Quote:

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 !!

You've got no clue mate!

S-ATA already has reached a point, where it could replace U160SCSI on the desktop.
The fastest (faster as current SCSI, yes) Desktop HDA is "Western Digital Raptor WD740GD 73.4GB" (SATA). Go and read www.storagereview.com before wasting your money.

As for video-editing:

Maxtor DiamondMax Plus 9 250GB, 8MB Cache, Serial ATA (6Y250M0)
This is well known to be optimal for any "streaming" data. It's actually the video editors hda of choice atm. At least for many. And yes, it is noiseless ;)

U160SCSI is okay for the server market, where many different users/processes access the disk at once. One can't say this about an NLE.

Go, check you facts and save lots of money before believing the hype.

Quote:

olegil wrote: @Crisot Hehe, the other day I benchmarked my AmigaOne copying a 128MB file over GigE. I think 35MB per second is rather nice, don't you?

Off-Topic:

No, it is not. This is the limit of CAT5e (350Mb/s). A 1Gb Ethernet should do waaaay more. Gb Ethernet in the desktop sektor still is a hype. It's the cables, that are missing.

/me, who has CAT5e "in wall" everywhere

Quote:

DrBombcrater wrote: It seems highly unlikely the 686B is responsible for the DMA problems. There are millions of them out there in x86 systems and as far as I know no data integrity problem has ever been traced back to the 686B directly. Badly set-up North Bridges did cause some problems, but nothing that couldn't be fixed by a patch.

Hmmm.... I remember back in the Athlon T-Bird days there was quite some fuzz about the 686B when used with a Soundblaster doing disk-corruption. So I checked UseNet and found this:

Some random UseNet post

and later:

The Register reports about VIA's confirmation

So, it could well be, that there are other 'cases', not only Apollo133 & Soundblaster.

@CodeSmith

Why don't you read the results of your own Google search? Yes, the 686B has bugs (just like every other IDE controller) but all of them can be worked around either by tweeking the drivers or by making sure the North Bridge is set-up correctly. There's no doubt that it is a picky device that needs careful handling and badly written drivers or poor board quality can screw it up badly, but if you treat it right the 686B will get the job done.

One theory I've developed over the past few years, having built and maintained several hundred systems with 686B chips in them, is that it takes the blame for many 'bugs' that are due to faulty motherboards. The 686B was used in huge quantity through the period where PC motherboards were often fitted with capacitors that contain incorrectly formulated electrolyte. When the caps go bad and burst, as they invariably do, the IDE ports are often the first thing to go wrong because the 686B is very sensitive to signal and timing tolerances.

In many cases only a few caps go bad, just enough to cause problems with the IDE bus without destabilising the rest of the system. I've seen many dozens of 686B-based boards go like this, including several of my own. When the caps were replaced all the IDE problems - corruption, lock-ups, failure to detect devices - all went away.

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

@amix

The 'Soundblaster' problem was down to the SBLive flooding the PCI bus with useless requests, causing the PCI arbitration logic in the North Bridge to starve the 686B of cycles. It also happened on some Intel chipsets (particularly the i815), but to a much lesser degree as Intel's PCI arbitration logic is smarter and less aggressive than VIA's. BIOS revisions and improved SBLive drivers pretty much cured it in the end.

(and before anyone jumps in and tells me I don't know what I'm talking about - I do. I wrote a patch for Amithlon to address this problem, and a few related ones, for systems that didn't have the required tweeks in the BIOS. I know more about VIA chipsets than I ever wanted to )

But it's an interesting point, and one I'd stupidly not thought about till now in relation to the A1. It would explain a great many things as the symptoms of a PCI arbitration problem would look very similar to the alleged ArticiaS DMA bug.

It'd be nice to know if anyone's seen the 'DMA bug' on an A1 with no PCI cards in it...

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

@elwood

Quote:

@NicoPPC Quote: Elwood, the ArticiaS IS buggy You can't say that because you're not an engineer. Don't trust others, just trust what you see.

My dear elwood (hello M8 ).. I'll explain you why Articia S is 100% buggy, really buggy, buggy buggy.

It is buggy because Genesi/BPlan said it...

He trust what he see, but what he see has to come from Genesi to be true :->

I hate this. The same guy saying to A "prove what you say" and to B "What you say is true". Don't say me I am false, the Pegasos experience NEVER proved Articia S is buggy.

This was the last off topic post for me, guyz, if you want to fight like kids, create another please. I don't know where? ANN? MOS.DE?

_________________
AmigaOne XE --- 7455A@1400 --- Radeon 9000 Pro --- RAM 2GB --- SSD 40GB

@thread

/me shrugs

A1 hasn'y presented any real problems to me for some 5 months now. The Articia certainly isn't a monster

@DrBombcrater

Exactly - but my point is that all the problems that are being blamed on the Articia chipset can be equally attributed to the VIA 686B (unless you're saying that the AmigaOne and the Pegasos 1 both suffer from shoddy workmanship). The reason why these problems are no longer a big deal on the x86 side of the fence is that VIA gave BIOS manufacturers instructions on how to work around the problems. These instructions, however, came with NDAs. This is why there is no version of Linux that fully addresses the problem (that's OK on x86, because by the time Linux gets hold of the system, the BIOS has already correctly initialized the 686B). The Hyperion crew have several times already bitched about VIA not giving them the information they needed to make the same fix on UBoot, so they have had to spend several months trying to reverse-engineer the fix (they are not allowed to look at x86 BIOSes for the info, because that would be a copyright infringement).

So, my point throughout all this has been that it's very likely that the problem has been with the VIA chipset all along, and that is a problem that VIA themselves have said is fixable in software.

@Crisot

There's no fight at all. I just wanted to be honest to the people that read this. They must know that until we have an OS4 IDE driver, there's no bug. Plain and simple...

_________________
Philippe 'Elwood' Ferrucci
Sam460 1.10 Ghz
AmigaOS 4 betatester
Amiga Translator Organisation

@DrBombcrater

Quote:

It'd be nice to know if anyone's seen the 'DMA bug' on an A1 with no PCI cards in it...

I havent got ANYTHING except me AGP graphics card. - how do I check it / how does it manifest itself ?

mind you, I dont exactly thrash me system (and Xine plays MPGs fine).
Plus, I only got an ancient flat 40-way IDE cable from me A1200 (cheapskate arent I !)

however, it sounds (to me) a lot like a combination of:
a) iffyness in the 686B
b) driver writers not getting the info to fix the problem cos it aint for Wintel,
c) CRAP soundcards.
d) experimental mobo syndrome.

I am after a soundcard at the mo, what would be recommended ? (not too expensive. and place it slightly higher so we get a two-level effect and a path. )

_________________
XH558 - the worlds last flying Vulcan.
ok, its actually XL426 in the picture but you know what I mean.

@Crisot

Quote:

My dear elwood (hello M8 ).. I'll explain you why Articia S is 100% buggy, really buggy, buggy buggy. It is buggy because Genesi/BPlan said it... He trust what he see, but what he see has to come from Genesi to be true :->

I trust that I had see on my Peg I non April and on AOnes.
Genesi has nothing to do here, as they haven't "show" me anything but:
- They said my Peg I (non April) was buggy
- They were right.
- Then, they propose me to replace it (for free)
- They did. and no more problems...

Now concerning, the AOne. It use the same NB than the Peg I without patch, and this NB is said to cause the problems on Peg I (non April).
There are the same problems on AOne.

That's it. Now, let's see if it's fixable in software without killing the perf

I trust that I see: and I saw that the AOne can't handle DMA safely. I will consider it can, if I saw it doing it properly.

And no, I don't trust Genesi as if I would be blind.

@Crisot: Nescafé

That's it.

@amix

Uhm, I regularly do 600Mbit on the same cable between faster processors, so you're the one who doesn't know what you're talking about.

The G3-SE is too slow to play ball with the P2 2.2G on gigE here
I posted complete benchmarking results in a thread last week, go look...

_________________
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.

@NicoPPC

Quote:

the AOne can't handle DMA safely

No, no and no !!!!!!!!!!!!!!!!!!!!!!!!!!!!
Not the A1 but the Debian IDE driver ! When will you understand ?

_________________
Philippe 'Elwood' Ferrucci
Sam460 1.10 Ghz
AmigaOS 4 betatester
Amiga Translator Organisation

@NicoPPC

Quote:

It use the same NB than the Peg I without patch, and this NB is said to cause the problems on Peg I

Thanks, you just disqualified yourself. The AmigaOne uses a later revision of the Articia S than on the Peg I. Only a handful of boards have been produced that use the same revision. Among other things the new revision had the ability to use two unregistered memory modules.

So please stop talking about things that you do know nothing aboute.

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

@Rogue

Quote:

Among other things the new revision had the ability to use two unregistered memory modules.

That's very nice to know, Rogue, thanks.

It would be nice if Eyetech would actually tell people about stuff like this. I, for one, am very fed up with discovering important facts about the A1 - like unregistered memory, the dead on-board sound, etc - from random forum and mailing-list posts. Such things are easy to miss and often hard to verify.

I guess anyone who's just dropped a chunk of cash on registered memory is going to be more than a bit cheesed off right now.

To Eyetech: If you can't even manage the basic step of putting up a web page listing detailed specification and compatibility information for your main product then, frankly, you should not even be selling sausages-inna-bun to passersby, nevermind something as complex as PPC motherboards.

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

@DrBombcrater

Quote:

I guess anyone who's just dropped a chunk of cash on registered memory is going to be more than a bit cheesed off right now

Well, I tried 4 unregistered memory sticks, none of them worked ok. The only one that didn't give problems was the expensive Kingston one, so if you have a working unregistered stick, you're just lucky IMHO.

Davy

_________________
Audio Evolution
http://www.audio-evolution.com

@DrBombcrater

Quote:

I guess anyone who's just dropped a chunk of cash on registered memory is going to be more than a bit cheesed off right now.

As Dave already pointed out, the ability is only theoretical. I have encountered two modules that I could actually use, but these where noname 64 MB modules. YMMV, I would in any case recomment to check with the dealer and possibly ask other AmigaOne users about their experience.

The big issue is that many onboard EEPROMS on SD-RAM modules actually have false or sloppy programming, making it hard for a non-x86 BIOS to actually program the memory controller correctly.

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

@DrBombcrater

In my opinion, Eyetech do a reasonable job in this arena. however, they arent a dealer anymore, but they should probably keep the dealers informed.

sausages-inna-bun ? thankyour Mr C.M.O.T. Dibbler !

I'm a bit frustrated by the broken onboard sound, but only a smidge. I gather AC97 is poo anyway.

whats the performance difference between reg and unreg memory anyway ?
as far as I can tell, its a grillion times faster at doing X11 than my humble (see below) AGA machine...

_________________
XH558 - the worlds last flying Vulcan.
ok, its actually XL426 in the picture but you know what I mean.

@Rogue

'YMMV' - "your mileage may vary" thats a new one on me ! (took some working out too !)

EEPROMS and memory controllers on memory sticks ? and heres me thinking they just contained RAM chips ... sheesh what next ?

_________________
XH558 - the worlds last flying Vulcan.
ok, its actually XL426 in the picture but you know what I mean.

@Rogue

That's very interesting. I didn't know that there was actual program code on RAM chips themselves.

_________________
Sam