My Friends! here the compiled files tested! (Include Original Source Code)
http://www.mediafire.com/file/bk3qwqqcd ... %26Bin.zip
Stockfish bench in i486 & Pentium 75mhz !
Moderator: Ras
-
- Posts: 116
- Joined: Mon Nov 28, 2011 10:17 am
-
- Posts: 373
- Joined: Thu Aug 14, 2008 3:21 am
- Location: Albuquerque, NM
Re: Stockfish bench in i486 & Pentium 75mhz !
One more note about this, there are two issues regarding toolchain.
The std=c++11 requirement is an issue forcing newer toolchains. This manifests by requiring a specific windows version, specifically one that has __fstat64 in its kernel. In theory, it should be possible to make c++11 builds that work even in '95, but I haven't bothered to seriously tackle this issue.
The other issue is optimizations and CPU support. This *should* be easily controlled by omitting -msse and -msse2, etc. But my experience says there's more to it than just that. Using older toolchains made a big difference wrt compatibility on old hardware, yet I'm not sure why since omitting the newer instruction sets should be all that is required.
The objective is to get modern gcc optimizations that work on very old hardware. It's harder than it sounds, but in theory, it should work. I've given up trying beyond what I've already built. Now, if I had access to a 386 or 486 machine, that might be all the motivation I need, but sadly such hardware is hard to come by, and is prone to being flaky just because of its age.
The std=c++11 requirement is an issue forcing newer toolchains. This manifests by requiring a specific windows version, specifically one that has __fstat64 in its kernel. In theory, it should be possible to make c++11 builds that work even in '95, but I haven't bothered to seriously tackle this issue.
The other issue is optimizations and CPU support. This *should* be easily controlled by omitting -msse and -msse2, etc. But my experience says there's more to it than just that. Using older toolchains made a big difference wrt compatibility on old hardware, yet I'm not sure why since omitting the newer instruction sets should be all that is required.
The objective is to get modern gcc optimizations that work on very old hardware. It's harder than it sounds, but in theory, it should work. I've given up trying beyond what I've already built. Now, if I had access to a 386 or 486 machine, that might be all the motivation I need, but sadly such hardware is hard to come by, and is prone to being flaky just because of its age.
-
- Posts: 116
- Joined: Mon Nov 28, 2011 10:17 am
Re: Stockfish bench in i486 & Pentium 75mhz !
Hi Eric! Exactly ...all you say is correct. Find the right combination of parameters Compiler and Libraries versions was terrible....
Thankfully it is already done...
Enjoy my contribution Friends!
Thankfully it is already done...
Enjoy my contribution Friends!
-
- Posts: 5284
- Joined: Thu Mar 09, 2006 9:40 am
- Full name: Vincent Lejeune
Re: Stockfish bench in i486 & Pentium 75mhz !
Some screenshots who got deleted from this thread :
SF6 386-33mhz (~1000 NPS)

SF6 486-33mhz (~2000 NPS)

SF6 pentium-75mhz (~6000 NPS)

SF6 386-33mhz (~1000 NPS)

SF6 486-33mhz (~2000 NPS)

SF6 pentium-75mhz (~6000 NPS)

-
- Posts: 5284
- Joined: Thu Mar 09, 2006 9:40 am
- Full name: Vincent Lejeune
-
- Posts: 512
- Joined: Tue Sep 29, 2020 4:29 pm
- Location: Dublin, Ireland
- Full name: Madeleine Birchfield
Re: Stockfish bench in i486 & Pentium 75mhz !
How well does the new Stockfish 12 work on Pentiums and i486?
-
- Posts: 5284
- Joined: Thu Mar 09, 2006 9:40 am
- Full name: Vincent Lejeune
Re: Stockfish bench in i486 & Pentium 75mhz !
Reply from here : http://talkchess.com/forum3/viewtopic.p ... 99#p863999Madeleine Birchfield wrote: ↑Mon Oct 12, 2020 5:13 pm How well does the new Stockfish 12 work on Pentiums and i486?
Some approximations here :
http://www.talkchess.com/forum3/viewtop ... =6&t=72485
http://www.talkchess.com/forum3/viewtop ... =2&t=63857
A version close to SF 11 was around 3035 at 20 kn/s.
A Pentium 75 would run SF 11 at speed around 6 kn/s. Rating for SF 11 on this P75 would be around 2900-2950.
SF 12 is around 50 RP over SF 11.
-
- Posts: 5672
- Joined: Tue Feb 28, 2012 11:56 pm
Re: Stockfish bench in i486 & Pentium 75mhz !
I'm afraid the network file may not fit on the harddisk of the i486.Vinvin wrote: ↑Mon Oct 12, 2020 7:25 pmReply from here : http://talkchess.com/forum3/viewtopic.p ... 99#p863999Madeleine Birchfield wrote: ↑Mon Oct 12, 2020 5:13 pm How well does the new Stockfish 12 work on Pentiums and i486?
Some approximations here :
http://www.talkchess.com/forum3/viewtop ... =6&t=72485
http://www.talkchess.com/forum3/viewtop ... =2&t=63857
A version close to SF 11 was around 3035 at 20 kn/s.
A Pentium 75 would run SF 11 at speed around 6 kn/s. Rating for SF 11 on this P75 would be around 2900-2950.
SF 12 is around 50 RP over SF 11.
-
- Posts: 512
- Joined: Tue Sep 29, 2020 4:29 pm
- Location: Dublin, Ireland
- Full name: Madeleine Birchfield
-
- Posts: 5672
- Joined: Tue Feb 28, 2012 11:56 pm
Re: Stockfish bench in i486 & Pentium 75mhz !
I did some googling. Hard disks in the early 90s seem to have been typically 40MB or larger, so the network could be stored.
I don't know when 32MB of RAM or more became the norm, but already the i486 seems to have supported 256MB of RAM (on some non-typical motherboards).
So it should be possible to run Stockfish 12 with NNUE on an i486 (but most i486 systems probably couldn't for lack of 32MB of RAM or more).