One more for the list:

Stock Core i5 760 - 12.667s.

k6-2+ @ 133mhz(66*2), gigabyte 5ax 17:30, no match for pentium anyways.

Just benchmarked my 3dfx-Glide machine that I built a while ago:

Pentium II-400 (4x100), Gigabyte GA-6BXE (Intel 440BX), 512MB (2x256) PC-100 CAS2
6m 20.434s

I'm honestly not too satisfied with the results, seeing as jwt27's PII-350 was about 40 seconds faster.

Svenne wrote:

Just benchmarked my 3dfx-Glide machine that I built a while ago: […]
Show full quote

Just benchmarked my 3dfx-Glide machine that I built a while ago:

Pentium II-400 (4x100), Gigabyte GA-6BXE (Intel 440BX), 512MB (2x256) PC-100 CAS2
6m 20.434s

I'm honestly not too satisfied with the results, seeing as jwt27's PII-350 was about 40 seconds faster.

What OS are you using? Apparently using 9x results in about 25% slower results or so then when using an NT OS

Maybe we should split the results up?

Tetrium wrote:

What OS are you using? Apparently using 9x results in about 25% slower results or so then when using an NT OS

Maybe we should split the results up?

thats not essentially the case. win9x results may be less consistant than nt versions, but the performance loss should be no more than 1~2% if you can get it running correctly.
1. optimize your bios settings for performance
2. clean up your windows and load least auto-running programs possible, and don't run any other programs when running superpi
3. for win9x, run some high load applications such as 3dmark or quake to warm it up before running superpi.

Another one:

Stock Pentium D 930 - 42.844s

ASUS P/I-P55T2P4 i430HX 512Kb 64Mb EDO w98se
Pentium 233MMX @262Mhz (75x3.5)
12m. 05.075s.

Dual Pentium 3 Xeon 550/1Mb SL3CE

3m 48.672s

Will retest when I get my Asus XG-DLS board and my dual 700/1Mb Xeons running.

PIII Coppermine 1000/256/133 @ 1127MHz / Tekram P6BX-An 440BX / 768MB PC133 @ 150FSB / WinXP

2m 11.278s

Just for shits and giggles, I ran this with some other stuff open, etc.

AMD A8-3500m @1.5ghz, 8gb ram, windows 7x64

41.499s at 1M

During normal day work on my desktop-machine:

12.558s 1M
(C2D E8500 @3,8 GHz (400x9.5), DDR2-800 4-4-4-12 2T)

ASUS P/I-P55T2P4 i430HX 512Kb 64Mb EDO w98se
Pentium 233MMX @262Mhz (75x3.5)
😮 Improved to 10m 48029s. !
😅 😅

k6-2+ @ 550 (100*5.5), mvp3 board with 2mb cache:
l3 cache on 5:19
l3 cache off 6:36

Results from my main rig:

TELVM wrote:

PIII Coppermine 1000/256/133 @ 1127MHz / Tekram P6BX-An 440BX / 768MB PC133 @ 150FSB / WinXP

2m 11.278s

Same system with PIII Tualatin 1400S/512/133 @ 1470MHz / 140FSB:

0.1m 42.387s

On my everyday computer : 10.873s (core i7 950 @ 3.83GHz, 166MHz bus, CAS 8 DDR3)

Cool, haven't been here in a couple months and this thread is still going! If I got time in the next couple days, I'll do a test on my new undervolted mobile-A64 3000+ @1GHz 0.875v.

How important are the FSB and the memory for SuperPI? On a 430TX-based board I got around 12m22s with a K6-2/400 (compared to a littler over 9m with an MVP3-based board in the original post)...

One difference is the FSB (my CPU runs 6*66), and then my Win98SE might be running outside cacheable RAM (I have 128MB on my system).

I'm still running my dual 550/1MB P3 Xeons for the moment, until I can swap their heatsinks onto my 700/1MBs, but I am curious to see if there is any difference between the SuperPi run I executed while running my Intel MS440GX board, and the ASUS XG-DLS board I'm now running. Might do another run before I move to the faster processors just to see if there is. Both boards are based on the 440GX chipset, so I'm not really expecting much difference between the two, but who knows?

iulianv wrote:

How important are the FSB and the memory for SuperPI? On a 430TX-based board I got around 12m22s with a K6-2/400 (compared to a littler over 9m with an MVP3-based board in the original post)...

One difference is the FSB (my CPU runs 6*66), and then my Win98SE might be running outside cacheable RAM (I have 128MB on my system).

FSB matters much, but running outside cacheable RAM matters MORE.

My Pentium MMX at 262Mhz and 75Mhz FSB (430HX) gives 10m 37.605s [latest score]. Taking into account that EDO is generally slower than SDRAM, your system's performance is hugely affected by running outside cacheable RAM.