Can't remember if I posted my PS1/pro results , here they are:

486slc2/40mhz IBM PS1/pro model 2123 ISA CL-5429 1MB: 5.7 without the 486slc2 1kb cache enabled
486slc2/40mhz IBM PS1/pro model 2123 ISA CL-5429 1MB: 13.6 with the 486slc2 1kb cache enabled

I have no idea about chipset, timings, cache. The machine came with a 386sx/20 and I use a clip-in 486slc2 at 40mhz (20x2)

Tried with 133MHz FSB yesterday, here's the result:

Intel - Pentium_II - 133 - 3.5 - 467 - MSI - MS_6119 - 440BX - 512k - AGP - 3dfx_Voodoo3_2000 - 16MB - jwt27 - 717.4

Intel-486DX2@66-33-2-66-Jetway-J403TG V2-UMC-256KB-VLB-S3 Vision864-2MB-aacheron-45.4
Intel-486DX2@66-33-2-66-Jetway-J403TG V2-UMC-256KB-VLB-S3 805p-2MB-aacheron-45.4
Intel-486DX2@66-33-2-66-Jetway-J403TG V2-UMC-256KB-VLB-Trident TGUI9440AG-1MB-aacheron-45.4

something strange here, same result with different VGAs, may be VLB bus freq. is limiting VGAs speed... I will try to find correct jumper on mobo... Most mobos has a 33MHz bus limit for VLB, is it?

Updated V3 bios today, 3DBENCH score increased from 539.9 to 570.7 (and 717.4 to 759.4 with FSB at 133) 😀

Hi, shouldn't this thread be stickied? I find it really useful.

Because I was lazy I ran 3Dbench2 on a Desktop Sandy Bridge Pentium G620 2,6 GHz with 1333MHz memory CL9,9,9. Radeon 6450. In pure DOS 7.10.
Without fastvid about 350, with fastvid about 720.
Surely it means nothing. no way to see a 4th digit overflow? 😜 I prefer PC Player bench for any pentium. Which again is very dependent on FSB-speed and MTRR settings.

I just made a spreadsheet for these results on Google Docs. So far I only added the 386 and P2 results from the start post (using OCR, so don't mind the spelling errors), but hey, it's a start!

Add your results to the bottom of the list, and when you enter the last column, the list is sorted automagically 😀 Here's the link:
https://docs.google.com/spreadsheet/ccc?key=0 … R09TNUlWdXlLYXc

Wasn't sure what was best for a Winchip so....

3dbench gives me DX4-100 performance:

IDT-Winchip-66-3-200-Gigabyte-GA5AX-Ali-512kb-PCI-Cirrus Logic-unknown-ratfink-66.6

3dbench2 gives me faster-than-MMX-233:

IDT-Winchip-66-3-200-Gigabyte-GA5AX-Ali-512kb-PCI-Cirrus Logic-unknown-ratfink-150.7

Just tested my two DOS systems, here's the results:

Intel-486DX2-33-2-66-Unknown-4HL3S-Unknown-256kB-VLB-S3 805-1MB-CapnCrunch53-43.4

Intel-386DX-25-1-25-Unknown-Unknown-Unknown-128kB-ISA-TSENG ET4000-1MB-CapnCrunch53-8.4

Don't know the maker of my 486 board but that's the model #. Seems pretty close to aacheron's results with the same CPU, amount of cache, and his S3 805p.

The 386 board I don't know anything about; it has no make or model on it, and all I know is that it was apparently used for data collection with a CNC machine. PCCONFIG tells me it has 128KB cache though.

Any thought on grabbing the 3dBench results from the Ultimate 486 Benchmark Comparison? There's quite a few results contained in there which aren't published here. All the information required should be in the pdf.

I apologise for not looking after this project more than I did. I was pretty good initially but then life does its thing and you don't have the time anymore 😒

So I put my latest copy of the spreadsheet on Google docs. I hope I got the permissions right. Anyone with this link below should be able to edit the spreadsheet!

https://docs.google.com/spreadsheet/ccc?key=0 … QjdBV2ZqWnVzQkE

I have made about 25 entries to the 3dBench 1.0 section and 4 entries to the 1.0c section. I'm not sure where the magic sort button is, but you'll want to press it.

The sorting works just like in any spreadsheet. You highlight the data and then tell it to sort by a certain column (benchmarks score).

I will sort it out later 😀

I sorted it 😀 Not super-familiar with Google docs, and also I don't want to mess it up, but there should be a way to have it automatically sort new entries. I could do it in a jiffy in Excel, just haven't messed around with Google that much yet.

Thanks for putting that up! It's nice to look through all the scores and see how others' systems perform.

Google docs is quite simple but it's great with permissions and quickly letting people change things. Microsofts online apps are more advanced but I wasn't sure if people could make changes without signing in...

I find the results varies quite a bit with different graphics card.

These are with P-233MMX @ 292MHz (3.5x83) @ 2.8V

ATI Rage Pro 4MB PCI... 185.0
S3 Virge 86c325 2MB PCI... 205.6
3DFX Voodoo 3 3000 16MB PCI... 206.8
Nvidia TNT2 M64 16MB PCI... 218.9

Motherboard: QDI VX Explorer IV, Triton 82430VX, L2 Cache 512KB, CAS 2T, RAS 3T.

I'm also wondering what is with Super7 ALI5 chipset. Compating P-233MMX in the spreadsheet with my on VX chipset that gets 175.0 with TNT2 M64. 😕

cy

PCI vs AGP. AGP is faster that explains it 😀

Explains what?

Comparing P-233MMX @ 233MHz VX : Ali5 - my system has higher score with older VX chipset and a PCI card. I wanted to get a Ali5 board to get higher FSB for OC, but I guess in the end its not much difference for Pentiums.

Mau1wurf1977 wrote:

PCI vs AGP. AGP is faster that explains it 😀

AGP is only faster when the PCI version of the same card saturates the PCI bus and not a lot of cards did that yet in the time period of the hardware being tested. Super 7 boards also had issues that degraded AGP performance so in many cases PCI cards would be preferred over AGP.

kixs wrote:

Explains what?

Comparing P-233MMX @ 233MHz VX : Ali5 - my system has higher score with older VX chipset and a PCI card. I wanted to get a Ali5 board to get higher FSB for OC, but I guess in the end its not much difference for Pentiums.

183 when I did this video:

http://youtu.be/milD2ZAyZAY?t=5m6s

229 for 100x2.5 setting.