Like you i have a Radeon HD7870 and 3dmark won't run on it claiming i don't have enough video memory. I guess 2GB of video memory isn't enough for a 2001 3D benchmark 😁

Just set the affinity to cpu0 before you hit the ok button! 😉

Thank you, that worked like a charm!
I got:
-56700+/- #dmarks on my i5 750 (3.8Ghz) with a AMD HD7870;
-28800+/- on a AMD 5600K APU.

Until now i was never able to run 3D mark on these systems due to this error. Still i guess 3Dmark2001 scores are meaningless on newer systems because the 3Dmark score on the 5600K apu is a bit low since i was able to play MaxPayne 3 and BioShock Infinite on it reasonably well.

Has anyone got this to run on the Matrox Millennium G200? It complains of lack of memory. I think you need the little SGRAM snap-in upgrade, which I've never seen for sale anywhere.

@rick6: Yeah, but the 3dmark03 scales pretty damn good! It's veeeery vga dependant, which I like very much. My favorite of all 3dmarks.

Did not know that would work to get past that error.
Tried it on my windows 8.1 machine tho and it refuses to do any of the Max Payne tests.
Crashes every time it hits them. 2001 hates Win 8.1... The performance was very low on the Truck and Dragon tests too I doubt they were running correctly 🤣 .

It's getting 15 years old anyway 😜
But no problems whatsoever here. It really sucks on scaling though, and that's a fact!

RichB93 wrote:

A GTX 580 paired with a Core 2 Duo? Jesus, that must be a hell of a bottleneck.

Hm, then how about a C2D E8300 @ 3.15 (450x7) with 4x2GB Hynix DDR2 @ 900Mhz, paired with Gainward GTS250 512MB? Windows 7. No bottlenecks here?

We don't measure bottlenecks with this benchie! 😜

GeorgeMan wrote:

We don't measure bottlenecks with this benchie! 😜

Yeah, I know, I was joking 😀

Still, I do wonder where the C2D becomes a bottleneck and you need an upgrade. Cause the C2D will live longer than the Athlon in people's cases.

K6-2+ @ 585 MHz, GA-5AX, 128 mb, GF2 MX400, Windows 98SE, no AGP drivers.

Skyscraper wrote:

K6-2+ @ 585 MHz, GA-5AX, 128 mb, GF2 MX400, Windows 98SE, no AGP drivers.

Hm, Skyscraper, can you run same test @ 100 FSB with onboard cache turned on and 512 or 768 RAM?

Your setup puzzles me? What's the point in having only 128MB RAM with that specific CPU? (except for running a 130FSB, if it's only stable with 1 stick of 128MB).

Edit: Why no AGP drivers?

I have this board in revisions 2.2 and 4.1. I planned to use the 2.2 for a long run (it's 100FSB max but that's OK for me).

Space Cowboy wrote:

Hm, Skyscraper, can you run same test @ 100 FSB with onboard cache turned on and 512 or 768 RAM? […]
Show full quote

Skyscraper wrote:

K6-2+ @ 585 MHz, GA-5AX, 128 mb, GF2 MX400, Windows 98SE, no AGP drivers.

Hm, Skyscraper, can you run same test @ 100 FSB with onboard cache turned on and 512 or 768 RAM?

Your setup puzzles me? What's the point in having only 128MB RAM with that specific CPU? (except for running a 130FSB, if it's only stable with 1 stick of 128MB).

Edit: Why no AGP drivers?

I have this board in revisions 2.2 and 4.1. I planned to use the 2.2 for a long run (it's 100FSB max but that's OK for me).

The 128 MB memory is because the board was on the testbench and was also tested with CPUs without on chip L2. The board can only cache 128MB with its bugged Aladdin 5.
I posted a thread here in the Marvin section with a few benchmarks at both 4.5*130 MHz = 585 MHz without on board L3 and at 600 MHz with L3, not 3dmark 2001 but 3dmark 99 which is a bit more relevant. I use 3dmark 2001 for stability testing so I diddnt bother to save a screenshot of the score at 600 Mhz but if I remember correctly it was a little bit lower.

My experience of AGP drivers for socket 7 boards in Windows 98 is not very good...

Well i've tried a few runs on a Pentium 4 560 3.6Ghz (Yeah, this exact one http://www.tomshardware.com/reviews/p4,919.html).

And these are the contestants:

Here you have with the little fan a Geforce 256 32mb SDR, with the big dark heatsink a Geforce2 GTS 64MB DDR, the passively cooled is a Geforce 4 MX440SE 64MB DDR and the big red video card is a ATI HD3850 AGP 512 DDR3.

And here are the scores accordingly, all running under Windows 7:

Yeah this last one is a big disappointment to be honest. Under WinXP i was able to score 17000ish which is still disappointing. The board could be one to blame but if you remember correctly Pentium 4 Prescott were a bit beyond the limit of the Pentium 4 arch. Maybe the hd3850 doesn't scale well with these ancient toasters from 2003.

rick6 wrote:

Yeah this last one is a big disappointment to be honest. Under WinXP i was able to score 17000ish which is still disappointing. The board could be one to blame but if you remember correctly Pentium 4 Prescott were a bit beyond the limit of the Pentium 4 arch. Maybe the hd3850 doesn't scale well with these ancient toasters from 2003.

Hm, if I'm not wrong, I had far better results from a GF6800LE (unlocked to 12 pipes) and overclocked Athlon XP. (17-18K)

There's something bugged with the card, not the chipset. There's something generally wrong with that PCI-E to AGP bridges (whatever they are) ...

I remember I had an HD 2600XT before, and was not really happy with it. Makes sense for HD movies only 😀

A little while back I put my Dell E310 system (Pentium 4, 512MB DDR2 533 etc) to the test with its onboard Intel GMA 900 graphics. In 3DMark 01 It scored pretty badly, but then I put in an FX 5200 PCI and its score was close to the GMA 900!. It's surprising how crappy the FX 5200's are! The PCI bus might have something to do with it but I still hate that card.

I need to hook it up again and upload the results, maybe later on 😀

Space Cowboy wrote:

Hm, if I'm not wrong, I had far better results from a GF6800LE (unlocked to 12 pipes) and overclocked Athlon XP. (17-18K)

There's something bugged with the card, not the chipset. There's something generally wrong with that PCI-E to AGP bridges

I guess the card is fine, i've had it for a while on a Athlon 3700+ and had some really nice scores, above 24000 something if i'm right. I just might have some sort of issues between the card and the VIA chipset from this ASrock board.

I'm going to try a Asus V9999 Geforce 6800 video card and see how it does.

Space Cowboy wrote:

Hm, if I'm not wrong, I had far better results from a GF6800LE (unlocked to 12 pipes) and overclocked Athlon XP. (17-18K)

Alright i think i figured what was going on more or less. The board always made a annoying coilwhine wich i was able to dampen a bit with hot glue, but it never went completly away. Today i noticed that when running 3Dmark each time i heard a coilwhine on the board the fps dropped considerably from 250 for example to 100 and then to 300 and them to even less than 90. So after reading that article from tomshardware i decided to turn off cpu throtling option in the bios and then the score jumped from those 14000 3dmarks to 17000ish. Not to great either but i believe that if i was running under xp the score would be even better.
I've tried 3dmark 2003 and i scored 23000 so i guess everything is pretty much in line to what it should be. Oh and the coilwhine went away after disabling cpu throtling!

If only i was able to figure why each time i unplug the computer from the wall i lose all my bios settings..even with new cmos battery.
EDIT: just got it fixed Ever had this annoying problem? Bios setting are lost even after replacing cmos baterry!

Pentium 4 2.4GHz Northwood, Soyo P4I-845PE, 512mb DDR-333 ram, NVIDIA GeForce 3 Ti 200 64mb AGP (30.82 Win9x drivers), DirectX 8.1b, Windows 98SE

Not bad for the Dell OEM GF3. Not bad at all 😎

Pentium 4 2.4GHz Northwood, Soyo P4I-845PE, 512mb DDR-333 ram, NVIDIA GeForce FX5500 256mb AGP (82.69 Win9x beta drivers), DirectX 8.1b, Windows 98SE

Now this I didn't expect. Looks like the FX5500 is indeed a dog - even the GF3 beats it handily 😦

Hm, interesting. I just did the same test with also a Geforce3 ti200 and a Geforce FX5500 on a Pentium 4 1.7Ghz sk423 and the scores were 6200\6300 for the for the Geforce3 ti200 and 5800/5900 for the Geforce FX5500. To think that that's a Geforce 5 midrange (?) being beaten by a low end Geforce 3... The Geforce Fx series bellow FX5600 seem to be really weak when compared to previous generations.