I added a Pentium III 933Mhz and Pentium II 400Mhz to the spreadsheet. Here's the Pentium II result:

17-Zip (A)  4.58 beta  Copyright (c) 1999-2008 Igor Pavlov  2008-05-05
2p7zip Version 4.58 (locale=C,Utf16=off,HugeFiles=off,1 CPU)
3
4RAM size:     128 MB,  # CPU hardware threads:   1
5RAM usage:      6 MB,  # Benchmark threads:      1
6
7Dict        Compressing          |        Decompressing
8      Speed Usage    R/U Rating  |    Speed Usage    R/U Rating
9       KB/s     %   MIPS   MIPS  |     KB/s     %   MIPS   MIPS
10
1118:     278   100    245    247  |     2673   100    227    227
12----------------------------------------------------------------
13Avr:          100    245    247               100    227    227
14Tot:          100    237    237

I've still got a K6-2, Celeron 333 and Pentium II 333 to test.

I might test some computers when i have time, i have been busy lately. I've got here an old 486 SX 25 and a old Pentium MMX Packard Bell to test.

I have the feeling some CPUs like the Pentium Pro will underperform due to no begin optimized for 16 bit code

Tried DOS version in some PCs, i tried on my HP DC7800 which i used to write the floppy and my AMD K6-2 build, also tried on my 486-SX 25, but it crashed, i guess 7 Zip needs FPU acording to the errors

386DX @40Mhz

17-Zip (A)  4.58 beta  Copyright (c) 1999-2008 Igor Pavlov  2008-05-05
2p7zip Version 4.58 (locale=C,Utf16=off,HugeFiles=off,1 CPU)
3
4RAM size:     128 MB,  # CPU hardware threads:   1
5RAM usage:      6 MB,  # Benchmark threads:      1
6
7Dict        Compressing          |        Decompressing
8      Speed Usage    R/U Rating  |    Speed Usage    R/U Rating
9       KB/s     %   MIPS   MIPS  |     KB/s     %   MIPS   MIPS
10
1118:       6   100      5      5  |       75   100      6      6
12----------------------------------------------------------------
13Avr:          100      5      5               100      6      6
14Tot:          100      6      6

486 DX2 @66Mhz

17-Zip (A)  4.58 beta  Copyright (c) 1999-2008 Igor Pavlov  2008-05-05
2p7zip Version 4.58 (locale=C,Utf16=off,HugeFiles=off,1 CPU)
3
4RAM size:     128 MB,  # CPU hardware threads:   1
5RAM usage:      6 MB,  # Benchmark threads:      1
6
7Dict        Compressing          |        Decompressing
8      Speed Usage    R/U Rating  |    Speed Usage    R/U Rating
9       KB/s     %   MIPS   MIPS  |     KB/s     %   MIPS   MIPS
10
1118:      19   100     17     17  |      249   100     21     21
12----------------------------------------------------------------
13Avr:          100     17     17               100     21     21
14Tot:          100     19     19

Added some more computers, somehow, my IBM Thinkpad A21m got an incredibly high score compared to my other P3s in 7-Zip DOS, it also even outperformed my P4 based Celeron 2.6 GHz despite the huge clock disavantage, idk if it was something on the CPU, BIOS or the motherboard/chipset (i know this laptop uses i440BX as chipset)

I'm in the process of adding every official K6-2 speed except for the 533 @ 97FSB. Also added an underclocked K6-2 @ 166 for comparisons to other cpus at the same clock. Motherboard is SiS 530 chipset with 512KB L2. I noticed Carlos K6-2 @ 500Mhz is faster than mine @ 550. He's on the MVP3 chipset with 1024KB L2, so that explains it. Looking at the compress and decompress scores, that validates it. My 550 has a higher decompress score, which depends on integer performance. His 500 has a higher compress score, which depends on memory speed and latency.

Also adding Pentium II from 166 to 333. I'm the lucky owner of an unlocked PII-333. 😉 Tested on a 440LX chipset with 512KB L2.

Thank you for adding your results, guys.

http://www.bttr-software.de/forum/board_entry.php?id=14569

I have a PII 400Mhz however when I ran the test in DOS mode it hangs. It works fine from dos from within Win95 but I do not trust those results. I'll try getting into DOS before getting into Win95.

DosFreak wrote:

http://www.bttr-software.de/forum/board_entry.php?id=14569

Thanks! I'll check it out. It would suck to have to run our benchmarks all over again, though. 😒

vladstamate wrote:

I have a PII 400Mhz however when I ran the test in DOS mode it hangs. It works fine from dos from within Win95 but I do not trust those results. I'll try getting into DOS before getting into Win95.

I only tested under DOS 6.22. It works from a DOS 6.x boot disk as well if you don't want to install a new DOS. Plus, see DosFreak's post above. He may have found a more recent 7-Zip that will work in DOS. If that's the case, we might end up starting over.

I tested the recompiled-for-DOS version 16.02 of 7-Zip from the link DosFreak posted. It works perfectly in DOS but does give significantly different (higher) results. If we decide to switch to it, we will have to dump the results we've made so far and start over. Carlos, if you have time, would you test out version 16.02 and provide feedback? I'm attaching to this post.

Thanks

clueless1 wrote:

I tested the recompiled-for-DOS version 16.02 of 7-Zip from the link DosFreak posted. It works perfectly in DOS but does give significantly different (higher) results. If we decide to switch to it, we will have to dump the results we've made so far and start over. Carlos, if you have time, would you test out version 16.02 and provide feedback? I'm attaching to this post.

Thanks

Idk if it will be a good idea redo all results, there are already 40 results on the sheet and redoing all of these 40 is gonna take some time

Also, added some more P3s on the Biostar M6VCG mobo atm

I added another tab and kept the existing results. If it's something we have time for we can slowly add them. If not, then we'll just keep using 4.53a version and ignore the other tab. That keeps our options open.

I'll test that new version too on my machines. I have 386, 486, P180MMX, PII 400, P3 866, P4 2.6Ghz.

There's definitely a difference between version 4.53a and 16.02. 16.02 recognizes specific cpus (for example, K6 3DNow!), so it may have optimizations for different cpus. I also see a big difference when comparing between K6-2 and Pentium II. With 4.53a, the PII-333 is 2% faster than the K6-2 300/100. With 16.02, the PII is 35% faster than the K6-2.
So it looks like PII scores are about 40% higher in 16.02 vs 4.53a, where the K6-2 scores are about 10% higher. It seems to be taking advantage of some PII optimizations in 16.02?

Just a question, should stil continue with 4.53a scores, or migrate to 16.02 completely? i can also do both, but is gonna take the double of the time per CPU/PC

If I had to choose, it seems 16.02 would be the more accurate. But if you can, it would be nice to replicate all your 4.53a entries to 16.02. Once that is finished, we can switch to 16.02 exclusively. I think there may be some good information in comparing them, information that would explain the different results in a way that makes sense.

clueless1 wrote:

If I had to choose, it seems 16.02 would be the more accurate. But if you can, it would be nice to replicate all your 4.53a entries to 16.02. Once that is finished, we can switch to 16.02 exclusively. I think there may be some good information in comparing them, information that would explain the different results in a way that makes sense.

I'll will redo all scores to 16.02 and ditch 4.53a

Carlos S. M. wrote:

clueless1 wrote:

If I had to choose, it seems 16.02 would be the more accurate. But if you can, it would be nice to replicate all your 4.53a entries to 16.02. Once that is finished, we can switch to 16.02 exclusively. I think there may be some good information in comparing them, information that would explain the different results in a way that makes sense.

I'll will redo all scores to 16.02 and ditch 4.53a

Somehow, i can't run 16.02, i get the follwing error:

1A:\>7za b -md18
2Load error: no DPMI - Get csdpmi*b.zip
3
4A:\>