Just added results for a VIA Eden "Nehemiah" 1GHz from an HP thin client t5630.
Numbers were improved a lot with MTRRLFBE, but decided to just post the results of the 100% stock system since I'm not racing anyone 😀 and it's not one of the most common CPUs either

Personal thought #1: I'm impressed, it's not bad at all
Personal thought #2: Need to do some reading on this tongue-twister MTRRLFBE. I had no idea about it until today that it just somehow managed to double 3DBench score.

My Packard Bell has a Pentium 120 processor, but it can be changed by jumpers to 75, 90, 100, 120, or 133 Mhz. Unlocked processor, no buses being overclocked. If I submitted results for all speeds, would you prefer them in Stock or Overclocked?

konc wrote:

Just added results for a VIA Eden "Nehemiah" 1GHz from an HP thin client t5630. Numbers were improved a lot with MTRRLFBE, but d […]
Show full quote

Just added results for a VIA Eden "Nehemiah" 1GHz from an HP thin client t5630.
Numbers were improved a lot with MTRRLFBE, but decided to just post the results of the 100% stock system since I'm not racing anyone 😀 and it's not one of the most common CPUs either

Personal thought #1: I'm impressed, it's not bad at all
Personal thought #2: Need to do some reading on this tongue-twister MTRRLFBE. I had no idea about it until today that it just somehow managed to double 3DBench score.

That is quick, it's pretty equal to my P3 933 on via chipset with savage4 listed just above it & that was with MTRRLFBE, would like to see your results with MTRRLFBE enabled.
MY EPIA C3 1 GHz should be here tomorrow so hopefully have some results from that one soon too.
Best,
Chris

BSA Starfire wrote:

That is quick, it's pretty equal to my P3 933 on via chipset with savage4 listed just above it & that was with MTRRLFBE, would l […]
Show full quote

That is quick, it's pretty equal to my P3 933 on via chipset with savage4 listed just above it & that was with MTRRLFBE, would like to see your results with MTRRLFBE enabled.
MY EPIA C3 1 GHz should be here tomorrow so hopefully have some results from that one soon too.
Best,
Chris

Yes, I was impressed too. Never thought of this little thing as a potential dos-gaming, 100% silent, super small and portable pc. I got it as a backup for another one which runs XP (a very specific application, not everyday computing) so I got the chance to measure it. Having in mind its really poor performance with XP, I wasn't expecting this. Its FPU though scores very low (MMX - K6-2 level), if those old dos benchmarks give correct results.

As for the number with this-thing-I-still-don't-know-what-it-does, easy. I took a note of them, so here you are

1Bench/Stock/MTRRLFBE as the only change
2----------------------------------------
33DBench         415.2   095.8 (obvious rollover here)
4PCP             182.6   217.2
5Doom realticks  564     527
6Quake           123.1   137.3

Just added the 2nd Celeron 333 score to the database. With FASTVID (MTRRLFBE yields identical scores except for a slightly lower 3dbench score).

Added my Satellite Pro 460CDX, TravelMate 313T and Presario 1694 laptops to the list.

Tried using the DOS32A DOS extender for the Doom benchmark and got 2,7% better result, from 837 til 815 realticks. Haven't published any results for this system in the database.

Specs: Athlon 1333MHz, MSI KT133A based motherboard, Radeon 9500, 256MB SDRAM.

Added a few,
Celeron Mendocino 466MHz, with TRIO 3D/2X.
Pentium 166 & 200 mmx with Cirrus Logic laguna 3D (CL-GD 5464).
P3 coppermine 933 on VIA apollo with a 3DFX Voodoo3 2000 AGP.

More added, PCCHIPS M590 with CPU's from a Cyrix/IBM 6x86MX 266 to 333 and AMD-K6-2 300 to K6-III 400.

Great contribution all of you!

What if I match dosbox cycles with the results of Doom and PCBench, can I presume the emulated dosbox machine and the real machine are equivalent?

I found that 55,000 dosbox cycles equal to Pentium-133 average on Phil's Excel list, would you say that it's a "safe bet" or roughly equivalent?

James-F wrote:

What if I match dosbox cycles with the results of Doom and PCBench, can I presume the emulated dosbox machine and the real machine are equivalent?

I found that 55,000 dosbox cycles equal to Pentium-133 average on Phil's Excel list, would you say that it's a "safe bet" or roughly equivalent?

70600.

Roughly, as in DOSBOX each Opcode is one cycle whereas on a real CPU different opcodes take different cycles.

Huh, I wanted to write few words about DOOM test not working, but I modified DOOM.BAT the way that I removed CALL from line 3, and it started working. But surely it is good software!

Also, I tried to share my results, but it seems there is only empty spreadsheet. It is correct?

I'm in the process of setting up a machine for VGA testing. I have a few interesting video cards and I want to see how they compare against each other.
The test machine will likely be an Intel i815E-based board (i815EEA) probaby with a Coppermine-800 or Coppermine-T-1000 P3. I have the following cards to test:

3DFx:
Voodoo2 (single, and in SLI) PCI
Voodoo3-3000 AGP
Voodoo5-5500 AGP

Nvidia:
TNT2 M64 PCI
TNT2 M64 AGP
Geforce2 GTS AGP
Geforce4 Ti4200 AGP
Quadro NVS280 (essentially a FX5200) PCI
Geforce FX5900 AGP
Geforce 7800GS AGP (because why not go overkill? 😀)
Geforce 8400GS PCI

ATI:
Radeon 7500 AGP
Radeon x800 Pro AGP

Originally I had planned to use a Soyo SY-TISU motherboard with a Celeron Tualatin 1.4GHz, which I had in storage for the past 12 years or so. The motherboard was working when I stored it, but now it appears to be dead (doesn't POST, no beeps, no video etc). So I had to scratch that plan and get another socket370 motherboard and CPU off ebay - which are still in the mail. I used a diagnostics card on the Soyo motherboard, and it appears that there's a problem with the reset circuitry. I don't know how to fix that.

Will use DOS 6.22 and Windows 98 SE. Should I try XP as well?

krivulak wrote:

Huh, I wanted to write few words about DOOM test not working, but I modified DOOM.BAT the way that I removed CALL from line 3, and it started working. But surely it is good software!

Also, I tried to share my results, but it seems there is only empty spreadsheet. It is correct?

Make sure you are on the right spreadsheet tab. I noticed sometimes when I load the spreadsheet it starts on an empty tab.

I wonder how you will get 3DBench running on Voodoo2 as it is no VGA card.
Also how comparable is running benches in the Virtual DOS Machine of WinXP?

Usually a clean boot in Real Mode is a good start, so HIMEM at best.

stamasd wrote:

I'm in the process of setting up a machine for VGA testing. I have a few interesting video cards and I want to see how they comp […]
Show full quote

I'm in the process of setting up a machine for VGA testing. I have a few interesting video cards and I want to see how they compare against each other.
The test machine will likely be an Intel i815E-based board (i815EEA) probaby with a Coppermine-800 or Coppermine-T-1000 P3. I have the following cards to test:

3DFx:
Voodoo2 (single, and in SLI) PCI
Voodoo3-3000 AGP
Voodoo5-5500 AGP

Nvidia:
TNT2 M64 PCI
TNT2 M64 AGP
Geforce2 GTS AGP
Geforce4 Ti4200 AGP
Quadro NVS280 (essentially a FX5200) PCI
Geforce FX5900 AGP
Geforce 7800GS AGP (because why not go overkill? 😀)
Geforce 8400GS PCI

ATI:
Radeon 7500 AGP
Radeon x800 Pro AGP

Originally I had planned to use a Soyo SY-TISU motherboard with a Celeron Tualatin 1.4GHz, which I had in storage for the past 12 years or so. The motherboard was working when I stored it, but now it appears to be dead (doesn't POST, no beeps, no video etc). So I had to scratch that plan and get another socket370 motherboard and CPU off ebay - which are still in the mail. I used a diagnostics card on the Soyo motherboard, and it appears that there's a problem with the reset circuitry. I don't know how to fix that.

Will use DOS 6.22 and Windows 98 SE. Should I try XP as well?

This benchmark database is for DOS. Also, the benchmarks are all 2D-based, so Voodoo2 will not run the tests.

clueless1 wrote:

This benchmark database is for DOS. Also, the benchmarks are all 2D-based, so Voodoo2 will not run the tests.

Well I see Phil's test package as a beginning step only. 😀 Will expand afterwards. Also, of course I won't use the V2 by themselves, they will be paired with a 2D card. Not useful for Phil's database, but I have further plans for that setup.

I did a benchmark of all my graphic cards, don't know if it is any useful to the main chart. If anyone want to see, just take a look.

Every one was tested on Intel Celeron FV524RX466 with 512 Mb RAM (no overclocking at all -> base 466 MHz) on motherboard SOYO SY-7VCA.