Carlos S. M. wrote:

Added multiple AMD K8 and K10 based CPUs to the list, AMD K8 destroys Intel Netrust in DOS, aparently, AMD K10 is clock per clock slower than AMD A8 in 7-Zip DOS when it comes to decompressing though

Everything destroys Netburst, and not just in DOS. 😁

Standard Def Steve wrote:

Carlos S. M. wrote:

Added multiple AMD K8 and K10 based CPUs to the list, AMD K8 destroys Intel Netrust in DOS, aparently, AMD K10 is clock per clock slower than AMD A8 in 7-Zip DOS when it comes to decompressing though

Everything destroys Netburst, and not just in DOS. 😁

Well yeah, but not with massive difference 🤣 like on these results

Added some Pentium 4s using a QDI Platinix 2A (Intel 845 SDRAM), the SDRAM memory made the P4 even slower compared to Pentium III

I added a P4 2.53, P3 933, and POD200MMX on a SiS 530 motherboard. Interesting to see 200MMX results with various L2 caches:
MVP3 with 1024KB L2
SiS 530 with 512KB L2
Triton with no L2.

All three are within 2.5% of each other on the Decompress (integer) score. It's on the Compress (memory/cache subsystem) that the differences show.

-In Compress, the SiS 530/512KB is 23% faster than the Triton/0KB
-In Compress, the MVP3/1024KB is 15% faster than the SiS 530/512KB and 42% faster than Triton/0KB.

That leads to a MIPS advantage of 11% going from 0KB to 512KB and 17% going from 0KB to 1024KB. Going from 512KB to 1024KB improves MIPS by 5.6%.

Added multiple Socket A/462 CPUs, 1 Duron 700, 1 Athlon 1.2 GHz (FSB 200) and a bunch of Athlon XPs (there two Semproms as well, but they are just rebranded Athlon XPs), Clock speed matters more than FSB speed aparently, my Athlon XP 3000+ (2.1 GHz/FSB 400) loses to my Athlon XP 3000+ (2.16/FSB 333) due to the clock speed advantage, also Athlon XP losts to the Athlon 64 clock per clock in both compression and decompression

clueless1 wrote:

I added a P4 2.53, P3 933, and POD200MMX on a SiS 530 motherboard. Interesting to see 200MMX results with various L2 caches: MV […]
Show full quote

I added a P4 2.53, P3 933, and POD200MMX on a SiS 530 motherboard. Interesting to see 200MMX results with various L2 caches:
MVP3 with 1024KB L2
SiS 530 with 512KB L2
Triton with no L2.

All three are within 2.5% of each other on the Decompress (integer) score. It's on the Compress (memory/cache subsystem) that the differences show.

-In Compress, the SiS 530/512KB is 23% faster than the Triton/0KB
-In Compress, the MVP3/1024KB is 15% faster than the SiS 530/512KB and 42% faster than Triton/0KB.

That leads to a MIPS advantage of 11% going from 0KB to 512KB and 17% going from 0KB to 1024KB. Going from 512KB to 1024KB improves MIPS by 5.6%.

I have an old Packard Bell with an Intel 440VX based board, it has 256 KB L2 acording to some specs, maybe, i could give a try to that system

I added a column at the end, "MIPS per Mhz" that might be fun to use when comparing.

Added Intel 486DX2-40 and 486DX2-50 to the chart.

Seems like this thread is dead, i need to wrtie a floppy disk again and bench systems again with the DOS version. My floppy with the 7z benchmark went bad + i had too much classwork, so this why i haven't done many sumbits recently

Carlos S. M. wrote:

Seems like this thread is dead, i need to wrtie a floppy disk again and bench systems again with the DOS version. My floppy with the 7z benchmark went bad + i had too much classwork, so this why i haven't done many sumbits recently

Must not be as interesting as the Windows 7-zip benchmark. I've already submitted all my systems. Thanks for all your contributions. 😀

Added results from a Duron 1000 and an Athlon XP 2600+ using an ECS K7S5A motherboard

Added results from an Athlon XP-M 2400+ in multiple clocks, XP 2600+, both XP 3000+s and XP 3200+ in an MSI K7N2 (nForce 2) motherboard, it seems to be barely faster than my MSI KT6V-LSR (VIA KT600) i used in previous tests, although the K7N2 refussed to boot with my Durons

The database is a bit sparce with slower systems. If anyone has 286-class, 386-class, 486-class or Pentium-class systems set up, would you mind benchmarking and adding your results to the Google spreadsheet in the first post? Instructions are in that first post, it only takes a few minutes. Well, it might take longer on a 286 or 386. In fact, I'm not even sure this benchmark runs on a 286, I don't have one to test on. Thanks guys.

clueless1 wrote:

The database is a bit sparce with slower systems. If anyone has 286-class, 386-class, 486-class or Pentium-class systems set up, would you mind benchmarking and adding your results to the Google spreadsheet in the first post? Instructions are in that first post, it only takes a few minutes. Well, it might take longer on a 286 or 386. In fact, I'm not even sure this benchmark runs on a 286, I don't have one to test on. Thanks guys.

i only got more Pentium and 486 systems. i Still have a Pentium MMX system Packard Bell (Packard Bell OEM board, Intel 430VX 256 KB cache), a socket 7 Pentium MMX i rescued from the trash recently (MS-5129, Intel 430VX 512 KB cache), a dual socket 7 system with dual MMX 233 (GA-586DX, Intel 430HX 512 KB cache), a 486 DX2 50 i rescued from the trash as well, but i need to upgrade the RAM since it only has 4 MB RAM (OPTI 495/493) also a Toshiba Tecra 750DVD i need to check the specs

None of these systems were benched because i didn't have them before except for the Packard Bell

The best compression I've seen on MSDOS 😀
Is it possible to create multiple volumes with a self extractor combined? It seems to create the .exe and multiple archives but won't let me decompress them.
Regards,

Added an entry as well for an AMD K6-2 500AFX CPU running at 450Mhz (75*6).

Also got the the error that cwsdpmi was needed (with Dos 7.1 from Windows 98SE). Running with either himem or XMS didn't help. It ran with a Dos prompt under Windows 98SE though. Eventually downloaded it from https://sandmann.dotster.com/cwsdpmi/. Only dwsdpmi.exe is needed from that zip file. After that it ran without problems.