You're missing Quake. 😀

But yeah, that P100's underperforming somewhere.

Software needs(ed) to be optimized for P5.

Yes. There is nothing wrong. Many simpler benchmarks and games like Doom run faster on a tuned late 486. The Pentium architecture with superscalar execution units, branch prediction, deeper pipelines, split-cache architecture, etc.needs certain complexity of software to show it muscles. P100 is objectively faster, just not at everything.

Thank you, all.

X2 at prev reply, but Any chance the p100 has fake cache? If not. Then try other benchmarks as stated.

Something is probably off with the Pentium machine. I suggest you try other benchmarks.

Also, I believe this motherboard chipset (SIS 5511 or 5513) might be one of those that actually supports running with a single 72-pin. If this is the case and you are running with a single SIMM, this may explain the performance.

Agree darry.

xeiter wrote on 2022-10-23, 20:39:

I have two systems out of which the 486 system outperforms the P100 system in 3dBenchmark. So I wanted to check with the community if it is how it is supposed to be or something is wrong.

You give lots info on the 486 board but nothing on the Pentium board ( no picture, no link) and what version of 3dBenchmark so is hard to say why such a difference.
A good P-100 should out perform a 486DX-100 by nearly 50% unless you compare good to inferior products...
Can you post picture/link to the pentium board AND run cachechk on both and post pictures. something is off but not sure why.. just my opinion.
good comparison of Pentium 66 versus AMD 486-133 under different tests is this: https://www.youtube.com/watch?v=tm5lIQ_47Ko

On a side note, 3dbench is heavily video speed dominated. What do these boards give in landmark or speedsys in this categori with those cards? I suspect the P5 result will be abysmal while, if my memory serves, the EXP8449 had top notch PCI speed for a 486.

https://dependency-injection.com/early-pentium-chipsets/ SiS 5511 is surprisingly not terrible, at least in Abit PG5. But there is potential for 50% performance difference between terrible and excellent P100 board.

Hey, I didnt have photos - sorry about that.

I took some photos now - please see them below. After I assembled the PC back I thought: "Hmm what if I load just bare DOS 6.22 without anything in autoexec.bat and config.sys. So I did that and P100 returned much better results. My autoexec.bat and config.sys are very raw - I just got them off some CD that helped me to get my machines going. I still have lots of work to do with setting things up.

I just started getting back into retro pcs and dont remember many things from 90s. I guess loading stuff on boot like memory managers does affect things and skew test results.

I am now getting 71 for 486 and 91 for P100 in 3dBenchmark 1.0.

Below are the photos - maybe someone can pick something up that I am still missing:

Pentium 100, 32MB, 256 Cache, S3 Trio 64 V+

https://stason.org/TULARC/pc/motherboards/F/F … 6F56-VER-2.html

486dx4-100 , 16MB RAM, 256KB Cache, S3 Trio 64 V+

https://stason.org/TULARC/pc/motherboards/D/D … 86-EXP8449.html

AUTOEXEC.BAT and CONFIG.SYS

Thank you for your help.

Horun wrote on 2022-10-24, 02:03:

Agree darry. […]
Show full quote

Agree darry.

xeiter wrote on 2022-10-23, 20:39:

I have two systems out of which the 486 system outperforms the P100 system in 3dBenchmark. So I wanted to check with the community if it is how it is supposed to be or something is wrong.

You give lots info on the 486 board but nothing on the Pentium board ( no picture, no link) and what version of 3dBenchmark so is hard to say why such a difference.
A good P-100 should out perform a 486DX-100 by nearly 50% unless you compare good to inferior products...
Can you post picture/link to the pentium board AND run cachechk on both and post pictures. something is off but not sure why.. just my opinion.
good comparison of Pentium 66 versus AMD 486-133 under different tests is this: https://www.youtube.com/watch?v=tm5lIQ_47Ko