vetz, could you provide a photographic example of the visual difference between fullscreen with and without UI? From running the two case scenerios, I did not notice any difference, but perhaps I do not know what to look for.

I am running under pure DOS 6.0

You are right. I had sound on.
I launched it as quake -nosound and it minimized the discrepancy
I am surprised the sound kills performance so much.

Now I have:

Normal Quake: 10.8 FPS
Phil's Quake 10.9 FPS

smeezekitty, thank you for checking this. I wonder what is going on with my performance issues when using Phil's Quake. The problem only manifests itself with Quake, and not any of the other benchmarks.

EDIT: What motherboard, chipset, CPU, and graphics card are you using?

I would compare the config.cfg in the id1 directory between the two versions to see what differences there are
--edit---
Same system I have been using for my 486 tests

Acer AP43 with SiS 496/497
AMD 486DX-120
S3 ViRGE 4MB

OK, it seems Vetz was correct. Phil's bench removes the bar at the bottom of the screen which shows how much life you have left, bullets, etc., whereas the default setting is to leave that status bar in place.

I do not understand why smeezekitty's results are unaffected by the two benchmark scenerios though.

vetz wrote:

If I remember correctly Phil's benchmark run at fullscreen without UI, while default settings have the UI enabled. This explains the 15% performance difference.

According to Phil in an earlier post here on Vogons were I pointed this out it was set to fullscreen to be more heavy on faster machines (same applies to Doom).

Of course, now I remember 😀

The benchmarks come as a large ZIP with a menu system ready to download. Just stick with that.

I added my Cyrix 5x86-133/4x system to the database using Phil's VGA batch file.

PCPBench = 23.6
3Dbench = 83.9
Quake = 15.8
Doom = 49.3 (1515)

SUM = 88.7

For a non-overclocked system, it is neck-and-neck with Vetz's POD83.

I also added an IBM 5x86-100.

feipoa wrote:

I added my Cyrix 5x86-133/4x system to the database using Phil's VGA batch file. […]
Show full quote

I added my Cyrix 5x86-133/4x system to the database using Phil's VGA batch file.

PCPBench = 23.6
3Dbench = 83.9
Quake = 15.8
Doom = 49.3 (1515)

SUM = 88.7

For a non-overclocked system, it is neck-and-neck with Vetz's POD83.

That's a pretty good score. You don't own a S3 Vision PCI card? Would be interesting to see how the system performed with the same graphics adapter as my. G200 PCI is a tad quicker than the S3 Vision864 VLB card. The only reason I get a higher total score is because of the FPU performance of the Pentium chip in Quake.

Actually, I was using a Voodoo for this particular test.

I do not own any S3 Visions. The S3 cards I own are:

Trio 3D (4 MB) - 86C365
ViRGE DX (4 MB) - 86C375
Trio64V2/DX (2 MB) - 86C775

EDIT: So the closest card to the Vision would be the Trio64V2/DX. I ran the tests with this card, however the results were identical, with the exception of DOOM, which scored 0.6 fps less when using the Trio64V2/DX.

OK. I had inadvertently reset the Phil's bench version of quake.
After replacing the config file with the one from the archive (going back to the status-less screen)

I have:
Normal Quake: 10.8 FPS
Phil's Quake: 9.7 FPS

Added a K5 I bought recently. Tried to add IBM 6x86MX PR233, but it was not stable on P55T2P4 board (430HX). At 2.5x75MHz it could not boot into DOS and hanged randomy in BIOS, at 2x66MHz it could boot windows 98 (still random hangs), yet none of the benchmarks in DOS worked 🙁 Is it a bad chip? Or, perhaps, the board does not like it? All the other parts are fine, working well with P233MMX at 3.5x75MHz.

feipoa wrote:

The opposite effect - very strange indeed. Are you running the benchmarks from within DOS or within Windows' DOS prompt?

Is it possible that you have sound enabled? Try removing the sound card or disabling sound in Quake (there is some command for this, but I don't recall right now).

I think you can just set the BLASTER env variable to blank as it looks for that on startup.

Does any version of doom work for benching?
I have Final Doom CD, not the original release, also do I patch quake?
I wanta give this a go on my Cyrix pr200 and p120.

Ok Final Doom is DoomII just went through its jewel manual... I did not know this 😒 .
I'm not a serious doom fan 🤣.
Does the shareware version of doom 1 work?

If you want to contribute to the database, there is a link on the 1st post that has everything you need in one neat package. 😉

Oh thanks, 🤣 I shoulda checked what was inside of that download first!.
Thought it was just the batch files to do the time demos!

Just type bench from a dos prompt, it's that easy 😜

Yes I tried to make it as easy as possible.

Video instructions here: http://youtu.be/AhCQ0VnLdlI

Packard Bell C110
DOS 6.22\Windows 3.11
Pentium 120, 64mb 60ns EDO. S3 VIRGE\DX 4MB
3DBENCH = 108fps
PCPBENCH = 27.2
DOOM = 52.97
Quake = 25fps

MTRRLFBE doesn't seem to do anything with my Virge, is it supposed to?

(Deleted. No longer relevant anyway)

Oh I wondered what it was for, it gives me an error LFB ADDRESS: F8000000h MTRR failed -8 and not supported.
Never used it before, is it a substantial performance increase on a P2?