good ram advice to BURN an old slot A Athlon TB

[billchase]

"VIA's, but W2K reports Ultra DMA as the current mode so it's working. I've also found some references on the net advicing to remove DMA tool before installing VIA drivers so I assume the drivers substitute it."

I actually I had heard that DMAtool was sometimes required to activate DMA
with the VIA drivers installed.  Maybe newer versions of the drivers eliminate
the need for DMAtool.  Probably wouldn't hurt to try it anyway.

Are you running with the ACPI HAL installed?  I have also not had any
luck with ACPI on older hardware including irongate boards. If you are going
to do another install of Windows anyway, try just using the Standand PC HAL
if you are not already.

Good luck with it, outside of what has been suggested, I am not sure what else
it could be.

C Snyder

[Jose]

Hi. Got bored and installed the 1Ghz Athlon. CLearly makes a difference but still not smooth!! Lowered the resolution to 800x600. Got a bit better but still not completely smooth.
Do you think getting a better video card would make a difference ? My current one is old and has only 8MB, but PowerDVD docs say 8MB is enouph. Also the card supports overlay and works in AGP 2x mode.

What the heck is going on  :-?

ACPI is working, but (weird) when I overclock the FSB to 110 (+10%) a blue screen with white font appears complaining the BIOS is not ACPI compatible.

[billchase]

ACPI will "work" on Slot A boards, but the reliabilty and
stability is questionable.  When I was running W98, having
ACPI enabled (I forgot to use setup /p i during the install)
caused a big performance hit and stability issues.  W2k's
implementation might be better, but I still wouldn't trust
it.  Your video card might be the bottleneck as well, but
regardless, I highly recommend ditching ACPI due to the age
of the machine.

Do you have NERO by any chance?  The Nero Infotool will help
confirm that DMA is indeed working.  I have heard of windows
incorrectly reporting DMA status on VIA equipped boards.
I apologize if I'm beating a dead horse on this DMA issue, but
my gut instinct tells me its still a possible factor.  

C Snyder  

[Jose]

I got Nero and run the Infotool. DMA is active, at least it says so. I also run SiSofSandra and compared my system to a few on the list... My system is practically equivalent to a P4 at 1.6Ghz! I bet it will beat the crap out of it when I overclock it.

Anyway, I'm gonna take it to a friends PC store and ask to try it out with a different Gfx card, WITHOUT hardware decoding of course...

:pint:

[billchase]

Well that certainly helps close the door on the DMA issue.
My last random suggestion would be to try using an 80wire
ide cable instead of the old 40wire if you are not already
doing so.  Have you also been able to test with another DVD
drive?  If the GFX upgrade doesn't work, this is the only
other hardware factor I can think of.

C Snyder
 

[Zac67]

Jose wrote:

Yeah, but 6MB/s should be more than enouph anyway, MPEG2 is around 3.5MBits/s IIRC, unless the machine is unstable.

Don't have the specs at hand right now, but afaik a DVD peaks at about 8.5 Mbps and may use a sustained rate of at least 6 Mbps. Since you've got Nero, use CD-DVD Speed to check the transfer rate of a data or video DVD - it should never drop below 1x. If the graph isn't smooth, there's something wrong with the DVD or the drive.
If it's OK the bottle neck is somewhere else: some process eating up CPU power, graphics driver problems, AGP trouble, DirectX, ...

[Jose]

After checking in device manager and also NeroInfoTool that DMA was on,  I wasn't even gonna run CD-DVDSpeed but I decided to give it a go...

Speed: Cpu Usage:
1x     12%

2x     22%

4x     98%:-o

8x    100% :-o  :-o

Still, DVD playback is 1x and running other tests it never dropped below that. The graph for speed wasn't regular but the peaks it had seem to be related to the kind of drive.

100% @ 8x is weird isn't it ? Now I have to determine if this the systems fault or not. Suggestions...(I checking for similar problems with this drive model now...)

[Jose]

The drive is a Samsung DVD-Rom SD-616E BTW...

[Floid]

I'm too lazy to think about this IDE problem, but...

-The Slot A boards used a slightly different electrical arrangement than the Socket A ones.  I forget the details, but it suggests that pushing the FSB is a little trickier (if you're used to Socket A boards, where it was a piece of cake).

-I can confirm that the K7M has some delightful ACPI quirks, some of which may be worse in the final beta BIOS release for it.  In fiddling with power management under DragonFly, I managed to get it into a nice 'fire hazard' state - the system woke, but not the CPU fan(s).

-Do you have a manual handy?  This is a pretty 'classic' board, and there's a PCI slot interrupt-sharing map to be aware of.

Edit:  Oh yeah, this thing can't handle '32x4' or whatever the common "high density" SDRAM is.  So if you want to use SDRAM sticks >128MB, you basically suffer the same limitations of the old 440BX.  If you spend through the nose for name-brand motherboard-matched RAM (e.g. Kingston), you'll be okay, of course.  (And I now notice someone else already said that, but the issue is more annoying and befuddling than 'sidedness.')

[Zac67]

Jose wrote:
Speed: Cpu Usage:
4x     98%:-o

8x    100% :-o  :-o

Seems like a DMA problem after all - try disabling UDMA in BIOS/DMATool, so hopefully Multiword DMA is used which might even work (and is fast enough for DVD).

[Jose]

@Floid

Oddly enouph, I'm using the latest beta BIOS and ACPI seems to work fine, but I'm using Win2K wich seems to have a way better implementation than Win98.
I remember when I was running Win98 the system wouldn't even be able to shut down ALOT of times :lol: That disappeared when I got Win2K.

@Zac67

Yes, but at 1x the CPU load, though too high for normal DMA, shouldn't overload the CPU enouph to prevent MPEG2 decoding at a decent speed.
So, I noticed PowerDVD has some other tests and I decided to run the System Diagnostic and then do a diagnostic to test my Gfx card (which was being reported as having overlay and everything).
Didn't pass ANY of the YUV tests or any...! So it's the Gfx card too...

Looking on eBay for a Radeon 9700 right now, the fastest AGP2x compatible one. Before anyone says anything, the card will still be useable if I upgrade in the future so I don't mind if it's a bit overkill for my current system. And its going cheap on eBay. I think I'll get the Gigabyte one that has way better colling.

Still, I'd like to resolve the DMA issue if possible ...    

[bloodline]

Jose wrote:

Looking on eBay for a Radeon 9700, the fastest AGP2x compatible one. Before anyone says anything, I guess the card will still be useable if I upgrade my system in the future so I don't mind if it's a bit overkill for my current system. And its going cheap on eBay.

Any AGP card is obsolete now, So don't pay over the odds.

[Jose]

@bloodline

Still plenty AGP boards out there selling for 64 bit systems :-)

[stefcep]

FWIW I play full screen DVD's with no frame skipping on a P3 650 and a geforce 32 meg 2xAGP using Win98SE.  The dvd is master on secondary cable.  Like everyone else has been saying you should have ample CPU power

[Jose]

Can someone recommend a test program that will test CPU load under DMA transfers to HD ? From SisSoftSandra (trial version here) results my HD manages 33MB/s, not bad since it's an ATA66 controler.

I'm assuming that if HD transfers don't have high CPU load then the fault for high cpu load in DVD accesses would be in the DVD itself or it's drivers..

[Jose]

Hi. Be amazed!

Just runned Quick Bench...

Sequential Speed: Read- 39.2 MB/s
        Write- 33.8 MB/s
...
...
Average CPU- 2%  !!!!!!!

The average CPu was a bit higher when I first launched the program (>10%), lowering right after if I kept idle, but I think that was due to other task having just been terminated. If I kept idle and stoped the test in the middle and restarted again the CPU load would be 2% in the very beginning.

Now what it could be? The DVDROm drive, it's drivers, or the cable...?