Jasin Natael wrote on 2021-11-15, 16:34:

That memory bandwidth increase is massive.

Yes indeed it is! I figured that's where all the performance gains came from.

I'm going to try running that CPU at 83mhz fsb at some point, would be cool if it's stable near the 500mhz mark but I don't expect it to be.

pete8475 wrote on 2021-11-15, 22:25:

Jasin Natael wrote on 2021-11-15, 16:34:

That memory bandwidth increase is massive.

Yes indeed it is! I figured that's where all the performance gains came from.

I'm going to try running that CPU at 83mhz fsb at some point, would be cool if it's stable near the 500mhz mark but I don't expect it to be.

I've run numerous K6-2+ and K6-3+ 450s @ 500 MHz in TXP4s and P55T2P4s over the years and have never had an issue. Might have to bump the voltage to 2.1 at worst.

Repo Man11 wrote on 2021-11-15, 22:53:

I've run numerous K6-2+ and K6-3+ 450s @ 500 MHz in TXP4s and P55T2P4s over the years and have never had an issue. Might have to bump the voltage to 2.1 at worst.

Awesome!

I'll give it a go after I eat dinner tonight.

EDIT - It seems to be working at 498mhz! On post the BIOS reports it's a K6-2+/500. I'll run the same benchmarks to see how much faster it is over the next bit. Hopefully it's stable!

TXP4 128MB PC100 CL2 SDRAM 500MHZ CPU (6x83MHZ FSB)
3dmark 99 - 2446, 6435 cpumarks
3dmark 2k - 2199, 148 cpumarks
3dmark 2001SE - 1135
Quake timedemo - 84.1 fps
Quake 3 (MX4000) - 30.6 fps (borderline playable territory, never thought I'd see that)
Topbench - 331
3dbench 1.0c - 255.1
Chris's 3d Benchmark - 507.8 (304.7 fps)
Chris's 3d Benchmark 640x480 - 116 (79.9 fps)
Speedsys 4.78 Processor Benchmark - 567.29
Speedsys 4.78 Memory Bandwidth - 239.70 MB/s

CPU-Z reports it's actually 83.5FSB and the CPU is running at 501mhz.

EDIT - Quake 3 crashes at 2.0v but runs repeated timedemos with no issue at 2.1v

pete8475 wrote on 2021-11-16, 00:37:

TXP4 128MB PC100 CL2 SDRAM 500MHZ CPU (6x83MHZ FSB) 3dmark 99 - 2446, 6435 cpumarks 3dmark 2k - 2199, 148 cpumarks 3dmark 2001SE […]
Show full quote

TXP4 128MB PC100 CL2 SDRAM 500MHZ CPU (6x83MHZ FSB)
3dmark 99 - 2446, 6435 cpumarks
3dmark 2k - 2199, 148 cpumarks
3dmark 2001SE - 1135
Quake timedemo - 84.1 fps
Quake 3 (MX4000) - 30.6 fps (borderline playable territory, never thought I'd see that)
Topbench - 331
3dbench 1.0c - 255.1
Chris's 3d Benchmark - 507.8 (304.7 fps)
Chris's 3d Benchmark 640x480 - 116 (79.9 fps)
Speedsys 4.78 Processor Benchmark - 567.29
Speedsys 4.78 Memory Bandwidth - 239.70 MB/s

CPU-Z reports it's actually 83.5FSB and the CPU is running at 501mhz.

EDIT - Quake 3 crashes at 2.0v but runs repeated timedemos with no issue at 2.1v

Comparing your scores to mine are interesting. ....
You are getting a higher CPU score in Speedsys, yet mine is higher in both Topbench and 3DMark 99
Your memory bandwidth is higher and so is your high res Chris' 3DBench score, yet my low res score is higher. (I'm guessing that is the video card)
My Quake score is a bit faster but your system seems overall a bit faster than mine.
Pretty cool stuff to compare.

My scores:
A few benchmarks are @ 83x6=500mhz

Superscape 1.0 Low Res: Comes back 0.0 but runs super fast?
Superscape 1.0 High Res: 587.5
Chris' 3D Bench Low Res: 627.6 (376.5FPS)
Chris' 3D Bench High Res: 116 (69.6FPS)
PC Player Bench Low Res: 162.3FPS
PC Player Bench High Res: 41.2FPS
DOOM Minimum: 2134 in 140 realtics
DOOM Maximum: 2134 in 510 realtics
Quake Low Res: 88.5
Quake Medium Res: 38.0
Quake High Res: 22.2FPS
Topbench: 519
Speedsys CPU: 564.46
Speedsys Memory Bandwidth - 186.69 MB/s
Speedsys Throughput - 122.57MB/s

Windows benchmarks:

3DMark 99 Max Default Run 800x600 16bit:
2750 3D Marks
6501 CPU 3D Marks

Edit: These benches were ran before I found that I had a driver issue with a soundcard. I know it was effecting my score in Unreal but not sure if it was anywhere else. I really need to run them again....but general idea.

Jasin Natael wrote on 2021-11-16, 18:35:

Comparing your scores to mine are interesting. .... You are getting a higher CPU score in Speedsys, yet mine is higher in both T […]
Show full quote

Comparing your scores to mine are interesting. ....
You are getting a higher CPU score in Speedsys, yet mine is higher in both Topbench and 3DMark 99
Your memory bandwidth is higher and so is your high res Chris' 3DBench score, yet my low res score is higher. (I'm guessing that is the video card)
My Quake score is a bit faster but your system seems overall a bit faster than mine.
Pretty cool stuff to compare.

Yes fun stuff for sure! I'll run those PC Player, Doom and other Quake benchmarks in a few minutes.

I haven't bothered with the PC Player one in a while because I get some kind of graphics corruption on the screen seemingly regardless of what graphics card I have in there.

EDIT - Results:
PC Player 320x200 - 145.3 fps
PC Player 640x480 - 41.0 fps (flickering at top and bottom of screen where the text is but the score is perfectly legible)
Doom low detail - 2134 in 215 realtics
Doom high detail - 2134 in 1049 realtics
Quake 360x480 - 35.0 fps
Quake 640x480 - 22.6 fps

Also about Topbench I can consistently get it to report a score of 451 if I reboot into DOS from 98SE but if I boot straight to DOS it scores 331. No idea why.

pete8475 wrote on 2021-11-16, 23:56:

Yes fun stuff for sure! I'll run those PC Player, Doom and other Quake benchmarks in a few minutes. […]
Show full quote

Jasin Natael wrote on 2021-11-16, 18:35:

Comparing your scores to mine are interesting. .... You are getting a higher CPU score in Speedsys, yet mine is higher in both T […]
Show full quote

Comparing your scores to mine are interesting. ....
You are getting a higher CPU score in Speedsys, yet mine is higher in both Topbench and 3DMark 99
Your memory bandwidth is higher and so is your high res Chris' 3DBench score, yet my low res score is higher. (I'm guessing that is the video card)
My Quake score is a bit faster but your system seems overall a bit faster than mine.
Pretty cool stuff to compare.

Yes fun stuff for sure! I'll run those PC Player, Doom and other Quake benchmarks in a few minutes.

I haven't bothered with the PC Player one in a while because I get some kind of graphics corruption on the screen seemingly regardless of what graphics card I have in there.

EDIT - Results:
PC Player 320x200 - 145.3 fps
PC Player 640x480 - 41.0 fps (flickering at top and bottom of screen where the text is but the score is perfectly legible)
Doom low detail - 2134 in 215 realtics
Doom high detail - 2134 in 1049 realtics
Quake 360x480 - 35.0 fps
Quake 640x480 - 22.6 fps

Also about Topbench I can consistently get it to report a score of 451 if I reboot into DOS from 98SE but if I boot straight to DOS it scores 331. No idea why.

Thanks for running those!
That is interesting about the Topbench, usually it is the other way around with DOS stuff.
I would say that our machines are pretty well matched, other than the video card and your much better memory bandwidth they seem to be on par with one another.

I thought I'd run some more test on my TXP4, and I found the memory bandwidth was noticeably lower when the motherboard cache was enabled when measured with SiSsoft Sandra 99. I wondered if it was an oddity with that software, so I tried Speedsys which confirmed it the difference. 187 MB/S with the motherboard cache disabled, 149 with the cache enabled.

pete8475 wrote on 2021-11-16, 00:37:

TXP4 128MB PC100 CL2 SDRAM 500MHZ CPU (6x83MHZ FSB) 3dmark 99 - 2446, 6435 cpumarks 3dmark 2k - 2199, 148 cpumarks 3dmark 2001SE […]
Show full quote

TXP4 128MB PC100 CL2 SDRAM 500MHZ CPU (6x83MHZ FSB)
3dmark 99 - 2446, 6435 cpumarks
3dmark 2k - 2199, 148 cpumarks
3dmark 2001SE - 1135
Quake timedemo - 84.1 fps
Quake 3 (MX4000) - 30.6 fps (borderline playable territory, never thought I'd see that)
Topbench - 331
3dbench 1.0c - 255.1
Chris's 3d Benchmark - 507.8 (304.7 fps)
Chris's 3d Benchmark 640x480 - 116 (79.9 fps)
Speedsys 4.78 Processor Benchmark - 567.29
Speedsys 4.78 Memory Bandwidth - 239.70 MB/s

CPU-Z reports it's actually 83.5FSB and the CPU is running at 501mhz.

EDIT - Quake 3 crashes at 2.0v but runs repeated timedemos with no issue at 2.1v

Could I ask you what your memory settings are? Your memory bandwidth is noticeably higher than mine. Edit: I realized I was mixing up the memory throughput and memory bandwidth.

Repo Man11 wrote on 2021-12-20, 22:51:

Could I ask you what your memory settings are? Your memory bandwidth is noticeably higher than mine. Edit: I realized I was mixing up the memory throughput and memory bandwidth.

I'll have tomorrow night, that PC is partly taken apart right now I was trying to get an MX4000 with DVI to work in it but had no luck so I started to replace the board with a Shuttle super socket 7 board that has AGP.

That said it w0n't take me long and I can take a pic with my phone of the settings, I'm just busy playing Skyrim on here right now and being kind of lazy. 😁

Pic as promised, as far as I can recall the only setting I actually changed is setting the CAS latency to 2T.

pete8475 wrote on 2021-12-21, 22:22:

Pic as promised, as far as I can recall the only setting I actually changed is setting the CAS latency to 2T.

Thank you.

Repo Man11 wrote on 2021-12-21, 22:46:

pete8475 wrote on 2021-12-21, 22:22:

Pic as promised, as far as I can recall the only setting I actually changed is setting the CAS latency to 2T.

Thank you.

No problem, I also noticed in your post you were actually looking at memory throughput so I'm running Speedsys right now and I'll screenshot the results in a moment.

EDIT - pic attached, seems to get 131.28 MB/S. The memory bandwidth number is lower than what it used to be, no idea why but this board is about to be replaced anyway so who cares.