New 64-bit Stockfish 1.5.1 (oct. 10) behaves same as 32-bit

[Spock]

They were all lost for Stockfish anyway, but time losses shouldn't happen at this time control. They are very rare

[mcostalba]

Hmmmm SF blocks on the last move before time control and it blocks when he has still many seconds on the clock.

Sf also rushes the last moves like the avaiable time is much less then what actually is. Could you please also post some game won by SF, just to see if it properly used all the available time.

Sorry, what is the GUI used ?


BTW does this laso at fixed time controls as example 4+0 ?

Thanks

[Spock]

ChessGUI - what I have used for all my chess960 testings for several months

It doesn't seem to be playing that well - barely stronger than Glaurung 2.2
Roughly how many ELO stronger would you expect this to be ?

I'm going to re-run in the Shredder GUI and see what happens there

[Jim Ablett]

Bravo _only_ to Jim.
Unfortunatly I cannot verify the speed of the 64 bit version because I have a 32 bit. But I _blindly_ trust Jim

Well then Bravo Jim!!!

As I said, since I could only compare in terms of kn/s with the previous 64-bit compile (the 1.5.1 of oct. 8), the question of compared speed is still open.
On the start_position, I found that speed was the same. I will see what happens in a few more positions (the ones I tested in http://www.talkchess.com/forum/viewtopi ... 13&start=0)

Maybe Jim can explain why he thinks his new compile is faster.

Hi Ernest,

You can run some benchmarks using Stockfish's built-in bench command on the command-line e.g > stockfish.exe bench 32 1 12 default depth

These are mine >

Code: Select all

original Intel 64 bit
----------------------
Total time (ms) : 26766    
Nodes searched  : 20934821
Nodes/second    : 782142

new msvc 64 bit
-------------------
Total time (ms) : 26515   
Nodes searched  : 20882348
Nodes/second    : 793161

latest Intel 64 bit (using latest version compiler update)
------------------------------------------------------------
Total time (ms) : 25843  
Nodes searched  : 20934821
Nodes/second    : 810077

I've got a pretty decent speedup with the new version Intel compiler I updated today.
You can download this new compile here if you wish >

http://www.mediafire.com/?zwmeijm33yx

Jim.

[mcostalba]

I've got a pretty decent speedup with the new version Intel compiler I updated today.
You can download this new compile here if you wish >

http://www.mediafire.com/?zwmeijm33yx

Jim.

Hi Jim,

this is a 3.5% speed up !!

This is really a nice improvment...for sure it translates in some elo points more.

Thanks a lot
Marco


BTW have you tried with MSVC 2010 ? On my PC it is the fastest binary I can get....but slower then yours of course

[Jim Ablett]

I've got a pretty decent speedup with the new version Intel compiler I updated today.
You can download this new compile here if you wish >

http://www.mediafire.com/?zwmeijm33yx

Jim.

Hi Jim,

this is a 3.5% speed up !!

This is really a nice improvment...for sure it translates in some elo points more.

Thanks a lot
Marco


BTW have you tried with MSVC 2010 ? On my PC it is the fastest binary I can get....but slower then yours of course

Hi Marco,

I installed and tested MSVC 2010.
When I compile it informs me that no profile guided optimization is available in the beta version

This is an absolute must for me in order to compile the fastest chess engine executables.

Looks good though. I'll just have to wait for the full release.

Jim.

[solis]

Just 10 games between Stockfish15.1.x64 and Deep Shredder 12.x64

Engine Score De St S-B
1: Deep Shredder 12.x64 6.0/10 ····· 4-2-4 24.00
2: Stockfish 15.1.x64 4.0/10 2-4-4 ····· 24.00

10 games played / Tournament finished
Name of the tournament: Arena games
Site/ Country: Q9550-XP64, United States
Level: Tournament 40/4
Hardware: Intel(R) Core(TM)2 Quad CPU Q9550 @ 2.83GHz4x 3500 MHz with 4,096 MB Memory
Operating system: Microsoft Windows XP 64 Bit Professional Service Pack 2 (Build 3790)

[Mincho Georgiev]

Hi Marco,

I installed and tested MSVC 2010.
When I compile it informs me that no profile guided optimization is available in the beta version

This is an absolute must for me in order to compile the fastest chess engine executables.

Looks good though. I'll just have to wait for the full release.

Jim.

Hm, that is really strange, because right now i actually build my testbed executables with MSVC2010beta 1, PGO mostly. Just to mention, that i am only using the executables (../bin), without the IDE, probably the restriction is valid only under the IDE.

[ernest]

I've got a pretty decent speedup with the new version Intel compiler I updated today.
http://www.mediafire.com/?zwmeijm33yx

OMG, Jim, what a pace, I have to run if I want to follow you!...

First there was a stockfish_151_x64_ja.exe from october 8
Then there was a stockfish_151_x64_ja.exe from october 10, which is still on your site http://homepages.tesco.net/henry.ablett/jims.html, which for the first time behaves identically to the 32-bit version, but which I did not find faster (in kn/s) than the previous from oct. 8. But I will check with your command-line method.
And now, this stockfish_151_intel_x64.exe from october 12, which I will of course also check (looks promising!)

[ernest]

this is a 3.5% speed up !!
This is really a nice improvment...for sure it translates in some elo points more.

Normally, +3.5% ==> +3.5 Elo