@harsh

Quote:

The mighty Commodore 64 did pretty well with serially connected drives even in a market that already had access to DMA-based devices

Just what does the C64 have to do with this? It could not multitask and therefore would have gained no benefit whatsoever from DMA hard disk access.

Paper cards with holes punched in them used to be an acceptable way of storing data, but that doesn't mean it's acceptable today.

Quote:

I've been tinkering with Amigas since Kickstarts were numbered in the upper .8x range. Back then, the argument was the same: The Atari 520ST was based on CP/M68K (a proven operating system based on the venerable CP/M for the Intel 8080 from which DOS was unabashedly ripped off), DMA hard drive access, VGA display, Yamaha sound chip, DOS format floppies and VT52 emulation built into the boot ROM. Clearly it would quickly and decisively win the battle against the Amiga 1000

This is completely irrelevant, but I feel compeled to answer in honour of the many ST vs Amiga flame wars of years past

As someone who owned and wrote software for the AtariST before going Amiga, I can state very clearly that it had no chance whatsoever of beating the Amiga. That 'proven operating' system was a pile of garbage that could not multi-task, had little real structure and was crudely ported from the PC. The display system was primitive, with only 16 colours, no hardware acceleration, and it needed a custom monitor to go above 640x200. The sound chip was a nasty device that played door-bell style 'music' and barely deserved to be called a sound chip in the first place, and the ST did not use DOS-format floppies. They were almost the same but not quite - the ST could read DOS floppies (as could the Amiga using CrossDOS) but ST formatted disks would not work in a PC because they used an invalid media descriptor byte.

The ST only sold because it was cheaper than the Amiga (that's why I bought mine, I just couldn't afford an Amiga then). As soon as the A500 came out ST sales went into decline, and when the A500 reached price-parity with the 520ST due to Commodore riding out a surge in memory prices better than Atari, the ST just died in its tracks.

Quote:

I submit that your theory about complete DMA support in the pre-release being critical to the long-term success of OS4 and the next generation of Amiga computers is similarly destined.

I normal circumstances no-one would care about DMA support in a beta version of the OS, but with rumours of a buggy DMA controller are floating around it becomes very serious indeed.

And to clarify, I don't personally care about DMA in the pre-release. What I'd like is a clear, unconditional statement from Hyperion that OS4.0 will work fully and correctly in DMA mode on the A1.

_________________
Who do you serve, and who do you trust? - Galen

@Rogue

Quote:

It already works. The sound card drivers as well as the ethernet drivers already use it, ok?

I'm sure the ethernet and sound drivers work just fine, they are not the issue here. The ArticiaS DMA problem is alleged to happen under very heavy DMA load with the CPU fully loaded. Using sound and ethernet is not going to provoke that as together they can barely push 11MB/sec through the North Bridge.

Modern hard disks, if you place one on each UDMA channel and copy files between them, can use more than 100MB/sec of bandwidth, saturating the bus and pushing the DMA controller as far as it will ever be required to go.

Does your DMA 'testbed' code allow you to do that?

Quote:

OS 4 is not tied to a specific harddisk. FYI, look at the reports of people having DMA problems and lockups on the Pegasos (even the Pegasos II), and tell me that this is a joke too.

I certainly regard the Pegasos as a joke, for many reasons. As it uses exactly the same chipset as the A1 and DMA glitches would surface on that system ,too. The Peg2 I can't comment on as I've seen no report of problems with it. Can you point me toward the ones you've seen?

Quote:

Is anyone of you actually listening to what I am saying? Probably not.

I'm listening attentively to everything you say, but you keep dodging the issue. I'm asking for a clear, precise answer to the question "Does OS4 work correctly on the A1 in DMA mode when the DMA controller and CPU are fully loaded?".

Waffling about sound and ethernet drivers does not answer that. Saying you have 'testbed code' does not answer it either. As far as I can see there are just three possible answers:

"We've tested under those conditions and DMA works fine."
"We've tested under those conditions and DMA failed."
"We have not tested under those conditions and don't know if DMA will work"

So which one is it, Rogue?

Quote:

Hardly. It takes more to get me upset, much more to get me screaming, and a lot more to make me throw stuff around.

Is that a statement or a challenge?

_________________
Who do you serve, and who do you trust? - Galen

@DrBombcrater

Quote:

Just what does the C64 have to do with this? It could not multitask and therefore would have gained no benefit whatsoever from DMA hard disk access. Paper cards with holes punched in them used to be an acceptable way of storing data, but that doesn't mean it's acceptable today.

Shhh....Don't tell that to Nintendo! Nintendo e-card

_________________
Men who have girlies in their avatars are Girliemen!

@Rogue

Quote:

Rogue wrote: @Darth_X Quote: Hidden cameras set up all over Hyperion's offices give over the shoulder's views of what the Hyperion developers are up to... This could be the biggest show in your life, if you had one

You are correct.. I don't have a life, I work fulltime nightshift!

_________________
Men who have girlies in their avatars are Girliemen!

@Rogue

Quote:

Rogue wrote: Quote: OBTW, the radeon PLL thing - was it an endianness issue when copying the BIOS tables to the video driver? I've got a bet with a friend, and I'd like to know who won But the good thing is they're working now.

so we can have our preview cds next week?

byebye...

_________________
Michael Merkel
(Michael.Merkel@gmx.net Home)
Member of Amiga-Freunde Pfalz, OS4 Welt

@Rogue

>look at the reports of people having DMA problems and lockups on the
>Pegasos (even the Pegasos II), and tell me that this is a joke too.

Please, don't blame the Pegasos when people are speaking about the
AOne DMA Problems:

So, let's make it clear:
- Pegasos with April does have WORKING DMA. I tested it by myself.
- Pegasos or any computer with VIA8231 SouthBridge DOES have IDE lock.
There are workaround in OS drivers. For example, I had one once under
Linux, I had no more disk operation during 5s, and then I could see
something like "IDE DMA TimeOut, resetting ..", then
everything continues ok.

I DID test a "fixed" AOne on Linux, and there WERE DMA trash. I
remind you it's was said to be FIXED.

By the way, MAI doesn't even claim anymore that the ArticiaS supports
DMA:
http://www.mai.com/products/articia%20s.html

Learving, the Pegasos as the the ONLY ArticiaS-based machine featuring
DMA.

Bye

@NicoPPC

Quote:

Learving, the Pegasos as the the ONLY ArticiaS-based machine featuring

Wrong

_________________
AmigaOne XE G3/933/VIA/FM801/R200 (fixed),
G3SE/600/Voodoo3/Sil680/RTL8139/SBLive! (noiseless!)
µA1-MK2/G3/933/R200/CMI8738
XE/G4 (broken 7450/800)

@NicoPPC

Quote:

NicoPPC wrote: I DID test a "fixed" AOne on Linux, and there WERE DMA trash. I remind you it's was said to be FIXED. Learving, the Pegasos as the the ONLY ArticiaS-based machine featuring DMA. Bye

That is a contradiction.


Bye.

@NicoPPC:
>Please, don't blame the Pegasos when people are speaking about the
>AOne DMA Problems

But the point is, that there are problems with IDE on the pegasos, too. And this problems have to be fixed, too. So give Hyperion some time to handle the DMA-issue an integrate it in AOS4.0.

>I DID test a "fixed" AOne on Linux, and there WERE DMA trash. I
>remind you it's was said to be FIXED.

AFAIK there is no totaly "fixed" Aone under Linux in the moment, because the "DMA"-kernel is still under development and in betatesting. Maybe there were some fixes to make DMA work better.

_________________
AmigaOne X5000/40, 2.2 Ghz, 4 GB RAM, Radeon R9 280X, M-Audio Revolution 5.1, 240 GB SSD

@DrBombcrater

Quote:

I've never even seen OS4 and certainly not tested it, but to suggest that PIO may be

and

Quote:

Would you be happy if MP3s started skipping because you copied some big files to

Contratiction in itself. How do you know mp3 starts skipping on the AmigaOne if you've never seen OS4 running on it? I can assure you, this doesn't happen, if in the current state.

_________________
AmigaOne XE G3/933/VIA/FM801/R200 (fixed),
G3SE/600/Voodoo3/Sil680/RTL8139/SBLive! (noiseless!)
µA1-MK2/G3/933/R200/CMI8738
XE/G4 (broken 7450/800)

@NicoPPC

Quote:

By the way, MAI doesn't even claim anymore that the ArticiaS supports DMA: http://www.mai.com/products/articia%20s.html

Just to clarify this. AFAIK the ArticiaS chip itself "never" supported DMA directly, because it simply doesn't have any DMA controller on the die. Other components of the motherboard (in other words, for example the sourthbridge) have to perform a DMA access, which looks more or less like a normal bus transaction for the ArticiaS. The ArticiaS just gives them the right to use the bus for a specific time. Nothing more....

Edited on 2004/03/05 16:18 GMT+1

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@DrBombcrater

Quote:

I'm sure the ethernet and sound drivers work just fine, they are not the issue here. The ArticiaS DMA problem is alleged to happen under very heavy DMA load with the CPU fully loaded. Using sound and ethernet is not going to provoke that as together they can barely push 11MB/sec through the North Bridge.

I think your original statement went something like "... if it [DMA] doesn't work". I just said that DMA indeed does work. I didn't comment on anything else.

Quote:

I'm asking for a clear, precise answer to the question "Does OS4 work correctly on the A1 in DMA mode when the DMA controller and CPU are fully loaded?".

I am not "dodging the issue". I told you that we have a testbed implementation of a DMA IDE driver. I said that this seems to work just fine. I don't even have that program, and I don't work on the IDE driver. No, we haven't done any stress tests. I never claimed that, and you never asked that until now. My take is that if there is any problem then these will be made public. IF there are.

DMA works, that is what I am sure about. That is a general statement. I am sure that it works because both the sound driver(s) and the ethernet driver(s) already make use of it. Whether there will be any problems with IDE is an entirely different topic, one that I haven't answered yet and will ONLY reply to when I actually know the answer.

What I am definitely not going to do is delay any work on OS 4 just to prove someone wrong or right. I've already explained in an earlier posting why this is the case.

Quote:

Is that a statement or a challenge?

Whatever, if you like such games...

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail

@NicoPPC

Quote:

Please, don't blame the Pegasos when people are speaking about the AOne DMA Problems:

I didn't "blame" the Pegasos for anything, I just contered the argument that DMA problems will kill off platform.

You might also want to be equally eager to jump in to such incorrectness when the next round of "Genesi didn't pay its employees" starts and certain people start saying that "Amiga Inc didn't do it either". I didn't see you there, though.

Quote:

Pegasos with April does have WORKING DMA. I tested it by myself.

As much as the AmigaOne, yes.

Quote:

Pegasos or any computer with VIA8231 SouthBridge DOES have IDE lock.

I think that is what I said, didn't I? Correct me if I'm wrong but the Pegasos II also is based on that, so where did I say something wrong?

Quote:

I DID test a "fixed" AOne on Linux, and there WERE DMA trash. I remind you it's was said to be FIXED.

That is a rather vague statement, isn't it? Can you quote the kernel revision you where using? Probably not.

Quote:

By the way, MAI doesn't even claim anymore that the ArticiaS supports DMA: http://www.mai.com/products/articia%20s.html

Well, my databook surely mentions DMA (page 45) and what do you think is "busmastering" about? "Anymore" would indicate that it said "DMA" before and now doesn't do so anymore.

Quote:

Learving, the Pegasos as the the ONLY ArticiaS-based machine featuring DMA.

As I already pointed out in the previous posting, the assertion is false. DMA is working nicely on the AmigaOne. Otherwise Sound drivers and network drivers wouldn't work at all, but in fact they do.

The claim is on the table that IDE DMA does produce corruption, under certain circumstances. That has yet to be proven. Until that time, your statement is rather empty.

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail

@Geri

Quote:

Just to clarify this. The ArticiaS chip itself never supported DMA, because it simply doesn't have any DMA controller on the die. Other components of the motherboard (in other words, for example the sourthbridge) have to perform a DMA access, which looks more or less like a normal bus transaction for the ArticiaS. The ArticiaS just gives them the right to use the bus for a specific time. Nothing more....

Correct. The Articia S has a PCI bus controller that arbitrates bus master cycles to busmaster capable devices, like sound cards, ethernet controllers, SCSI controllers and IDE controllers. The term "DMA" is nothing but the capability of a bus master to access main memory.

_________________
Seriously, if you want to contact me do not bother sending me a PM here. Write me a mail

@Rogue

Quote:

I am not "dodging the issue". I told you that we have a testbed implementation of a DMA IDE driver. I said that this seems to work just fine. I don't even have that program, and I don't work on the IDE driver. No, we haven't done any stress tests. I never claimed that, and you never asked that until now. My take is that if there is any problem then these will be made public. IF there are.

Thank you. That is the kind of answer I was looking for all along

I'll just keep the money I had earmarked for an A1 safely in the bank until such tests are done and I know the thng won't turn out to be the world's most expensive data shredder.

_________________
Who do you serve, and who do you trust? - Galen

@NicoPPC

The sound trashing problems on the Pegasos1 were not DMA related as was claimed now where they?

Hi,

I have no AOne at the moment .. ( wait for OS4 )
But ..
i will Use U160SCSI Drives and above with him ..
IDE is to slow and CPU hungry for RT-Video Editing

So .. IDE DMA Bug or not IDE DMA Bug .. if you will
POWER than use SCSI !!

[sorry for my suboptimal English]

R-TEAM

_________________
My Hardware Config and GFX-Work on my HomePage



Long Live T H E [|D|A|R|K^><^E|M|P|I|R|E|]

@R-TEAM

Wouldn't it be nice to have the possibility of doing:
gcc -O6 english.c -o posting

But alas, GCC needs some more work before it'll optimise the english language

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.

@Rogue

>You might also want to be equally eager to jump in to such incorrectness when the next >round of "Genesi didn't pay its employees" starts and certain people start saying that "Amiga >Inc didn't do it either". I didn't see you there, though

I'm not part of that people. I think your post could (and do) confuse people about the Pegasos.

You said Pegasos (I or II) has DMA problems. The only problem I can think of is IDE Lock. It's nothing comapre to DMA trash on AOne or no-fixed Pegasos (aka no April).

>As much as the AmigaOne, yes.

I will believe it when I wil I see it. And for the moment, I think it's not possible to perfrom safe DMA on AOne.

Bye

@NicoPPC

In Linux, no.

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.