Amiga.org
Amiga News and Community Announcements => Amiga News and Community Announcements => Amiga Software News => Topic started by: Norway on December 14, 2005, 01:51:23 AM
-
This new article on the offical AmigaOS 4 website, explains the classic memory alloction system and how the new slab allocation (object cache) system in impoves memory allocation.
"Back in the old days of the original AmigaOS, the system used to allocate areas of unused memory to new tasks was pretty simple. The size and position of blocks of free memory were kept in a list, and when memory was needed the system would traverse this list until it found a block that was sufficiently large. This block was then split, one part returned by the allocator for use, while the other part (whatever part of the block wasn't needed) was left in the free list. This method served its purpose well enough at the time, but with the increased demands of modern computing - and of course our desire to bring this new version of the Operating System to the cutting edge - AmigaOS4.0 has introduced a better way of doing things."
Read the rest here (http://os4.hyperion-entertainment.biz/index.php?option=content&task=view&id=22&Itemid=).
-
It's a little technically dense for me, but it seems appropriate for the primary target audience of the website - OEM customers.
Nice to see an update.
-
However, it's perfect for the geeks, though basically nothing but a pre-taste. A 10-page article would have been better ;-)
-
Ah, better? Doesn't look like Tigger is too impressed (http://flyingmice.com/cgi-bin/squidcgi/mbmessage.pl/amiga/124414.shtml).
Dammy
-
Well, it seems he's writing about the solution itself.
I was writing about the length of the article. Whether or not the approach is good is another topic. I merely came with a statement, that a longer article would be better.
EDIT: Better here is in the meaning: Better because it would reveal more of the approach taken. I'd like to get much more detailed information.
-
Ah, better? Doesn't look like Tigger is too impressed.
Kenny and Joanna too. There's a shock.
Perhaps they should rename it GroundhogBunny - same shyte over and over and over and over and over....
-
Back in the old days of the original AmigaOS, the system used to allocate areas of unused memory to new tasks was pretty simple.
This isn't even good english. The sentence doesn't make sense? "was pretty simple"? WHAT was pretty simple?
Shouldn't it be something like:
"the WAY THE / SOLUTION TO HOW THE / WHATEVER THE LIKE system used to allocate areas of unused memory to new tasks was pretty simple.
I'm no professional in english, apparently, but that just looked plain ugly.
-
looks alright to me.
'used to' as in 'utilised to' as opposed to
'used to' as in 'what it did before'.
ie:
the system used was pretty simple.
-
D'oh ! Dratted IE5 being as slow as molasses caused me to DP...
...double post that is !
-
I'm no professional in english, apparently,
Apparently not.
-
For the technoids who would like that 10+ page explanation, read here.... Slab Allocator - Sun Systems (http://citeseer.ist.psu.edu/rd/10936738%2C198499%2C1%2C0.25%2CDownload/http://citeseer.ist.psu.edu/cache/papers/cs/8756/http:zSzzSzluthien.nuclecu.unam.mxzSz%7EmiguelzSzbonwick.pdf/bonwick94slab.pdf)
This isn't even good english. The sentence doesn't make sense? "was pretty simple"? WHAT was pretty simple?
It may seem a bit of a "run-on", but legal. Dissecting you find that "the system" was pretty simple. I can see how it could be a tough read though.
Plaz
-
So what on earth is the magical MorphOS mechanism that beats out slab allocation?
-
@ dylansmrjo
See the links at the bottom of the third page. They describe the system in much more details.
-
@Floid
Dunno, but what has that to do with the topic at hand ? (hint sofar noone had even hinted MorphOS).
@uncharted
Tigger and Joanna do work in the embeeded market (and as far as I know both also use PPCs), so it might be that their constant negative attitude towards OS4 in embeeded markets is just based on better knowledge than on a Hyperion-death-wish.
Those 2 bring forward (from time to time) good arguments why OS4 is ill-suited for THEIR type of products and "embeeded" in general, but I have yet to see a deteiled explaination why anybody should choose OS4 over OSes allready in those markets.
And just to make it clear, I haven't looked into the article, I don't plan to do, I'm just sick and tired of the sameold bringing absurd conspiracy-theories whenever someone dare to even slightly question the holy-triumvirat.
-
Hm. Okay, the way it was written made me misinterpret the sentence.
"the system". So they refer to "the system used" as simple. But then I would have written it like "the system used, to allocate areas of unused memory to new tasks, was pretty simple.". Without using "," before and after "to allocate areas of unused memory to new tasks" I don't really find it an easy read.
So... "was pretty simple" is now what determines how "the system used to" should be interpreted. Since it on the first read could easly be interpreted as someting like "(once upon a time) the system did allocate memory", until reaching the end of the sentince, instead of "the system (that was) used to" I would still say that it is a rather vague sentence. I think at least SOME of my past english teachers would agree with me... :lol:
-
Ah, better? Doesn't look like Tigger is too impressed.
--------------------------------------------------------------------------------
Kenny and Joanna too. There's a shock.
Perhaps they should rename it GroundhogBunny - same shyte over and over and over and over and over....
Gee Uncharted can you explain to me why you are impressed by OS 4 using slab, and please inform us all about the first time you knew you were using a slab memory system, and why you think its a great and innovative system for use in OS4.
-Tig
-
So what on earth is the magical MorphOS mechanism that beats out slab allocation?
I actually don't know (or even care) how MorphOS does it, probably still like the Amiga used to do it, though its possible they may have some Slab like implementation. My issue is this, when I started using computers, at which point most of the guys from Hyperion were breast feeding or not yet born, computers didnt have moniters, you typed a line it showed up on the piece of paper if you were lucky, we also had cards, but thats another story. Moniters for terminals were a big technical innovation, they came after the theory of slab programming, which is used in XP, Linux, (huge list of OSs deleted). Its not innovative for an OS to take an idea older then the amiga computer and which is used by everyone else as our memory management system, its ok to do that, but dont write an article about the innovation of the memory system, when its just like everyone elses. Any real technical person is going to look at that article, laugh at it (as I did) and say why should I use this OS, especially as poorly described as the system is. Someone said the site is for OEMs to look at, I hope not, no OEM is going to look at those first two articles (Ram Disk and then new Slab article) and think that they want to use this OS on there hardware, and since it doesnt run on any manufactured hardware at all now, it doesnt really have a market anyways.
-Tig
-
Yeah, I sort of figured I could actually click on links :crazy:
Very interesting reading. I was a bit surprised to see it was based on a technology used (and created) by Sun and used in Linux as well. We are talking major OS'es here - especially on servers.
Perhaps Amiga OS4 developers suffers from grandeur :lol:
No just kidding. I like it. I really like this.
-
Well, you can always get an microA1 - they are actually available and ships with a pre-release of AOS4.
I've notice you like to flame AOS4, and go ahead. If that makes you life easier, then be my guest.
The rest of us will be pleased to see AOS4 using techniques from the large OS'es rather than an outdated primitive technology which was known to be bad at the time it was implemented.
It's not innovative, but that's not the point. What's interesting is that we get some description of the system, though in a very light edition.
-
and please inform us all about the first time you knew you were using a slab memory system, and why you think its a great and innovative system for use in OS4.
Oh man, I hate that. A pop quiz.
I guess it's inovative that OS4 will include it when it hasn't before? But since (as you mention) it's a model that's being used in other OS's, it's not really new. I also notice from my reading that red hat implements slab in concert with MMU and a zoned buddy allocator. Since OS4 doesn't have these extra components at it's disposal, could it be called innovative when they get it to work without an MMU? Go easy please. I'm no expert, just musing through the data and try to make some sence.
Plaz
-
Well, you can always get an microA1 - they are actually available and ships with a pre-release of AOS4.
Actually there not, and havent been for months, but hey try again. If you want to argue this point, please inform me the name of a dealer where I can buy an Amigaone today and get it fedexed to me. AmigaOnes are nat available now, and won't be available in the future. This is the real problem with OS4, the other one is that HYPErion is trying to live up to its name even though they dont have anything they can sell as the moment.
I've notice you like to flame AOS4, and go ahead. If that makes you life easier, then be my guest.
I flame idiots, you are getting close to playing in fire now, Hyperion is taking a very old Memory management idea and trying to jam it into the Amiga system, its likely to be slower, and not offer alot of benefit. I and others have doubts whether they have actually implemented it in this manner now, and fear this system will break lots of current software.
The rest of us will be pleased to see AOS4 using techniques from the large OS'es rather than an outdated primitive technology which was known to be bad at the time it was implemented.
Why exactly do you believe that the current system is either primative or bad?? Its worked for a long time, and given the way the Amiga works, it definitely had advantages over what Hyperion claims they are doing for OS4.
-Tig
-
@Plaz
could it be called innovative when they get it to work without an MMU?
Are you saying OS4 doesn't use the MMU? Why not?
-
Gee Uncharted can you explain to me why you are impressed by OS 4 using slab,
I never said I was impressed, in fact I never commented on the system itself.
and please inform us all about the first time you knew you were using a slab memory system,
Who says I am using a slab memory system?
and why you think its a great and innovative system for use in OS4.
I never said I thought that.
You must be confused, I was commenting on my lack of surprise at the responses posted by you and others.
-
I find myself reading those OS4 pages over and over :)
I thought it was well written and explained. I got a lil confused with the byte shifting for CPU cache part but appart fromthat, i found it very imformative. :))
-
If you want to argue this point, please inform me the name of a dealer where I can buy an Amigaone today and get it fedexed to me.
Well, I know of at least 3 dealers in Scandinavia and Northern Europe who are capable of shipping mikroA1. There are however no AmigaOne's but there are mikroA1's to get (at a rather high price due to the unfortunate low production).
I don't know about the situation in USA and I really don't care. Whatever goes on overthere makes no impact on my life.
I flame idiots, you are getting close to playing in fire now, Hyperion is taking a very old Memory management idea and trying to jam it into the Amiga system, its likely to be slower, and not offer alot of benefit. I and others have doubts whether they have actually implemented it in this manner now, and fear this system will break lots of current software.
Well, then you ought to start flaming yourself. The management system is a tad old, but it's fine in many OS'es, and there is no reason it shouldn't work fine in AOS. It is way better than the old system, which resembles the memory management in Win9X way too much.
The old way was wrong even in 1985. It works but so does a Trabant. It doesn't make it a good car, though. Nor is the old system good in any way.
The amount of software for OS4 is minimal and any broken software can easily be repaired, unless it is coded in crappy way. However, nothing seems to indicate it would break anything. You're just trolling as usual.
Why exactly do you believe that the current system is either primative or bad?? Its worked for a long time, and given the way the Amiga works, it definitely had advantages over what Hyperion claims they are doing for OS4.
The old system is bad because it wastes too much memory and is inflexible. As I wrote in the former paragraph it looks a lot like other outdated and long time since replaced systems.
Don't look at AOS4 with your eyes on a 20 year old Amiga. Look at it with your eyes on more modern hardware (though the mikroA1 is using some pretty dated technology - at least the Peg2 is using DDR-RAM).
-
Uncharted: Ah.. this thread explains all those funny messages that suddenly appeared on MB... I was wonderhing how those people were so suddenly become interested on that site. :-P
Yeah.. I don't think they (Hyperion) should be hunting new Buzz.Phrases for their OS.. They are allready way too late on fixing critical bugs and delivering the OS, so trying to implement this kind of 'high-tech' stuff only delays it further.
And please do note... word high-tech is on quotes in purpose.. This Slab-allocator (even though itself has certain adwantages over The Amiga way) will not be any miracle cure for underlying problems of their OS (and system in whole), it's just something Flashy they found up on net and decided that it would look cool enough to introduce on website in purpose on buying some time, before the few remaining user's loses their interest.
-
@Tigger
My issue is this, when I started using computers, at which point most of the guys from Hyperion were breast feeding or not yet born, computers didnt have moniters, you typed a line it showed up on the piece of paper if you were lucky, we also had cards, but thats another story. Moniters for terminals were a big technical innovation, they came after the theory of slab programming, which is used in XP, Linux, (huge list of OSs deleted). Its not innovative for an OS to take an idea older then the amiga computer and which is used by everyone else as our memory management system, its ok to do that, but dont write an article about the innovation of the memory system, when its just like everyone elses.
Maybe you should check out what is written and "being presented" before you make an ass of yourself.
The VERY first refence link "[1] The Slab Allocator: An Object-Caching Kernel Memory Allocator (1994)" shows the date "1994". If you looked at that you see in the first section for "Citations" is a link to "The Art of Computer Programming (context) - Knuth - 1968" Yeah that's right "1968"!!
I mean clearly they don't want you to know this has been around for a long time, clearly they want you to think they invented it, it's plainly obvious this is just marketing and they are trying to hide the truth. CONSPIRACY!! I say.
AmigaGuy :roll:
-
@JoannaK
before the few remaining user's loses their interest.
Get a grip. We haven't EVER had memory proctection. We haven't had it for 20 years!! If you're still an "Amiga" user it's probably not, after all these years, going to make any difference if they implement it now, in the future, or ever. I highly doubt this is to "keep the few remaining users", that's as absurd as it gets around here.
AmigaGuy :lol:
-
Amigaguy: Well. I was not thinking about memory protection(*) while I was writing that comment... There are a lot more pressing and critical problems they have not been able to solve in last three or so years. Memory protection won't be the key issue on determing the OS4 future..
(*) It's an real *old* (we thought that a lot in those days as first Turbo cards with MMU:s appeared for A2000, so it must be at least 10 years old debate) and well known issue that one can't make decent memory protection without breaking Amiga Message passing system entirely (or slowing it down to grinding halt). So I didn't even expect them to be able solve that part of the puzzle on OS4.. :-)
-
@JoannaK
Why does OS4 even matter to you. If it's so bad why even bother to comment?
-
@JoannaK
Why does OS4 even matter to you. If it's so bad why even bother to comment?
Because this is a discussion forum, where people discuss things?
You don't have to like something to comment on it.
Why not discuss the article - offer your own views on her observations, and explain why you disagree?
Or you could just continue telling anyone you don't agree with to shut up.
-
@Lando
To quote the preacher in From Dusk Till Dawn
I didn't make a statement. I asked a question.
-
Are you saying OS4 doesn't use the MMU? Why not?
No, I didn't. I just asked an obscure (and facecious) newbie question about using slab without MMU since the Hyperion article didn't mention the MMU as part of their article. I know MMU has been around for ages, but I don't pretend to know much about it's "theory of operation". I do need to be careful about my dry humor.
Plaz
-
Maybe you should check out what is written and "being presented" before you make an ass of yourself.
Maybe you should. First of all why is this a ground breaking system like hyperion claims if its from 1968. Secondly, don't you think that the fact the example they give breaks worse on a slab system then on the current amiga system makes it a bad example to use on there website. Don't believe me, work it out yourself. The example that Hyperion uses for why they are using a slab system, works worse on a slab system then on the the current implementation.
-Tig
-
Well, I know of at least 3 dealers in Scandinavia and Northern Europe who are capable of shipping mikroA1.
Which part of give me a name was confusing to you??
The management system is a tad old, but it's fine in many OS'es, and there is no reason it shouldn't work fine in AOS. It is way better than the old system, which resembles the memory management in Win9X way too much.
The current system isnt like win9x, and there are several reasons the slab system will preform slower on the new system. Is slower better to you??
The old system is bad because it wastes too much memory and is inflexible. As I wrote in the former paragraph it looks a lot like other outdated and long time since replaced systems.
Again, you understand the example they give actually works better with the old system, rather then with the new system right??? I mean super OS genius such as yourself caught on to that right away, right?? But you still think the 3 pages on the website is ok, I mean posting an example that shows you dont know how slab works is good for a website preaching you use slab??
-Tig
-
First of all why is this a ground breaking system like hyperion claims if its from 1968
Again, did you READ anything?? Could you point me to the phrase where it says "Ground Breaking" because I think you're full of {bleep}.
I can't see anywhere in the article that they say they invented this technology or that it's ground breaking. Just more crap from the usual suspects.
What a joke.
AmigaGuy
-
Again, you understand the example they give actually works better with the old system, rather then with the new system right???
One has to admire such bloody-mindedness as yours. Okay, so the extreme example chosen to illustrate fragmentation doesn't show the advantages of slab allocators and objects caches, but that's no reason to dismiss the concepts.
But let's go with bloody-mindedness for now.
The pattern of memory allocation and de-allocations as chosen in that example wouldn't happen in the real world. But let's assume they did. You say the original AmigaOS allocation scheme would work better. How so? Let's assume that the allocations are done with AllocMem and that AllocMem and friends offer the same alignment on OS4.0 that they did on earlier versions of AmigaOS. Then the old system and the slab allocator would perform roughly equally. And neither system would fail allocating the 2-byte blocks, since all allocations are done in multiples of 4 bytes anyway. I suppose you mean the new system performs worse since the slab system has a greater storage overhead involved somewhere. However, since the implementation details haven't been made available, we don't know what this overhead might be. Remember, an old-style free list containing every other 4-byte block is already considerable overhead.
Now let's consider allocations done with AllocVec. With the object-caching system proposed, it's no longer necessary to store the length of an allocation before each memory block. For small allocations such as these, the length of the block can be a property of an object cache. Surely now, even considering such a pathological case as the example you're making such as fuss about, the new system potentially performs much better since it has much less overhead per allocation.
In the real world, the memory system in OS4.0 provides 8-byte alignment for the PPC's FPU (Or is it 16 bytes, as required by AltiVec? I'm not clear on this). Assuming 8-byte alignment, then old and new are roughly equally again in terms of memory usage. But that's a function of the new alignment of allocations - not the fact that OS4.0 uses a slab allocator and object-caching.
In the real world, the simple fact is that the slab system is in general much faster than the old system because there's no need to walk the (potentially arbitrarily long) free memory list every time you want to perform an allocation.
Cheers,
Rich
-
:pint:
Cheers folks.
It's been a while since I enjoyed reading an *amiga* related debate on this site.
-
I can't see anywhere in the article that they say they invented this technology or that it's ground breaking. Just more crap from the usual suspects.
From the article:
"our desire to bring this new version of the Operating System to the cutting edge - AmigaOS4.0 has introduced a better way of doing things"
And this
"This is just one more example of the technical innovations we are bringing to AmigaOS4.0 to make the greatest Operating System of the past into the greatest Operating System of the future."
Again really think these 1968 concepts are cutting edge and technical innovations??? And apparently you are the one who didnt read the article.
-Tig
-
One has to admire such bloody-mindedness as yours. Okay, so the extreme example chosen to illustrate fragmentation doesn't show the advantages of slab allocators and objects caches, but that's no reason to dismiss the concepts.
You are missing my point. First of all this an old idea, its not innovative, etc as Hyperion says on there website. Secondly, there example shows they basically do not understand slab, I do, I've used it on more then one occasion, thats not the issue, look at my first post on moobunny. I didnt know whether to laugh or cry, because here is the OS company, selling there OS by showing they dont know why you would use slab over the current implementation. And because slab has more overhead then the current implementation, its really worse in the case they gave. If you were an OEM looking for an OS, (say like I have done on several occasions) would this make you not consider OS4?? Absolutely, if the company selling an OS, doesnt understand the concepts they are selling, I am surely not going to trust there OS on my product. Slab is used by many OS's, but for the most part they were all designed to use it from the ground up, I have a feeling that between the JIT and this memory change, alot of old stuff is going down, I'm talking 1.4 level of breakage, which ended up with us going 2.0 instead, which was less groundbreaking as an OS, but more compatible with the current software. I really dont imagine Hyperion is going to put anything like clickity-split into the OS, do you??
-Bill
-
Again really think these 1968 concepts are cutting edge and technical innovations??? And apparently you are the one who didnt read the article. -Tig
Okay, so you are say that Amiga OS 4.0 is the only current OS that is using this nearly 40 year old idea?
Well, for Amiga OS it is inovative. Of course by your argument Linux, BSD and UNIX are in the same boat of old tech. Oh, Guess that would incluse MacOS too. As it is based on BSD now which was based on UNIX. Which we all know is ancient.
So, that would make Window one of the only true modern cutting edge inovative OSs! :-D
Was the Theory of Relativity inovative? It was based on over 50 years of other peoples idea, theories, and experiments.
Now you are sounding like a broken record, skipping and repeating. I think you've made your statement and opinion clear. Time to move on to something else. So some people don't argee. Repeating yourself isn't going to work.
-
So you're telling me you're wrong and they didn't say it was "ground breaking"? "Ground breaking" meaning something that's never been done before??
"our desire to bring this new version of the Operating System to the cutting edge - AmigaOS4.0 has introduced a better way of doing things"
"A better way of doing things" - Isn't it better to have memory protection than not? "Their desire to bring AmigaOS to the the cutting edge" - Adding memory protection is helping to get that desire fullfilled. One step at a time...
"This is just one more example of the technical innovations we are bringing to AmigaOS4.0 to make the greatest Operating System of the past into the greatest Operating System of the future."
So you're saying SLAB wasn't a technical innovation? Strange I wonder why it got so widely adopted. Hmmm... I suppose the internet wasn't a technical innovation since it was around in the 60's and if someone made a device that allowed you to get on the internet on your "cell" phone, now that wouldn't be an innovation right, because the internet has been around since the 60's. :roll:
Still missing their 'ground breaking' claims or that they have invented something.
AmigaGuy
-
Dear Tigger,
The Knuth68 reference is just one of many, why do you stick to it all the time?
Fine, I know that the new system isn't not innovative (sounds more like market speech to me) but it's surely an improvement over the freelist approach - how long they have been in use, since 1950 or?-)
I don't think Hyperion is just blindly adopting something new and cool without serious testing.
-
"A better way of doing things" - Isn't it better to have memory protection than not? "Their desire to bring AmigaOS to the the cutting edge" - Adding memory protection is helping to get that desire fullfilled. One step at a time...
Why all this memory protection comments, didnt you read the article?? It has nothing to do with Memory Protection, in fact the phrase memory protection never shows up in the entire article. Its hard to take your arguements seriously when we are talking about a Memory Allocation article and you decide we are talking about Memory protection.
So you're saying SLAB wasn't a technical innovation?
No, I'm saying putting slab in AmigaOS in 2005 isnt a technical innovation. But you know that, you are just arguing. Slab is an old concept its used in lots of OSs (including Windows). Rubber tires were a technical innovation at one time on cars, putting rubber tires on a car these days is not a technical innovation. If I wrote that my car had the technical innovation of rubber tires, you'd think that was silly. If Hyperion writes that they have implemented the technical innovation of slab memory management, you dance around excited.
-Tig
-
Okay, so you are say that Amiga OS 4.0 is the only current OS that is using this nearly 40 year old idea?
No, in fact if you were reading the thread you'd see I pointed out that most OS's use this method.
So, that would make Window one of the only true modern cutting edge inovative OSs!
Well it uses Slab too, so using Hyperion logic, apparently it is cutting edge and super innovative, of course it and I believe all of the others that actually use it, were designed to use it from the start, not trying to shoehorn it in ala what Hyperion is trying to do with AmigaOS.
-Tig
-
In the end the point is that Amiga OS 4.0 now has memory allocation algorithms that aren't {bleep}, and are comparable to what is used in other operating systems today.
What part of this is bad? Would you rather they had left it as it was?
For people who don't care about AmigaOS4 I wonder why you seem to care so much. I think you do it to troll, you're not adding anything to the discussion, you appear to have a personal vendetta against the Hyperion people, people that have at least put a lot of their time and money into a platform that you claim to love. Indeed you ignore technical responses that don't suit you, and nitpick on people's words. Grow up.
-
What part of this is bad? Would you rather they had left it as it was?
They took a working system with known inefficiencies and have replaced it with another system with different inefficiencies and potential issues running current software. Is that really better then fixing the real issues with the OS so it can be done.
For people who don't care about AmigaOS4 I wonder why you seem to care so much.
I and a couple of hundred other developers paid for a copy in OS4 over 3 years ago, it would be really nice to actually see the OS get completed.
I think you do it to troll, you're not adding anything to the discussion, you appear to have a personal vendetta against the Hyperion people, people that have at least put a lot of their time and money into a platform that you claim to love.
So now its trolling to point technical problems with what people post on there website?? The only Hyperion person I dont like is the Belgian midget, and thats just because he told everyone he was going to kick my ass at Amiwest, till I actually showed up and he realized that Lee as the little guy of the Huntsville Mafia could break him into pieces and me I wouldnt even break a sweat doing it.
Indeed you ignore technical responses that don't suit you, and nitpick on people's words. Grow up.
I havent ignored any technical responses. In fact the only real technical response to me has been Rich, and Rich agrees that Hyperion example is poor to say the least. The rest of Richs description is basically correct, his comment is that slab memory overhead could be less then of the current system (actually I think hes arguing that they are equal), traditionally slab memory overhead is usually larger then that of the amiga system, and if they actually have less memory overhead then on the current system, wouldnt that, not the poor example that slab doesnt help with be on the site?? Now why dont you grow up, and if you really dont like the thread, dont read it, or post in it.
-Tig
-
Now why dont you grow up, and if you really dont like the thread, dont read it, or post in it.
I like the subject of the thread, so why shouldn't I post in it?
My point was that people are making too much effort in deriding it, beyond simple academic interest. For some it is 'it's AmigaOS, so let's find an issue with whatever it is and then harp on about it for ever and ever'.
Even I agree the example was poor, but a poor example does not negate the fact that this replacement method is most likely a hell of a lot better than the old method. And better is good. It isn't as if people haven't been running AmigaOS4 beta for a long time anyway. There must be a good reason that many other operating systems use the same or a similar allocation algorithm - probably the usage situation is a lot different these days, and a list based mechanism will be a lot slower with hundreds/thousands of entries, whereas the slab system, whilst taking a little more memory, will be a lot faster?
Was the Belgian guy Ben Hermans? He did go on a bit on ANN in the day, a bit too far. Wasn't he let go from Hyperion?
I'm sorry you lost your money on the Amiga Inc coupon thing (I presume). Note that whilst being interested in AmigaOS4 I don't have an interest in Amiga Inc themselves, whatever they're doing these days. They messed around the community too much.
-
this replacement method is most likely a hell of a lot better than the old method. And better is good. It isn't as if people haven't been running AmigaOS4 beta for a long time anyway.
First of all, I'm not sure that this is actually in the Beta, secondly, where I think this is most likely to break things is JIT, which of course isnt in the version being used by most of the people.
There must be a good reason that many other operating systems use the same or a similar allocation algorithm
A couple of reasons, but they dont have (to quote Bernie) our one big hunk of memory design, and you might note all of them support multi-processor (which is why I've used a derivation of it in our stuff in the past). So amigaos doesnt have the reason to use it that others do, and again changing the memory allocation system would have been way down on my list of things to do to upgrade OS4.
Was the Belgian guy Ben Hermans?
Absolutely.
I'm sorry you lost your money on the Amiga Inc coupon thing (I presume)
Actually it was what I like to call the pre-coupon scam. They sold 1000 copies of AmigaDE Partypack, which was to include a free OS4 for $100 to a bunch of developers. The product was fairly useless, and its now actually 4 years later (that was actually 2001) and we still dont have an OS4. Of course being one of those that sat through fleecys developer seminar/beastiality conference in 2000, I'm not sure why I expected more from them.
-Tig
-
>First of all, I'm not sure that this is actually >in the Beta, secondly, where I think this is >most likely to break things is JIT, which of >course isnt in the version being used by most of >the people.
I think one of the Friedens stated somewhere that it does not break compatability with software. I understand the old allocation methods still work (excuse my lack of technical knowledge).
Also with OS4's JIT engine, software known to have problems with the JIT are put on a black list which can be ammended. IIRC When a black listed program is run it is referred to the interpretive 68k emulator.
Therefore programmes that work now should still work when the final version arrives.