I'm using 6.22 with a barebones autoexec/config. Just XMS loading. I actually have a hard disk with DOS installed on it, and I connect it to each machine I test.
I'm using 6.22 with a barebones autoexec/config. Just XMS loading. I actually have a hard disk with DOS installed on it, and I connect it to each machine I test.
Using DOS boot flopy made from a Windows 7 machine (probably DOS 7.1), no XMS aparently
Tried 7z-zip 16.02 in DOS with CWSDMPI and it worked on my thinkpad, but i tried in another board (QDI Advance 9) and it got errors for some reason, 4.53a runs without problems on that system
I've only ever tried with DOS 6.22. I've tried with it on hdd and boot floppy. So far it has worked on every machine: 486, Pentium, PII, K6-2. I wonder if there is something about Win7's DOS that causes this issue? Do you have access to an older DOS?
I've only ever tried with DOS 6.22. I've tried with it on hdd and boot floppy. So far it has worked on every machine: 486, Pentium, PII, K6-2. I wonder if there is something about Win7's DOS that causes this issue? Do you have access to an older DOS?
I used a Windows 98 based boot floppy on that test, but somehow, is not working in all machines ¿do you have an floppy image of what DOS floppy are you using? I don't have any DOS 6.22 machine here, Windows 95B is the oldest i got
I've only ever tried with DOS 6.22. I've tried with it on hdd and boot floppy. So far it has worked on every machine: 486, Pentium, PII, K6-2. I wonder if there is something about Win7's DOS that causes this issue? Do you have access to an older DOS?
I tried the 6.22 Bootdisk, still begs for cwsdmpi, even with EMM386 enabled
Update. tried another version of cwsdpmi and worked now
I've only ever tried with DOS 6.22. I've tried with it on hdd and boot floppy. So far it has worked on every machine: 486, Pentium, PII, K6-2. I wonder if there is something about Win7's DOS that causes this issue? Do you have access to an older DOS?
I tried the 6.22 Bootdisk, still begs for cwsdmpi, even with EMM386 enabled
I don't enable EMS. Just Extended. If you use Phil's MS-DOS boot menu, it's option 5 that I use (Extend Memory with Mouse).
Try redownloading the file. Here's a link to my post with the 16.02 version: 7-Zip Benchmark, DOS Edition
Some things to verify the file:
file size is 521,104 bytes
when you run the benchmark, it says on the top line that it is version 16.02. Here's what the output should look like:
I've only ever tried with DOS 6.22. I've tried with it on hdd and boot floppy. So far it has worked on every machine: 486, Pentium, PII, K6-2. I wonder if there is something about Win7's DOS that causes this issue? Do you have access to an older DOS?
I tried the 6.22 Bootdisk, still begs for cwsdmpi, even with EMM386 enabled
Update. tried another version of cwsdpmi and worked now
You should not have to use CWSDPMI at all. I do not and it works fine without it.
Added a bunch of Pentium 4s using an Intel D850EMV2 (Intel 850E w/ PC800 Rambus), Clock per clock wise, Pentium 4 is considerably slower than Pentium 3 and it's results looked kinda embarassing, it needed ~2 GHz to outperform my 1 Ghz P3+VIA 693A combo on integer (decompression), if that 1 ghz P3 was copuled with a faster board like a 440BX or i815, it might end in a tie or even outperforming the P4
Added a Pentium 3 Tualatin (P3-S 1.4 GHz) and one of my Pentium 4 HT builds (3.2 Ghz HT Northwood), i tried with both HT on and off, but disabling HT didn't make a difference.
For older apps/games old version might be more useful. But if possible do both versions.
After benching all my systems, it looks like my 486 and Pentium cpus give nearly identical results between the old 4.53a version and the new 16.02 version. It's only with Pentium II and K6-2 and faster that I start to see a divergence in scores. I believe because the new version of 7za.exe is aware of the various cpu optimizations while the old version is not. If that's indeed the case, I see no reason to maintain scores for both versions. But I'm not going to delete the old scores, at least not for awhile until we've had time to study and digest the scores.
The P4 pretty much sucked unless code was optimized for SSE2. If 7za for DOS doesn't select the appropriate code path, you get a good illustration of this.
Added results for multiple K6-2, Pentium MMX and classic Pentium using my Epox EP-MVP3G2 motherboard, also some kinda crazy overclocks like K6-2 at 600 MHz or Pentium MMX at 300 MHz
Added multiple Pentium 3s on Dell Optiplex GX150 (Intel 815E chipset), clock per clock, it was faster than my other P3 runs on the QDI Advance 9 which uses the slower VIA 693A (Apollo Pro 133), not all are shown since the Dell lacks any OC/UC option. I might do more tests with the Biostar M6VCG (VIA 694X based) later
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