A couple of benchmarks, all carried out under Windows 98SE. For the 3DMark entries, the standard benchmarks were utilized, in whatever resolution and color depth each one uses by default.

All game benchmarks were carried out using 32-bit color depth. Game versions used:

• Quake 2 v3.20
• Quake 3 v1.32c (using demo FOUR with all in-game graphical settings at maximum)
• Unreal Tournament 99 GOTY (Direct3D renderer, all in-game graphical settings at maximum including Detail Textures)
• Drakan v445 (Direct3D renderer)

OS and drivers used

• Windows 98SE retail
• DirectX 9.0c
• Unofficial ICH9 drivers for Win98
• ATi Catalyst 6.2 reference drivers
• Official Audigy2 ZS installation CD (using VxD drivers)

Note: the Audigy2 ZS drivers are installed as per this guide of mine. Also, Everest mis-identifies the PowerColor X800 GTO 16 as an X800 XL due to its 16 pipes which are officially unlocked from the factory. More info about that can be found in this Anandtech review.

Beautiful! I have a very similar setup with a Gigabyte board. I run an L5240 @ 3Ghz and it runs a lot cooler than an E8xxx. I'm talking in the 30s at idle and I haven't measured under load but I can't imagine it going over 50 and this is with some very conservative cooling. It's a 45w TDP part... It requires BIOS and CPU modding though, but it's well worth it, in my opinion.

Why aren't you using two RAM sticks? No matter how much you limit the memory to, you will still get the speed boost benefit of interleaving (read: dual channel mode).

Can you also do a DEMO001/sound disabled benchmark please? Thanks

mockingbird wrote on 2023-09-10, 22:44:

Why aren't you using two RAM sticks? No matter how much you limit the memory to, you will still get the speed boost benefit of interleaving (read: dual channel mode).

For some reason, using two 2GB sticks in dual-channel mode severely reduces GPU performance under Win98 on this motherboard. It's been discussed here.

Can you also do a DEMO001/sound disabled benchmark please? Thanks

Sure, here you go:

Quake 3 version 1.32c using DEMO001

• 640x480x32 - 819.7
• 800x600x32 - 606.8
• 1024x768x32 - 466.9
• 1280x1024x32 - 340.1
• 1600x1200x32 - 247.8
• 1600x1200x32 + 16xAF - 208.5

For this test, all in-game graphical settings were at maximum and sound was disabled.

EDIT - updated benchmarks with 2x1GB memory running in dual-channel mode.

Here's Deus Ex running at 1600x1200 with 4xAA and 16xAF on this system:

This is using the stock Direct3D renderer with detail textures, vertex fog and volumetric lighting all turned on. Doesn't drop below 100 FPS on Liberty Island, which is an awesome feat for a Win9x rig.

Also, sorry for the poor brightness in that image, but Deus Ex doesn't correctly capture gamma settings when it takes a screenshot. The picture is much brighter in-game, but you should still be able to make out the framerate counter on the right side of the screen.

Very nice build, congrats! 😀

bloodem wrote on 2023-09-11, 10:55:

Very nice build, congrats! 😀

Thanks man! Couldn't have done it without your help! 😀

BTW, do you have any advice on cooler maintenance for X800 cards? I'd like to re-paste mine and possibly also lubricate the fan, but this thing looks like it wasn't built to be easily disassembled.

Joseph_Joestar wrote on 2023-09-11, 11:11:

Thanks man! Couldn't have done it without your help! 😀

BTW, do you have any advice on cooler maintenance for X800 cards? I'd like to re-paste mine and possibly also lubricate the fan, but this thing looks like it wasn't built to be easily disassembled.

You're welcome! 😀
I was actually going to mention the X800 cards and their coolers... I have a few X800XT and X850XT cards, and they are annoyingly LOUD and HOT.
I'm not familiar with the cooler on your particular card, my cards have the typical plastic shroud (as seen in these photos from TechPowerUp).
Those are fairly easy to clean: once you remove all the screws from the card's back, the whole cooler/heatsink assembly comes right off, and on its bottom side it has a bunch of tiny screws that hold the whole plastic enclosure together.
I imagine that yours should be similar, even if it has a different appearance on the surface.

The best option I can think of for these cards would be the Arctic Cooling ATI Silencer 4 & ATI Silencer 5 coolers, but sadly they are very hard to obtain. I actually have quite a few NOS ATI Silencer 3 rev.2 coolers (compatible with the Radeon 9800XT cards), and I was thinking of trying to exchange one of them for a new ATI Silencer 4 and/or 5 cooler.

bloodem wrote on 2023-09-11, 11:42:

I was actually going to mention the X800 cards and their coolers... I have a few X800XT and X850XT cards, and they are annoyingl […]
Show full quote

I was actually going to mention the X800 cards and their coolers... I have a few X800XT and X850XT cards, and they are annoyingly LOUD and HOT.
I'm not familiar with the cooler on your particular card, my cards have the typical plastic shroud (as seen in these photos from TechPowerUp).
Those are fairly easy to clean: once you remove all the screws from the card's back, the whole cooler/heatsink assembly comes right off, and on its bottom side it has a bunch of tiny screws that hold the whole plastic enclosure together.
I imagine that yours should be similar, even if it has a different appearance on the surface.

Cheers! I managed to clean and re-paste it:

Interestingly, it looks like the previous owner (or someone else) had already done this, but it may have been some time ago. Anyway, with fresh thermal paste and a newly lubricated fan, the noise went down about 5% according to my highly scientific testing method of "put your ear to the case". 😁 Also, temperatures under full load dropped by 4-5°C which is nice as well. But yeah, I wouldn't exactly call these cards quiet, even under the best conditions.

Joseph_Joestar wrote on 2023-09-11, 09:24:

Sure, here you go: […]
Show full quote

Sure, here you go:

• 640x480x32 - 750.3
• 800x600x32 - 604.9
• 1024x768x32 - 467.4
• 1280x1024x32 - 340.1
• 1600x1200x32 - 247.9
• 1600x1200x32 + 16xAF - 206.7

Very nice!

Interesting, I'll have to remove a stick with my setup and see if that gives me a boost...

A faster GPU will improve the FPS - not that it's required 🤣

Joseph_Joestar wrote on 2023-09-11, 13:08:

Cheers! I managed to clean and re-paste it:

Interestingly, it looks like the previous owner (or someone else) had already done this, but it may have been some time ago. Anyway, with fresh thermal paste and a newly lubricated fan, the noise went down about 5% according to my highly scientific testing method of "put your ear to the case". 😁 Also, temperatures under full load dropped by 4-5°C which is nice as well. But yeah, I wouldn't exactly call these cards quiet, even under the best conditions.

Looks good! Just curious, what temperatures are you seeing now during full load?
I tested a PC with an X800XT about a month ago using AtiTool (which only works in WinXP), and I was seeing 85 degrees C, even with a 21 - 22C room temperature.

bloodem wrote on 2023-09-11, 14:02:

Looks good! Just curious, what temperatures are you seeing now during full load?
I tested a PC with an X800XT about a month ago using AtiTool (which only works in WinXP), and I was seeing 85 degrees C, even with a 21 - 22C room temperature.

I've been using GPU-Z for my measurements, so I'm not sure how accurate that is.

Anyway, according to the GPU-Z log file, my freshly re-pasted X800 GTO starts off at 39°C when idle and goes up to 79°C under full load. This is with a 26°C room temperature.

mockingbird wrote on 2023-09-11, 13:56:

Very nice! […]
Show full quote

Joseph_Joestar wrote on 2023-09-11, 09:24:

Sure, here you go: […]
Show full quote

Sure, here you go:

• 640x480x32 - 750.3
• 800x600x32 - 604.9
• 1024x768x32 - 467.4
• 1280x1024x32 - 340.1
• 1600x1200x32 - 247.9
• 1600x1200x32 + 16xAF - 206.7

Very nice!

Interesting, I'll have to remove a stick with my setup and see if that gives me a boost...

A faster GPU will improve the FPS - not that it's required 🤣

Don't worry, you're not missing much with dual channel.

My specs are Wolfdale @ 3Ghz, (9 x 333Mhz, FSB/DRAM ratio 5:6). Benchmark is Q3A 1.32c, DEMO001, 1024x768 with all options set to max.

With single 1GB stick (average of three runs): 512.2 FPS
With two 1GB sticks in dual channel mode: 520.03 FPS

And I can't attribute the slight gain to dual channel mode or perhaps more RAM. So real world - you're fine with single channel. And an X800XT will get you about 40-50 more FPS on this platform, compared to the X800GTO.

Joseph_Joestar wrote on 2023-09-11, 14:29:

I've been using GPU-Z for my measurements, so I'm not sure how accurate that is.

Anyway, according to the GPU-Z log file, my freshly re-pasted X800 GTO starts off at 39°C when idle and goes up to 79°C under full load. This is with a 26°C room temperature.

Thank you! So it's basically in line with what I've been seeing. Your GTO card probably has slightly lower temps due to the lower clocks.
It's insane to me that these 20 year old GPUs run this hot, while my modern RTX 4070Ti stays at ~ 55 - 60 degrees C in full load (well, it's true that it does have an absolutely ginormous cooler which is almost an engineering marvel in and of itself).

mockingbird wrote on 2023-09-11, 15:25:

Don't worry, you're not missing much with dual channel.

Indeed, the memory bandwidth is very high even in single channel, not to mention the fact that the CPU has 6 MB of L2 cache with insane throughput (for the time), which decreases the reliance on memory speed (especially for "puny" workloads such as retro gaming). 😀
Anyway, I haven't tested it, but the performance issues are most likely caused by the 4 GB of RAM, not by the dual channel operation. So 2 x 1GB sticks should work just as well.

bloodem wrote on 2023-09-11, 15:37:

Indeed, the memory bandwidth is very high even in single channel, not to mention the fact that the CPU has 6 MB of L2 cache with insane throughput (for the time), which decreases the reliance on memory speed (especially for "puny" workloads such as retro gaming). 😀
Anyway, I haven't tested it, but the performance issues are most likely caused by the 4 GB of RAM, not by the dual channel operation. So 2 x 1GB sticks should work just as well.

Wow, you're right! With 4GB installed (Win98 shows 3327MB), I take a massive hit. The FPS drops to 368.5 from 520.03!

By the way, I thought I'd experiment a little since (some of?) the 771 Wolfdales are multiplier unlocked (from 6x to 9x in my case being presented in the BIOS). I have some DDR2 1066 RAM, so I ran the FSB at 444mhz (which with a 7x multiplier gave me a slight overclock to 3.1Ghz but more importantly set the memory speed to 1066Mhz instead of 800Mhz). The gains were only slight. I ran DEMO001 twice before the system crashed, and I got 534.05fps, so only around 15fps more. I don't know if the crash was because of some divider being unhappy with that odd FSB or whatnot.

I think with some more voltage I can push this Xeon a lot farther since it's only a 45w TDP part, but that's well beyond the scope of this...

EDIT: At 433x7 (3.03Ghz, 1039Mhz memory frequency), I get 528.7 FPS. So it is memory frequency which yields further improvement on a more linear scale than does CPU frequency with this benchmark. So I imagine with DDR3 1600 you'd get slightly better performance. Now I feel compelled to test with my Ivy Bridge motherboard...

EDIT 2: I tested with an Ivy Bridge @ 3.2Ghz with one 2GB DDR3 1600 stick. The result was 537.5, a very small increase over Wolfdale and not worth it, in light of the fact that it's not Windows 98 friendly like the older platform was, without getting into too much detail. I think Wolfdale and Regor are the most economical options for a very fast Windows 98 system.

Joseph_Joestar wrote on 2023-09-11, 14:29:

Anyway, according to the GPU-Z log file, my freshly re-pasted X800 GTO starts off at 39°C when idle and goes up to 79°C under full load. This is with a 26°C room temperature.

That cooler is just too weak, ATi reference design for Radeon X850 Pro/X800GTO VIVO used copper base and it mostly hovers around 70-72.

The Serpent Rider wrote on 2023-09-11, 17:54:

That cooler is just too weak, ATi reference design for Radeon X850 Pro/X800GTO VIVO used copper base and it mostly hovers around 70-72.

Good to know, and yeah, I didn't see any copper on the bottom side of this oddly designed cooler assembly. It baffles me that this thing actually pushes hot air into the case instead of expelling it out. The PowerColor engineers should have gone with the latter, even if that meant taking up two slots. No idea what they were thinking here.

Also, the airflow in this cheapo Blueberry case is not great, but it's the only spare one that I currently have. I'll upgrade to something more appropriate when time permits.

bloodem wrote on 2023-09-11, 15:37:

Anyway, I haven't tested it, but the performance issues are most likely caused by the 4 GB of RAM, not by the dual channel operation. So 2 x 1GB sticks should work just as well.

My 2x1GB memory kit just arrived and I can confirm the findings made earlier by @mockingbird. It's the 4GB memory amount that lowers performance under Win9x, not dual-channel.

BTW, I noticed maybe a 1% performance increase with this memory kit, but that's probably due to slightly better timings that it uses. The difference is more noticeable in lower resolutions.

Small update: X800GTO VIVO/X800GT02/X850 Pro use cooler with a copper base and 2 heatpipes attached.

The Serpent Rider wrote on 2023-09-23, 15:41:

Small update: X800GTO VIVO/X800GT02/X850 Pro use cooler with a copper base and 2 heatpipes attached.

Looks like PowerColor cheapened out on the cooler, at least compared to what you're showing.

The justification could be that my card is using the R430 chip (110 nm) while most X850 cards use the R480 chip (130 nm). Theoretically, the R430 should be slightly easier to cool, due to the improved production process, but who really knows.