Stockfish on AMD K6-III

Discussion of chess software programming and technical issues.

Moderators: hgm, Rebel, chrisw

User avatar
Murat
Posts: 170
Joined: Thu Mar 09, 2006 5:57 am
Location: Canada

Stockfish on AMD K6-III

Post by Murat »

Hello everyone,

which version of stockfish I can run on an AMD K6-III using windows XP 32bit please?

Thanks in advance.
tttony
Posts: 268
Joined: Sun Apr 24, 2011 12:33 am

Re: Stockfish on AMD K6-III

Post by tttony »

Here http://abrok.eu/stockfish/ and grab the Windows 32 compile, it should works
User avatar
Murat
Posts: 170
Joined: Thu Mar 09, 2006 5:57 am
Location: Canada

Re: Stockfish on AMD K6-III

Post by Murat »

Thank you for responding Tony.

None of the latest compiles work. As far back as v2 is instant crush. I went all the way back to 1.91 legacy. It still crushes after a few moves. Using Arena. This is very old cpu and also the laptop has 256 MB ram.
User avatar
Murat
Posts: 170
Joined: Thu Mar 09, 2006 5:57 am
Location: Canada

Re: Stockfish on AMD K6-III

Post by Murat »

In case any one is wondering, after downloading every 32bit compile of Stockfish v1.4 is running stabile under Arena. Can't go any higher.

All the best
tttony
Posts: 268
Joined: Sun Apr 24, 2011 12:33 am

Re: Stockfish on AMD K6-III

Post by tttony »

I've ran SF3 in PIII PC

AMD K6-III it's like PII

I thought that any Windows x32 compile will run in old hardware, it seems that only run in not too old hardware, maybe P4 and above?
User avatar
Murat
Posts: 170
Joined: Thu Mar 09, 2006 5:57 am
Location: Canada

Re: Stockfish on AMD K6-III

Post by Murat »

I think having 256 meg total memory is the reason for crushes.
Very little left after windows XP loads.

But I am happy with v1.4, has more then enough strength.

Thanks again
bnemias
Posts: 373
Joined: Thu Aug 14, 2008 3:21 am
Location: Albuquerque, NM

Re: Stockfish on AMD K6-III

Post by bnemias »

Murat wrote:I think having 256 meg total memory is the reason for crushes.
SF is pretty good at catching out of memory and relaying that back to the shell. What's the actual error?

I'm guessing the problem is lack of SSE support by the cpu. I've run into this before, but I can't remember which box had that issue. My solution was to do a custom build. Unfortunately, the code after SF6 requires a toolchain with c++ 11 support. That is difficult to resolve if you want support on old hardware and OS.

I have some custom builds that I can dig up if you want to try them. But first, run from a command prompt and see what the actual error is.
User avatar
Murat
Posts: 170
Joined: Thu Mar 09, 2006 5:57 am
Location: Canada

Re: Stockfish on AMD K6-III

Post by Murat »

Hello Eric,

after doing a basic stockfish bench 8 1 30 anything over v1.4 crushes with following.

Code: Select all

Exception Information
Code: 0xc000001d	Flags: 0x00000000
Record: 0x0000000000000000	ADDRESS; 0X0000000000419ba6

System Information
Windows NT 5.1 Build: 2600
CPU Vendor Code: 68747541 - 69746E65 - 444D4163
CPU Version: 000005D4  CPU Feature Code: 008021BF
CPU AMD Feature Code: C08029BF

That error was from v1.91 in particular. Version 1.4 and older run the command prompt test without a problem.

Thank you for trying to help
bnemias
Posts: 373
Joined: Thu Aug 14, 2008 3:21 am
Location: Albuquerque, NM

Re: Stockfish on AMD K6-III

Post by bnemias »

I sent you a pm with a temporary link to some custom builds I had, including SF6. I expect them both to work.
User avatar
Eelco de Groot
Posts: 4561
Joined: Sun Mar 12, 2006 2:40 am
Full name:   

Re: Stockfish on AMD K6-III

Post by Eelco de Groot »

bnemias wrote:
Murat wrote:I think having 256 meg total memory is the reason for crushes.
SF is pretty good at catching out of memory and relaying that back to the shell. What's the actual error?

I'm guessing the problem is lack of SSE support by the cpu. I've run into this before, but I can't remember which box had that issue. My solution was to do a custom build. Unfortunately, the code after SF6 requires a toolchain with c++ 11 support. That is difficult to resolve if you want support on old hardware and OS.

I have some custom builds that I can dig up if you want to try them. But first, run from a command prompt and see what the actual error is.
I think this was the oldlocks issue, Windows XP does not support the more modern way of locking threads introduced at some point in the code and you had to switch back to 'oldlocks' when making a compile. I don't remember where exactly in the code, sorry. But it was not difficult to make a custom build using old way of locking.
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan