I tried changing some BIOS settings (increased AGP aperture, disabled USWC) and it crashed again. This time, after the RESET I go some corruption during the BIOS splash screen, so it doesn't look good for the card, and I'm calling it quits for now . The card is going into my probably defective bin.

Have you tried cleaning the edge connectors on the card?

I would also try with a different power supply.

Could be something with the remainders of the Geforce drivers still being on the system... at least for the issues in Windows. That used to cause problems back in the day.

cyclone3d wrote on 2020-06-01, 01:37:

Have you tried cleaning the edge connectors on the card? […]
Show full quote

Have you tried cleaning the edge connectors on the card?

I would also try with a different power supply.

Could be something with the remainders of the Geforce drivers still being on the system... at least for the issues in Windows. That used to cause problems back in the day.

I'll give cleaning the edge connectors a shot, thanks .

I'm pretty much all out of spare 5V-heavy PSUs . The only one I have left is an Aopen 250W unit that probably wouldn't cut it with a Radeon 9700 . While not completely excluding a PSU issue, the corruption I saw at boot while the card was not stressed or stressing the PSU points toward a card issue at this point . That same corruption pretty much excludes a driver issue .

I'll probably re-test the card in my other retro rig at some point (that will test with another PSU) .

darry wrote on 2020-06-01, 01:54:

I'll give cleaning the edge connectors a shot, thanks . […]
Show full quote

cyclone3d wrote on 2020-06-01, 01:37:

Have you tried cleaning the edge connectors on the card? […]
Show full quote

Have you tried cleaning the edge connectors on the card?

I would also try with a different power supply.

Could be something with the remainders of the Geforce drivers still being on the system... at least for the issues in Windows. That used to cause problems back in the day.

I'll give cleaning the edge connectors a shot, thanks .

I'm pretty much all out of spare 5V-heavy PSUs . The only one I have left is an Aopen 250W unit that probably wouldn't cut it with a Radeon 9700 . While not completely excluding a PSU issue, the corruption I saw at boot while the card was not stressed or stressing the PSU points toward a card issue at this point . That same corruption pretty much excludes a driver issue .

I'll probably re-test the card in my other retro rig at some point (that will test with another PSU) .

Setting the card to AGP 1X in ATI's SMARTGART fixes the freezing . After effecting the change and rebooting, I was able to run 3DMARK2003 three time in a row without issue . I can only imagine that the corruption I saw on the BIOS splash screen was due to the card being left in a weird enough state by the crash that pressing the RESET button was not enough for it to recover .

EDIT: Before forcing AGP 1X it also crashed in 3DMARK 2001, so it was not purely a 3DMARK2003 issue .

Glad you got it working, phew!

I guess that it won't be bottlenecking that system. 😁

texterted wrote on 2020-06-01, 08:47:

Glad you got it working, phew!

I guess that it won't be bottlenecking that system. 😁

No it definitely won't be GPU bottlenecked, though the system is more of a testbed currently so the Radeon 9700 is unlikely to become a permanent addition to it . I primarily wanted to make sure the 9700 was still functional after removing the shim .

I had never thoroughly tested a 9700 in a BX board before and definitely wasn't expecting AGP issue with an Intel chipset . It just goes to show how fast things were moving back then, for a 3 or 4-year delta between hardware to be an issue .

I had that 9500 of mine running in my Gigabyte BX board for a short while. I didn't have any problems with it. Probably a driver issue with yours, I was using much earlier drivers than the 6.2's.

texterted wrote on 2020-06-01, 11:04:

I had that 9500 of mine running in my Gigabyte BX board for a short while. I didn't have any problems with it. Probably a driver issue with yours, I was using much earlier drivers than the 6.2's.

It could be a driver issue, though what clued me in to this was reading some old forum posts on the Internet from 2004 mentioning this workaround on 440BX boards . This was before Catalyst 6.2 's release, so if the issue is driver related, it probably started in older drivers or was always there to begin with .

Aye, all part of life's rich tapestry!

Totally unconnected, but what's the last two digits of the P/N on your 9700? Just that I've a warning in the manual of my Asus MB in my Athlon rig. About using 9x cards that end in less than "30". Dunno why, it just says to prevent instabillity.

texterted wrote on 2020-06-01, 11:53:

Aye, all part of life's rich tapestry!

Totally unconnected, but what's the last two digits of the P/N on your 9700? Just that I've a warning in the manual of my Asus MB in my Athlon rig. About using 9x cards that end in less than "30". Dunno why, it just says to prevent instabillity.

What chipset is your Asus MB on the Athlon rig?

texterted wrote on 2020-06-01, 11:53:

Aye, all part of life's rich tapestry!

Totally unconnected, but what's the last two digits of the P/N on your 9700? Just that I've a warning in the manual of my Asus MB in my Athlon rig. About using 9x cards that end in less than "30". Dunno why, it just says to prevent instabillity.

That is a very good question. The card is a Dell OEM Radeon 9700 TX .
Part number on the silkscreen on the PCB itself is 109-94200-01 .
Part number on the sticker is 1029421902 000001 .
There is another sticker that says REV A02 .

There is also a part number in the video BIOS, but I do not think that's trustworthy (as in may not be original video BIOS), because when I check that part number on techpowerup.com , it says card should run core at 276Mhz and mem at 263Mhz , but my card runs at 263Mhz for both like the other Radeon 9700 TX whose part number does not match the video BIOS' .
113-94216-102 is what I have in video BIOS .
113-94216-108 is the other one whose core clock speed would correspond .

Chipset on my A8V Deluxe is a VIA K8T800 PRO.

I've got a Radeon 9800 Pro that has a P/N ending in "00". If I just put it in the A8V with everything at default it will overclock the card, making it very unstable! Both GPU and memory are overclocked.
I have to set the FSB to 201 in the bios (from the default 200) for the 9800 to run at default speeds.

Weird things these Radeons sometimes!

texterted wrote on 2020-06-01, 21:02:

Chipset on my A8V Deluxe is a VIA K8T800 PRO. […]
Show full quote

Chipset on my A8V Deluxe is a VIA K8T800 PRO.

I've got a Radeon 9800 Pro that has a P/N ending in "00". If I just put it in the A8V with everything at default it will overclock the card, making it very unstable! Both GPU and memory are overclocked.
I have to set the FSB to 201 in the bios (from the default 200) for the 9800 to run at default speeds.

Weird things these Radeons sometimes!

I noticed that my Asus P3B-F's memory self-test at boot goes by almost instantaneously when I have the Radeon 9700 installed, compared to the Geforce FX 5500 , Geforce 6200 and Geforce 4 Ti 4200 that I have also tried .

Have the other cards got external power connectors?

Edit, well I know the ti4200 doesn't.

I'm thinking of grabbing an MX440 for my BX rig.

texterted wrote on 2020-06-01, 21:22:

Have the other cards got external power connectors?

Edit, well I know the ti4200 doesn't.

I'm thinking of grabbing an MX440 for my BX rig.

No, the 9700 is the only one with a power connector .