Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: Boot_WB on November 19, 2005, 10:34:29 PM
-
Am probably going to get a spider for my mediator when I get it fixed - does anyone know what the transfer speed is like using an IDE-USB or SCSI-USB converter.
According to Elbox the speeds are - potentially - remarkable (can't remember exact figures, but seemed better than powerflyer on the face of it).
So, given a standard IDE hd to boot off, and a nice big SCSI hd through the USB why would anyone still want a powerflyer?
Please don't flame my ignorance, I've never yet had USB on Amiga experience.
-
USB
is sloooow
is crap
wastes massive amounts of cpu power
Above comments apply to 3 Ghz Pentium 4 as well as .05Ghz 68060.
You should never use USB for high speed data transfers; that is exactly what USB is NOT made for.
USB is ok for, say, a digital camera that you connect once per day and transfer some pics over.
The worst possible use for USB that I can think of is to use it as a hard drive interface.
-
Fair enough, I'll shelve my plans for the spider for now.
-
I have a portable hd of 250 Gb that uses the usb 2.0 port and is very fast.
So i don't understand your statement.
-
Is that on your amiga, or your XP?
Elbox state compatible with USB 2.0 (480Mbps) and USB 1.1 (12Mbps) devices.
I know 'compatible' doesn't necessarily mean 'runs at this speed,' but surely it's not cripplingly slow like native IDE port.
Furthermore the mediator has DMA transfers between devices, so a usb cdrom - usb hdd should surely be a better, faster option than IDE or SCSI.
If anyone has a spider - hdd setup i'd be most interseted in a diskspeed test.
-
I am pretty sure that what they mean with compatiblity is that you can use USB 2.0 devices in the backwards compatiblity mode basically making it usb 1.1 and henceforth very slow.
-
usb 1.1 IS usb2 compatible. you an even say it is usb2.0. the hi speed portion of the usb2 standard is not supported afaik.
-
The spider might be a little bit faster then the Algor/Highway. At best about double the speed, resulting in about 2 MB/s transferrate maximum (the external USB-drive I have gives about 700kb/s on my Algor).
DMA isnt an issue here, as the Mediator can not DMA into FastRAM, but only into the RAM of the PCI GFX-card, from there it as to betransfered the "traditional" way.
Bottomline: USB on (classic) Amiga isn't appropriate for often used drives let alone the systempartiton at all. Besides you can't boot from the spider.
-
@ all
Lordy, I wish Elbox would state that in their description - something along the lines of "USB 2.0- compatibility at USB 1.1 speeds"
So, this speed limitation will also apply to the possibility of using a USB scanner, digital camera etc.
All of a sudden, USB on amiga 1200 seems less of a good idea.
With the mediator 4000 it seems it has DMA access to the motherboard space (presumably limited to the 16mb on board, not the extended fastram on a turbocard). Does the mediator 4000 then allow USB2 speeds?
Presumably if the Shark is released (or indeed ever exists) IT will provide dma access between PCI and fastram.
Can't seem to find anything on the Elbox site which mentions this though....
Pity I won't have the megabucks required to buy one.
-
Well, I'd recommend you don't trust Elbox marketing talk anymore you trust anyone elses.
It's quite obvious they'll fail to mention the things that make their hw less appealing. They seem to have a habit of recruiting some rabid followers to trash the competition, seem to have no problem in trashing your data, and calling their customers "underprivileged".
http://www.iki.fi/sintonen/elboxdecrypt/
http://www.theinquirer.net/default.aspx?article=6302
-
I'm stunned.
I'd read mention of the RDB trashing code on the Yahoo poseidon user's group, but didn't know the details.
I don't see how they can claim this doesn't destroy your data. I have a couple of hard drives where the rdb has become trashed (notably a toshiba which reads as a Uoshiba in HDtoolbox), and firstly - they come up as unsinstalled hdd, secondly - the hdtoolbox program errors out (error 4 I believe).
It does seem a rather rabid response from an allegedly professional company.
-
@Piru
You mean exactly what bPlan/DCE did with their G-Rex.
where they appropriately 'forgot' to mention that they had
only one DMA capable and one Bus-master capable slot.
not to mention how you yourself behave when it comes to MOS
(talk about rabid follower)
-
@trgse
Hello there, mr 'Adding broken Anti-MorphOS-code (http://www.amiga.org/forums/showthread.php?t=1883) to GPL applications'.
not to mention how you yourself behave when it comes to MOS
(talk about rabid follower)
How do I behave exactly? Would you be kind enough to describe in detail.
-
:flame:
-
Hmmm... there doesn't seem to be a smiley for ducking out of the crossfire. :-D
-
Lordy, I wish Elbox would state that in their description - something along the lines of "USB 2.0- compatibility at USB 1.1 speeds
they need to just use the approved logos.
usb.org (http://www.usb.org/developers/)
wait i read this on their site
• Five USB 2.0 Hi-Speed ports: four external and one internal
• Compatible with both USB 2.0 (480 Mbps) and USB 1.1 (12 Mbps) devices
• Compliant with EHCI, OHCI and PCI 2.2 standards
• All ports can handle high-speed, full-speed and low-speed transaction
does that mean they are outright lying.
-
As far as I understand the Spider / Mediator combo is actually running USB 2 but only as fast as it is able to get it going, resultin in not really that much more then USB 1.1 speed.
The same as with the 100 MBit drivers, where a 100 MBit card is faster then 10 MBit but by far not 10 time but only about 1.5 - 2 times.
Let's settle for "intentionally misleading information"
-
• All ports can handle high-speed, full-speed and low-speed transaction
By sticking to the letter say only say that the ports (-> the card itself) is able to handle those speeds. They don't say a word about the performance in the Mediator unter AmigaOS :rtfm:
-
@Lemmink
Indeed.
Elbox is also advertising Mediator 1200#? "transfer rate up to 264MB/s", but failing to mention it's only for PCI<->PCI transfers in optimum cases.
Sure it's a nice speed, but quite useless since most of the time you actually need the data to be moved to/from the system memory (or cpu). For that the actual speed is about 1/20th (or less) of the advertised speed.
-
I have a Prometheus PCI bridge, they advertised the transfer speed to be 12 mb/s from System to video card, I'm guessing it'd be similar for the Mediator too?
Robert
-
@Boot_WB
Powerflyer transfers are higher as the data does not need to be processed through the stack eating the CPU power.
@Piru
> Well, I'd recommend you don't trust Elbox marketing talk anymore you trust anyone elses.
You really do not know that Mediator/Amithlon drivers for the Spider card are USB 2.0 (EHCI) and USB 1.1 (OHCI)?
At least you, a person who is involved in the promotion of pirating the Spider drivers, should know if Spider drivers are 2.0 or 1.1? :lol:
@jkirk
> usb 1.1 IS usb2 compatible. you an even say it is usb2.0. the hi
> speed portion of the usb2 standard is not supported afaik.
So you know wrong.
Spider is High-Speed USB 2.0 card. Spider card includes the USB 2.0 and USB 1.1 controllers. With the Spider card Elbox supplies drivers for both controllers: EHCI (USB 2.0) and OHCI (USB 1.1).
The speed of the USB 2.0 devices connected to the Spider card is limited by the turbo card connected to the Amiga. Amiga turbo cards have not enough power to process USB data faster. The speed of the USB 2.0 devices connected to the Spider on the system running Poseidon with Spider drivers under Amithlon is exactly the same like with the Windows USB 2.0 drivers.
@Piru
> Elbox is also advertising Mediator 1200#? "transfer rate up to 264MB/s", but failing
> to mention it's only for PCI<->PCI transfers in optimum cases.
The same applies e.g. for Bugatti (www.bugatti-cars.de). They are advertising Bugatti Veyron 16.4 running at velocity 400km/h, but failing to mention it's only in optimum cases. :crazy:
Do you think Bugatti should mention in their site that you will not be able to drive their Bugatti Veyron at 400km/h due to the road quality or speed limitation in your country? :lol:
-
@tjaoz
a person who is involved in the promotion of pirating the Spider drivers
I am doing no such thing.
You must be seriously confused somehow.
The same applies e.g. for Bugatti (www.bugatti-cars.de). They are advertising Bugatti Veyron 16.4 running at velocity 400km/h, but failing to mention it's only in optimum cases.
Do you think Bugatti should mention in their site that you will not be able to drive their Bugatti Veyron at 400km/h due to the road quality or speed limitation in your country?
But is the Bugatti Veyron limited to 20km/h?
-
Piru wrote:
Elbox is also advertising Mediator 1200#? "transfer rate up to 264MB/s", but failing to mention it's only for PCI<->PCI transfers in optimum cases.
Sure it's a nice speed, but quite useless since most of the time you actually need the data to be moved to/from the system memory (or cpu). For that the actual speed is about 1/20th (or less) of the advertised speed.
It is less than that. I have a Mediator in my A1200T/060. I have run busspeedtest on it many times. I consistently get 9 MB/s which is only 2 MB/s faster than I can transfer data from slow old chipram.
There is a jumper on the Mediator that I can change which lets me then get 11-12 MB/s transfers but then the computer randomly crashes. I think due to the elbox gfx card driver is not compatible with 12 MB/s transfers.
Elbox annoyingly and intentionally does not mention these facts on their website.
-
@Piru
>> a person who is involved in the promotion of pirating the Spider drivers
> I am doing no such thing.
Yes, you do.
Even in this thread you link to your website which encourages pirating Spider drivers, where there are notes how to crack Spider drivers. Moreover, you also offer software tools needed for pirating these drivers.
> But is the Bugatti Veyron limited to 20km/h?
Sure, it is. In residential areas. :lol:
-
@ChaosLord
Busspeedtest to the graphic card in Mediator measures the maximum speed at which the turbo card can read the $200000 -> $9FFFFF area. Some A1200 turbo cards do at 6MB/s, others at 8-9MB/s, still others at 12-14MB/s, and some at 18MB/s. It all depends on the hardware design of the turbo card.
If busspeedtest in your A1200/060 reads 9MB/s, it means that your turbo card cannot access this area faster. It is the hardware limitation of your turbo card -- not limitation of Mediator.
Maybe you should submit complaints with the designer/producer of your turbo card, if they "annoyingly and intentionally do not mention these facts on their website"? :lol:
-
@tjaoz
Even in this thread you link to your website which encourages pirating Spider drivers
No, it does not.
where there are notes how to crack Spider drivers.
No there isn't.
Moreover, you also offer software tools needed for pirating these drivers.
No, I don't.
Again, I don't know how you manage to see the things you claim to see, but clearly you're confused somehow. Maybe you haven't actually checked out what the code in question does, or have problems understanding the technical issues handled there?
Clearly if there is something illegal on my web page, you should contact my ISP and Elbox right away. My ISP's abuse email address is abuse@kymp.net.
Remember to provide the evidence.
Busspeedtest to the graphic card in Mediator measures the maximum speed at which the turbo card can read the $200000 -> $9FFFFF area. Some A1200 turbo cards do at 6MB/s, others at 8-9MB/s, still others at 12-14MB/s, and some at 18MB/s. It all depends on the hardware design of the turbo card.
If busspeedtest in your A1200/060 reads 9MB/s, it means that your turbo card cannot access this area faster. It is the hardware limitation of your turbo card -- not limitation of Mediator.
So basically you're suggesting the A1200 CPU Slot interface could do upto 264MB/s with proper accelerator...
You must be kidding.
-
With the mediator 4000 it seems it has DMA access to the motherboard space (presumably limited to the 16mb on board, not the extended fastram on a turbocard). Does the mediator 4000 then allow USB2 speeds?
Mediator4000 can't do DMA transfers to the motherboard.
-
Spider is High-Speed USB 2.0 card. Spider card includes the USB 2.0 and USB 1.1 controllers. With the Spider card Elbox supplies drivers for both controllers: EHCI (USB 2.0) and OHCI (USB 1.1).
but if the intended application cannot benefit from that speed why claim it? you can't tell me it was designed for use between two pci slots. this is just a misleading statement nothing more.
Do you think Bugatti should mention in their site that you will not be able to drive their Bugatti Veyron at 400km/h due to the road quality or speed limitation in your country?
no i don't since it is obviously not designed for street driving at 400 km/h. but if on a closed circuit track you could not reach the 400 km/h then i would have major problems with the claim.
to put this into perspective. the government advertises you are not allowed to go faster than the limit. however you have nobody telling you that this card is limited in speed on an amiga w/pci slots. therefore i do believe they have a responsibility to tell us that speed is limited.
-
:whack: :flame: :destroy: :argue:
Yikes, (ducks under table to avoid chairs being thrown across bar)
I'm almost afraid to ask, for fear of inciting further aggression in the community :-)
From: http://www.elbox.com/products/mediator_pci_4000d.html
"
• DMA to A4000 motherboard space
Mediator PCI 4000D includes a busmastering controller supporting transfers between PCI cards and the A4000 motherboard space.
"
Is this, then, inaccurate?
If the 4000 supports dma to the mediator from fast & chip memory (and a noticable speed difference in memory-pci transfers) I'd consider making the switch from my 1200 - otherwise I probably won't bother.
-
tjaoz wrote:
If busspeedtest in your A1200/060 reads 9MB/s, it means that your turbo card cannot access this area faster. It is the hardware limitation of your turbo card -- not limitation of Mediator.
More than once I got the kind folks here to run all kinds of custom benchmarking tools I'd written to assess the real speed at which the CPU can access VRAM on their RTG cards (many mediators there) etc.
8-10MB/s is quite normal for VRAM write speeds just about every mediator system tested, regardless of CPU / mediator system.
-
tjaoz wrote:
Even in this thread you link to your website which encourages pirating Spider drivers, where there are notes how to crack Spider drivers. Moreover, you also offer software tools needed for pirating these drivers.
Whereas deliberately embedding malware into drivers that could trash someone's hard disk *even if they are not trying to decrypt the driver* (no memory protection in amigaos so no guarentee the trashing code cant be invoked due to an accident) is perfectly legal and moral eh?
-
Sony haven't been talking to Elbox, have they?
I don't want to knock Poland, but Polish companies seem to have rather loose views on ethics. That's just my personal experience and obviously a gross generalization.
-
jkirk wrote:
Spider is High-Speed USB 2.0 card. Spider card includes the USB 2.0 and USB 1.1 controllers. With the Spider card Elbox supplies drivers for both controllers: EHCI (USB 2.0) and OHCI (USB 1.1).
but if the intended application cannot benefit from that speed why claim it? you can't tell me it was designed for use between two pci slots. this is just a misleading statement nothing more.
I suppose if they ever release the SharkPPC :crazy: ...
-
Is this, then, inaccurate?
If the 4000 supports dma to the mediator from fast & chip memory...
A4000 with Buster 9 and 11 does support DMA to fast memory on the motherboard and the chip memory, as well as between Zorro III cards. DMA from Zorro III to CPU board memory must be supported by the CPU card (e.g. all phase5 cards do).
Regarding the Mediat*r4000: it has six Zorro III slots (plus one for the bridging card), but the Buster has only five bus request lines... strange. But to decide if it does DMA from Mediator to Amiga, check the following:
- does it work with Buster 7? If so, it is non-DMA, as Buster 7 doesn't support Zorro III DMA *at all*. There's no way of checking the Buster revision in software, so you would have to activate DMA by a software switch.
- does the software ask about Buster 9 or 11? DMA is buggy on Buster 9, and you need to take some workarounds for this, so if the software doesn't care about these differences, DMA is unlikely.
- check the bridging card between Zorro III and PCI: check pins 60 (/BRn) and 64 (/BGn) on the Zorro connector. These two lines are mandatory for Zorro III DMA (as well as E7M on pin 92); if they are not connected, Zorro III DMA is *not* possible at all. If they are connected and DMA is active, then short low pulses on /BRn (140ns, synchronized to raising edge of E7M) must be there, as well as longer phases with /BGn being low (which is DMA active).
Michael
-
@mboehmer_e3b
Greetings person with the unpronounceable name. :-D
I would like to thank you for such an informative message.
I wish more people would write messages like yours, all chock full of delicious and yummy technical infos!
-
Hi ChaosLord,
Greetings person with the unpronounceable name.
C'mon, it's not my fault that your grand-grand-grand-parents did forget the umlauts when they left Europe centuries ago :-D
I wish more people would write messages like yours, all chock full of delicious and yummy technical infos!
Sorry if it was too technical, but I was tired of some guys repeating what they see written on webpages, believing it without ever thinking about it :getmad:
Michael
-
I suppose if they ever release the SharkPPC
that is a mighty big if.
-
@Piru
> Again, I don't know how you manage to see the things you
> claim to see, but clearly you're confused somehow. Maybe
> you haven't actually checked out what the code in
> question does, or have problems understanding the
> technical issues handled there?
In the text on your website, someone (you?) writes what should be changed in the Elbox driver code to use this driver without Elbox restrictions. (These restrictions allow their usb driver working ONLY with the Spider card hardware.)
Next he (you?) describes what problem he (you?) found in the Elbox code during cracking it.
At the end, he (you?) writes: 'This archive was all made by me and contains code written by me. Nothing from Elbox.'
In the same way you can describe how to crack any other commercial software in order to break its restrictions.
Do you remember the email on the Mediator list with info how to break the Poseidon registration restrictions? This mail was treated as abusive despite that it only had info on which byte to change in the Poseidon code to remove restrictions on the unregistered version. You think such instructions are legal? :shocked:
Your website holds not only instructions what should be changied in the code of the Elbox driver, but you also include a tool for decrypting the code of this driver, which is necessary for cracking it.
> Clearly if there is something illegal on my web page,
> you should contact my ISP and Elbox right away. My
> ISP's abuse email address is abuse@kymp.net.
Sorry, it is not my busines...
Still, I wonder why moderators of Amiga.org allow you to post links to such texts here.
> So basically you're suggesting the A1200 CPU Slot
> interface could do upto 264MB/s with proper accelerator...
Why not?
The $200000 - $9FFFFF area is not used by A1200 motherboard at all. So what speed turbo card accesses this area depends on the hardware design of the turbo card only.
All accesses from the turbo card to this area are directed only to the external hardware (like Zorro II, Zorro IV or Mediators) connected to the CPU slot. If external hardware acknowledges cycles immediately (the case of the Mediator WAIT jumper opened), there is obvious that the speed at which this hardware is accessed is limited by the turbo card interface hardware only. :-D
-
@Crumb
> Mediator4000 can't do DMA transfers to the motherboard.
You are wrong.
-
@Trev
> I don't want to knock Poland, but Polish companies
> seem to have rather loose views on ethics.
It is funny to hear about ethics from someone from the US under Mr. Bush. :lol:
-
tjaoz wrote:
In the text on your website, someone (you?) writes what should be changed in the Elbox driver code to use this driver without Elbox restrictions. (These restrictions allow their usb driver working ONLY with the Spider card hardware.)
Really ?? All I could see was the bleeding obvious "look for the NEC USB pci subvendor ID" . Care to point out where Qwe (not Piru) points out the offset to the ID and what value it should be changed to ?
Do you remember the email on the Mediator list with info how to break the Poseidon registration restrictions? This mail was treated as abusive despite that it only had info on which byte to change in the Poseidon code to remove restrictions on the unregistered version. You think such instructions are legal? :shocked:
Like I said above, unlike the instructions on how to crack Poseidon, the article on Pirus site doesn't give any *useful* info on how to do the same to the Spider driver .
Even if you follow the instructions to the letter, the best you'll end up with is a messy asm source (cos IRA can't tell the difference between data and code in a code hunk, and doesn't know many 020+ instructions) that probably won't even recompile without changes.
@Piru: Did you get my PM last week??
-
@tjaoz
First of all, I am not 'Qwe'. I thought it was quite obvious, I am Harry Sintonen or 'Piru', not 'Qwe'.
In the text on your website, someone (you?) writes what should be changed in the Elbox driver code to use this driver without Elbox restrictions. (These restrictions allow their usb driver working ONLY with the Spider card hardware.)
I don't see such instructions or source code. I see some speculation what could be the reason for the protection to exist in the first place (1st paragraph of the 'History' section).
Where do you see these alleged instructions?
Do you remember the email on the Mediator list with info how to break the Poseidon registration restrictions?
No.
It could have something to do with the fact that I a) don't own a Mediator, nor have ever owned one b) don't read the Mediator mailing list.
but you also include a tool for decrypting the code of this driver,
That is true.
which is necessary for cracking it.
That is true, aswell.
It would be necessary to decrypt the driver if you were to crack it. I see no such thing anywhere on the webpage, however. If someone uses this tool for illegal purposes, it is not my fault, now is it? Also note that I do not host any Elbox software (pirated or otherwise), or tools/instructions on how to crack Elbox sw.
Anyway, the tool was obviously written to prove the claim that the driver contained such RDB trashing code. It can't even produce a working decrypted binary!
As a conclusion: Regardless what motives lead to finding of the RDB trashing code, it does not invalidate the findings.
People deserve to know.
> Clearly if there is something illegal on my web page,
> you should contact my ISP and Elbox right away. My
> ISP's abuse email address is abuse@kymp.net.
Sorry, it is not my busines...
Well, STFU then and let the moderators take care of the situation?
Still, I wonder why moderators of Amiga.org allow you to post links to such texts here.
I would think the reason they don't prevent me from linking to it is that they agree there is nothing dubious on the abovementioned web page.
> So basically you're suggesting the A1200 CPU Slot
> interface could do upto 264MB/s with proper accelerator...
Why not?
Because it's physically impossible?
The $200000 - $9FFFFF area is not used by A1200 motherboard at all. So what speed turbo card accesses this area depends on the hardware design of the turbo card only.
So the turbo card magically gets brain wave access to the area, without any need to send signals via wires or such silly stuff?
All accesses from the turbo card to this area are directed only to the external hardware (like Zorro II, Zorro IV or Mediators) connected to the CPU slot. If external hardware acknowledges cycles immediately (the case of the Mediator WAIT jumper opened), there is obvious that the speed at which this hardware is accessed is limited by the turbo card interface hardware only.
Newsflash: There are certain timing requirements for accelerators using the CPU slot interface, and if you break them the accelerator will not work.
This is quite obvious.
@Doobrey
Oops, missed it, sorry. You should have a reply now.
-
@tjaoz:
> > Mediator4000 can't do DMA transfers to the motherboard.
>
> You are wrong.
Then proove it.
Elbox claims that the Mediator4000 *D* (the Zorro III thing) could theoretically do DMA transfers between PCI and the mainboard, *but* *no* *existing* *driver* *is* *using* *it*. That's what Mr. Elbox himself said to me in Graz some years ago.
Why could that be?
-
(These restrictions allow their usb driver working ONLY with the Spider card hardware.)
If ones takes some minutes of thinking, one can come to the following conclusions:
- Spider(II) are normal bulk NEC USB cards
- software must distinguish a Spider(II) from NEC cards
- Spider(II) must have something different
- Spider(II) work under Windows and Linux
- difference is no hardware modification
- PCI cards have configuration ROM like Zorro cards
- the difference must be in some Flash memory
- the NEC USB controller has no Flash memory
- there's a SPI EEPROM on the card (like on NICs)
- compare PCI listings from a Spider(II) and a NEC card
- find the difference...
So no need to mess around with software code :-)
Nevertheless pirating code is illegal. As well as altering user data without permission.
Michael
-
@platon42
Thanks for the info re: a4000 dma. Don't think I'll bother making the switch to a4000 then. Too much cost, and memory/accelerator on a1200 is easier & cheaper than the a4000 equivalent.
Think I'll just get a powerflyer (which also seems to run at a LOT less than the16MB/s Elbox quotes :-) - the review quoted on the elbox website shows transfers of around 5MB/s, and if that's their PR review then you can bet normal transfer rates are slower) or blizzard scsi until the magical mysterious Shark pops out of the clouds into reality :lol: .
-
tjaoz wrote:
The $200000 - $9FFFFF area is not used by A1200 motherboard at all. So what speed turbo card accesses this area depends on the hardware design of the turbo card only.
All accesses from the turbo card to this area are directed only to the external hardware (like Zorro II, Zorro IV or Mediators) connected to the CPU slot. If external hardware acknowledges cycles immediately (the case of the Mediator WAIT jumper opened), there is obvious that the speed at which this hardware is accessed is limited by the turbo card interface hardware only. :-D
So things like glue logic on the card, phsyical connections etc simply cease to exist eh?
-
@Piru
> First of all, I am not 'Qwe'. I thought it was quite obvious,
> I am Harry Sintonen or 'Piru', not 'Qwe'.
Yes, I understand. You are 'Piru'. 'Qwe' is not you.
So, who is 'Qwe"? Your keyboard? :lol:
> I don't see such instructions or source code. I see some speculation
> what could be the reason for the protection to exist in the first
> place (1st paragraph of the 'History' section).
So if cracking instructions for Poseidon stack would be given in the form of speculations, would it be ok? :-?
> So the turbo card magically gets brain wave access to the area,
> without any need to send signals via wires or such silly stuff?
Are you really ignorant or do you only pretend to be so? :-o
> Newsflash: There are certain timing requirements for accelerators using
> the CPU slot interface, and if you break them the accelerator will not work.
:roflmao:
What requirements for accessing the $200000 - $9FFFFF area?
Could you be more specific?
-
@platon42
>>> Mediator4000 can't do DMA transfers to the motherboard.
>> You are wrong.
> Then proove it.
What for?
I know that Mr E3B has been trying to repair the Prometheus design for many months, as the original Prometheus does not have working PCI-to-PCI, and does not have DMA to Amiga motherboard at all.
Sorry, but I do not intend to help him. He should learn himself. :-P
-
I know that Mr E3B has been trying to repair the Prometheus design for many months, as the original Prometheus does not have working PCI-to-PCI, and does not have DMA to Amiga motherboard at all.
Hey, you are da man, you must be ingenious to find out... it was pronounced on several new pages.
But your information is outdated, as PCI-PCI DMA is running for months now on my Prometheus...
About direct PCI-to-Zorro III DMA conversion... well. I know it is not worth the time to explain to you, but if you were familiar with Buster design, you should know that it is not possible due to timeout constraints on both the PCI bus and the Buster side.
Michael
-
mboehmer_e3b wrote:
But your information is outdated, as PCI-PCI DMA is running for months now on my Prometheus...
Out of curiousity, is there any timeframe for when ordinary users will be able to send their boards in for reflashing?
About direct PCI-to-Zorro III DMA conversion... well. I know it is not worth the time to explain to you, but if you were familiar with Buster design, you should know that it is not possible due to timeout constraints on both the PCI bus and the Buster side.
Does this mean that PCI busmaster devices cannot access motherboard resources?
I don't know much about PCI and Zorro III timing so feel free to dismiss this outright, but couldn't Zorro III DMA be implemented with a series of individual PCI accesses since PCI is much faster than Z3. It seems a single PCI read or write operating takes around 5 cycles or so (or can it be longer depending on the device?). Can the buster chip not wait this long for each data transfer?
-
@mboehmer_e3b
> Hey, you are da man, you must be ingenious to find out... it
> was pronounced on several new pages.
What was pronounced on several news pages? That the original Prometheus does not have working PCI-to-PCI, and does not have DMA to Amiga motherboard at all?
When I wrote about these facts on Amiga.org there were always some "eyewitnesses" saying that PCI-to-PCI works well in their Prometheus boards. :lol:
> But your information is outdated, as PCI-PCI DMA is running for
> months now on my Prometheus...
I suggest you read my post again, and this time try to read it with understanding. I clearly wrote about the 'original' Prometheus, not about your experimental toy. Whether it works or not. :-D
> About direct PCI-to-Zorro III DMA conversion... well. I know it is not worth
> the time to explain to you, but if you were familiar with Buster design, you
> should know that it is not possible due to timeout constraints on both the
> PCI bus and the Buster side.
Many designs have been completed for Amiga, which others believed to be impossible.
You'd better wrote that making DMA to Amiga motherboard is too complicated for you. It's obviously not as easy as fix the PCI-to-PCI DMA in Prometheus. :lol:
-
@MskoDestny
> Out of curiousity, is there any timeframe for when ordinary
> users will be able to send their boards in for reflashing?
You can send your Prometheus board for reflashing since January 2002. :roflmao:
See details here: http://groups.yahoo.com/group/Amiga-Prometheus/message/1065
Forgive me, but I couldn't resist. :roll:
-
I have a powerflyer and trust me i wish i had never sold my blizzard scsi kit(didnt want to spend money on scsi kit at the time). blizzard scsi kit is true dma access. the powerflyer, whilst you get goodish transfer speeds, just watch your cpu usage hit 100% whilst transfering files.
-
Hi tjaoz. Welcome back.
Yes, we all know how this discussion ends.
Let's see, Cocolino mouse adapter, Elbox PCMCIA Right angle adapter, PowerFlyer, 4xEIDE'99, Mediator Multimedia CD, yes, all Elbox products have direct DMA to Amiga motherboard and it's absolutely pointless arguing that any other product is equal in performance in any way. Old news, move along. Next!
-
@tjaoz
I am curious. You seem determined that motherboard DMA is working for mediator4000. Could you provide us with a list of which products compatible with the mediator4000 use DMA to access the motherboard resources and some known examples where the feature gives a clear performance advantage over other solutions?
Presumably, DMA'ing to the motherboard memory is only useful for systems that have it. How does the feature help on systems that have local memory on accelerator cards?
Even if PCI -> motherboard memory DMA works, if you have an accelerator card in your system, the chances are you will end up having to use the CPU to shunt data between the local memory and the motherboard area accessible to the PCI.
-
@tjaoz:
>>>> Mediator4000 can't do DMA transfers to the motherboard.
>>> You are wrong.
>> Then prove it.
>What for?
Because otherwise, I suggest you STFU. The fact is, you cannot prove it. Remember, things like that could cause a war. Or was it you that claimed there were Weapons of Mass Destruction in Iraq?
And for the rest of the phrases: Smoke screen going off-topic. As usually. If you don't have any point to make, try jumping at something completely different. Pathetic.
-
Out of curiousity, is there any timeframe for when ordinary users will be able to send their boards in for reflashing?
We are running tests now with some external testers, and are evaluating the handling of the boards (shipment etc).
Please give us some time... it's far better to have it tested into the deep than giving away a fix which does not keep it promises (like the last reflash from the original designer :-(
Michael
-
...there were always some "eyewitnesses" saying that PCI-to-PCI works well in their Prometheus boards.
Yes, and always some people have seen a working Shark or Dragon... har har har... but seriously. The last DMA fix from Matay did work between PCI slots, but could be interfered from Zorro accesses, locking up the PCI bus.
You'd better wrote that making DMA to Amiga motherboard is too complicated for you.
Your struggling against reality given by existing bus and chip specifications is admirable.
Michael
-
mboehmer_e3b wrote:
We are running tests now with some external testers, and are evaluating the handling of the boards (shipment etc).
Please give us some time... it's far better to have it tested into the deep than giving away a fix which does not keep it promises (like the last reflash from the original designer :-(
Michael
I understand your desire to be absolutely sure given the costs involved with shipping boards all over the place. I also understand your reluctance to give a date since one problem could potentially push back the date significantly. I was just wondering if you could give some kind of educated guess. If you don't want to, I understand.