seem to be an advocate of it.
The argument could be made that an enhanced 68K would prove just as powerful, but I'm not weighing in on that.
I am 100% sure that no solution we come with on our own will even come close to a modern X64 CPU though.
-
Right?
wrong..
-
It just prove that Amiga NG was RIGTH way to develop new amiga.
...and for those Intel haters I think that AmigaOS should run on Intel cpu's and emulate everything else for sake of compatibility.
:-)
-
Amiga users everywhere lay awake at night wondering how they will get NetBSD running on their systems....
Sigh. Well, one of us does...
-
I dunno. AmigaOS x86/64 might mean less expensive hardware. Hell, at this point maybe an up-and-coming processor like ARM might work out. Though that might make Amiga too much like an Acorn.
-
I dunno. AmigaOS x86/64 might mean less expensive hardware. Hell, at this point maybe an up-and-coming processor like ARM might work out. Though that might make Amiga too much like an Acorn.
I think in all of those discussions, people need to ask question what makes an Amiga?
I have to say. If my AmigaOS4.1 would run as good as my WindowsXP on PIII 600Mhz + RivaTNT2. I really would not care what in under the hood.
Considering that basically there is not Amiga "proprietary" CPU/HW since ALL of this HW is available for other systems, really what is such a big deal about making AmigaOS4.1 for Intel based HW? It is not like I'm using Amiga specific LAN, sound card, GPU, ram, HDD etc. etc. even Amiga specific HW such as BVision was developed for PC computers etc. etc.
Can someone give me a reason why we have to have either 060 or PPC for AmigaOS. Some different reason then: Windows will not run on it. This was important 25 years ago...
What is most important/problematic: Waiting for under powered new PPC HW or waiting for AmigaOS that is widely available?
-
@kreciu
I suspect the sheer work of making it x86 based would that many years, so not really realistic, espesially given aeon and acube has sunk alot into PPC.
Maybe Im wrong, but thats the first thing that comes to mind regarding platform move.
-
How's AROS doing these days?
-
How's AROS doing these days?
I was thinking about MorhpOS when I read your post... I think we need more diversity in Amiga environment, that will help a LOT.
-
I am 100% sure that no solution we come with on our own will even come close to a modern X64 CPU though.
In many respects, Apollo extends the life of the Amiga while the PPC merely prolongs its death.
I'm assuming MorphOS is your NG OS of choice. Do we know how far off it is from x64?
-
I think we need more diversity in Amiga environment, that will help a LOT.
I may have missed the obvious sarcasm.
How would increased diversity help? Having multiple half-baked OS options has been a disaster for this market.
-
I was thinking about MorhpOS when I read your post... I think we need more diversity in Amiga environment, that will help a LOT.
hehehe I rather like MorphOS and would not be upset at all to see it running on x86/64 hardware. My thought was that AROS was targeted for the x86 all along so it might be the most likely candidate.
-
What is most important/problematic: Waiting for under powered new PPC HW or waiting for AmigaOS that is widely available?
It's pretty obvious that what makes the most sense has never been the guiding force behind the Amiga in the Amino era.
It's mind boggling to me that transitioning away from 68k became an exercise in tightly-coupling the OS with yet another CPU suffering from accelerated decrepitude.
Fifteen years ago we knew that going platform agnostic was the way to go - even if that meant favouring the PPC.
I suspect the sheer work of making it x86 based would that many years, so not really realistic, espesially given aeon and acube has sunk alot into PPC.
There is a compelling case for rewriting the OS rather than porting it and maintaining backward compatibility through emulation. See also: Apple.
The laughable thing about all of this is that 30 years ago a small group pumped out the first few generations of the software in far less time than it has taken to move the needle forward in the next 25. We're talking about transitioning or replacing something that came out decades ago, with a small footprint and limited complexity.
-
hehehe I rather like MorphOS and would not be upset at all to see it running on x86/64 hardware. My thought was that AROS was targeted for the x86 all along so it might be the most likely candidate.
Only problem with AROS is that it doesn't bring this emotional "load" as AmigaOS (same with MorphOS).
-
It's mind boggling to me that transitioning away from 68k became an exercise in tightly-coupling the OS with yet another CPU suffering from accelerated decrepitude.
Fifteen years ago we knew that going platform agnostic was the way to go - even if that meant favouring the PPC.
I believe a TON of work has been done on OS4 to do exactly that - limiting the areas of tight coupling to the CPU/architecture. It is something reasonably obvious to do. For example, a lot of old assembler code was replaced by C. That decouples from CPU. I expect if you compared a lot of OS4 code to OS3 code one would see significant cleanup all over.
However, I suspect the endianness is still expected to be the same.
-
Only problem with AROS is that it doesn't bring this emotional "load" as AmigaOS (same with MorphOS).
well, when it runs on your amiga your usual amiga software it actually feels like an amiga. okay, stupid theme and an ugly default icon set aside, which both can be replaced. on the other hand default zune settings look usually better than mui and still are configurable without stealing a keyfile you may not be able to register anymore.
-
Only problem with AROS is that it doesn't bring this emotional "load" as AmigaOS (same with MorphOS).
I don't see why not, you can even run AROS on 68k. There were a few mistakes they made along the way because they had all that x86 cpu grunt, plus a few other things that need optimising.
The main problem with AROS is that it's too similar to AmigaOS, so multicore & memory protection is an issue.
-
wrong..
Good, I am glad you think so.
-
Good, I am glad you think so.
im doing it. the development may find place in cross compiler environment, as well as inititial testing. but the hard part in many cases especially for low level stuff needs to be confirmed on an actual sytem via serial debug. unfortunatelly it confirms that the emulation is these days still not enough for everything.
-
In many respects, Apollo extends the life of the Amiga while the PPC merely prolongs its death.
I'm assuming MorphOS is your NG OS of choice. Do we know how far off it is from x64?
Interesting thought. I'd always kind of considered legacy to be the 'Living Dead' of the Amiga world.
(https://www.emojirequest.com/img-facebook/ZombieEmoji.jpg)
So Apollo, to me, would just be a fast zombie.
NG isn't wed to an ISA or a specific platform like legacy is, so altering it doesn't involve the massive re-writes necessary to rework our original platform.
-
I am 100% sure that no solution we come with on our own will even come close to a modern X64 CPU though.
An atom on an A1200 accelerator card booting a 68k emulator would be very interesting.
-
The main problem with AROS is that it's too similar to AmigaOS, so multicore & memory protection is an issue.
yes, on amiga-m68k target smp, while possible ,if only there was hardware for it, would cause incompatibility. not so on x64.;)
-
Only problem with AROS is that it doesn't bring this emotional "load" as AmigaOS (same with MorphOS).
Thats because it runs on x86, more than enough CPU to handle the emotional load. LOL
-
Gunnar made a general comment about the direction and intent of the Apollo core, which will adress relatively frequently asked questions;
Is was always clear that: APOLLO's roadmap includes to run everything any 68K can run. This means also including running Linux.
But we have a clear policy about future features:
We _NEVER_ promise the users features that are not available today but only on our roadmap.
Users always only guarantee stuff working today 100%.
This means even if we might even run internally Linux for testing.
As long we are not sure we can release it now.
Our official answer will always be "no you can not run Linux".
http://www.apollo-core.com/knowledge.php?b=2¬e=10876&x=1&z=Jg0cBR
-
It's a weird week. Trump says he's only ever tried to be friends with Kim Jong Un and Gunnar says it's clear that Apollo will run standard Amiga Linux builds (unless he's talking about an Apollo only Linux build that is only compatible with user land 68k Linux software, but that would be misleading and dishonest....).
In september Gunnar posted this:
"68080 follows this MOTOROLA tradition.
68080 has an MMU and its updated to be modern."
Which makes his comment about Linux compatibility less clear.
I guess I better wait until the V4 is released and see whether the MMU and FPU is compatible. Going back to the point of this thread, is it the "V2" or "V4" core that is being offered?
-
Maybe one day it will run Apollo/DomainOS :p
-
Going back to the point of this thread, is it the "V2" or "V4" core that is being offered?
There is only one core which is adapted for different hardwares, namely the V2 and the V4. What features can and will be included into a core for a new hardware will depend on that hardware. If it has a relatively small FPGA, the FPU would likely have more microcoded (i.e. transparently software-emulated) instructions than if a larger FPGA was used.
-
And Gunnar has had 2 cups of coffe today it seems. He decided to go into more details about the MMU mechanics of legacy and Apollo;
http://www.apollo-core.com/knowledge.php?b=4¬e=11026&z=U8rPdR
The MMU topic seems to be confusing.
Lets us try to explain it.
An MMU is a piece of logic which allows to do three things.
a) It allows to map blocks of memory to another address.
A typical used block size is 4 KB.
Modern MMU also support bigger blocks e.g 64KB or 1MB.
Support for bigger blocks gives for some applications advantages.
b) the MMU allows to control cache setting for blocks.
c) The MMu allows to write/protect memory
APOLLO 68080 has a build in MMU.
Where is the difference between APOLLO and previous 68K CPUs?
Previous 68K CPUs did NOT come with a memory controller inside the CPU.
This means the memory access logic, the memory layout and control of this was always done OUTSIDE of the CPU.
This means while the CPU could inside change their logical view of the addresses with the MMU - it did not have any understanding of how this is mapped outside - as this understanding was controlled by the mainboard chips of the system.
Also the other way around - if the mainboard chips did do some DMA access - they could not know how these areas are mapped inside the CPU MMU.
This old concept did create many problems.
You all have seen those on the AMIGA.
DMA channels on the AMIGA did always work on physical address.
While programs would work on virtual addresses.
This mismatch was not solvable on AMIGA.
If an application was requesting DMA and was using its virtual view of addresses this was leading to crashes.
APOLLO fixes this problem.
APOLLO has the SAGA DMA channels inside the CPU core.
This means both DMA and CPU can use the same representation of the world.
This does solve a number of old problems.
The CPU does see automatically if DMA channels update memory which is cached by the CPU caches.
This means the CPU has always a coherent view of the memory.
This solves many old DMA issues and also allow the usage of APOLLO with many CPU cores. All CPU cores and the DMA engines share one coherent view of the world.
Another big difference to old 68K CPUs is that old 68K CPUs
did only support 1 bus.
APOLLO is designed as a system supporting several memory controllers.
Old MMUs did had a linear address view in 1 dimension.
APOLLO has a multidimensional view.
All these concepts which APOLLO offers are new to the 68K world.
These concepts do exist already today in other enterprise systems.
Coherent multi core systems supporting several memory channels or memory hierarchies are state of the art today.
Apollo is from ground up designed to support this too.
AMIGA OS is not designed for this today.
AMIGA OS can not easily use SMP.
AMIGA OS can not control NUMA memory views.
To provide a old fashioned 1 dimensional memory view to old AMIGA MMU applications - a clean solution would be to provide an sub MMU with 1 dimension inside the multidimensional MMU view.
This feature would transparently support old applications.
This can be done but is _NOT_ in planned for any releases coming soon.
-
And Gunnar has had 2 cups of coffe today it seems. He decided to go into more details about the MMU mechanics of legacy and Apollo
Essentially, he still doesn't want to do it.
"To provide a old fashioned 1 dimensional memory view to old AMIGA MMU applications - a clean solution would be to provide an sub MMU with 1 dimension inside the multidimensional MMU view.
This feature would transparently support old applications.
This can be done but is _NOT_ in planned for any releases coming soon."
The rest seems like a sales pitch & I don't think I am the target audience of that pitch.
If it has a relatively small FPGA, the FPU would likely have more microcoded (i.e. transparently software-emulated) instructions than if a larger FPGA was used.
Are you referring to FEMU as micro code? Because that is misleading, I assume it's because the original 68882 was microcoded (as was the 68000) and it makes it an easier sell.
How would that work if you wanted to put the FPGA on a chip carrier with 68040 compatible bus for inserting into a Power Mac or Next Cube?
-
And Gunnar has had 2 cups of coffe today it seems. He decided to go into more details about the MMU mechanics of legacy and Apollo;
http://www.apollo-core.com/knowledge.php?b=4¬e=11026&z=U8rPdR
Unfortunately, again misconceptions here. Of course AmigaOs has a mechanism to handle the differences between logical and physical addresses for DMA devices. This mechanism consists of two exec functions, namely CachePreDMA() and CachePostDMA(). And they are even used by the DMA-capable devices CBM manufactured. Not by some third-party devices drivers whose authors "believed to know better", and they are patched away to almost-nothing by some other group of users that "believe to know better what they are good for".
*Sigh*.
-
Unfortunately, again misconceptions here. Of course AmigaOs has a mechanism to handle the differences between logical and physical addresses for DMA devices. This mechanism consists of two exec functions, namely CachePreDMA() and CachePostDMA(). And they are even used by the DMA-capable devices CBM manufactured. Not by some third-party devices drivers whose authors "believed to know better", and they are patched away to almost-nothing by some other group of users that "believe to know better what they are good for".
*Sigh*.
Gunnar answers your post on the bottom of this thread.
http://www.apollo-core.com/knowledge.php?b=4¬e=11026