DOS ain't dead

2008-03-27

p7z457b.zip (1.3MB)

Not a whole new version, only tiny code changes. djgpp 2.04 build. Also, binaries are compressed with UPX 3.02, not 3.00 like the 2007 release. BTW, I tried unpacking - it worked.

Download from:
http://mik.dyndns.org/dos-stuff/bin/
or
http://mik.mkw.ru/dos-stuff/bin/

Source diffs in:
http://mik.dyndns.org/dos-stuff/src/p7zip/
or
http://mik.mkw.ru/dos-stuff/src/p7zip/

> Not a whole new version, only tiny code changes. djgpp 2.04 build. Also,
> binaries are compressed with UPX 3.02, not 3.00 like the 2007 release.
> BTW, I tried unpacking - it worked.

It seems a lot bigger in size, and yet even the old binary says "go32stub, v 2.04T". So I dunno what he used before. Maybe he turned on -O3 this time?? (Yes, I know 2.04 has a much larger LIBC.A, but still ....) It will be interesting to see if there would be any speedups (even for a DJGPP port) when 4.58 is released.

---
Know your limits.h

> > Not a whole new version, only tiny code changes. djgpp 2.04 build. Also,
> > binaries are compressed with UPX 3.02, not 3.00 like the 2007 release.
> > BTW, I tried unpacking - it worked.
>
> It seems a lot bigger in size, and yet even the old binary says "go32stub,
> v 2.04T". So I dunno what he used before. Maybe he turned on -O3 this
> time?? (Yes, I know 2.04 has a much larger LIBC.A, but still ....) It will
> be interesting to see if there would be any speedups (even for a DJGPP
> port) when 4.58 is released.

Well, I compared the new 4.57 compile with the old one: it seems a good amount faster (esp. on my 486, but 586 also showed some little increase). So, if you can afford the extra .EXE bloat, it's worth the "upgrade", IMO. (Still a bit slower than Win32's 7ZA + HX though, heh.)

EDIT: Try "oldp7zip b -md20" and then "p7zip b -md20" and see for yourself. Even my newer cpus show that it's faster. (Kudos, Michael!)

EDIT#2: Oops, forgot to mention: 486 is much slower running .EXEs with UPX's LZMA-packing, so it will be much much faster if you unpack ("upx -d") and repack ("upx --best") on a newer machine first. This is what I did (else you might wonder why it's seemingly slower!). Of course, I'm probably one of only a few still (crazily? muahahaha!) using such dinosaurs.