Which Toga for this tournament ?

[Chris Formula]

thanks for voting for 3.1.2SE. for recent cpu's with SSE2 support, use the P4 compile. for generic cpu, please use the P3 compile.

[Denis P. Mendoza]

thanks for voting for 3.1.2SE. for recent cpu's with SSE2 support, use the P4 compile. for generic cpu, please use the P3 compile.

Adding to this, I have found some ways to squeeze some more speed on this strong single-cpu engine and uploaded it here:

http://kirr.homeunix.org/chess/discussi ... 165a196e1b

It's faster than the previously released all-purpose versions.

Coincidentally, I also found out that Jim (Ablett) also made his ICC speed compile of Toga 3.1.2SE. Now, there will be two versions to choose from. Thanks Master Jim!

Here is the link as well:
http://www.zshare.net/download/786431232e485a/
or
http://mihd.net/pacoh7

[Ovyron]

Thanks for these new compiles , I tested their speed on my machine...

Toga II 3.1.2SE p3

00:00:00.0 0.06 1 2 Na3
00:00:00.0 0.34 1 3 Nc3
00:00:00.0 0.40 1 13 d4
00:00:00.0 0.20 2 45 d4 d5
00:00:00.0 0.34 3 212 d4 d5 Nf3
00:00:00.0 0.20 4 815 d4 d5 Nf3 Nf6
00:00:00.0 0.30 5 3549 d4 d5 Nf3 Nf6 Nc3
00:00:00.0 0.20 6 8925 d4 d5 Nf3 Nf6 Nc3 Nc6
00:00:00.1 0.20 7 24889 d4 d5 Nf3 Nf6 Nc3 Nc6 Bf4
00:00:00.1 0.20 8 51095 d4 d5 Nf3 Nf6 Nc3 Nc6 Bf4 Bf5
00:00:00.2 0.15 9 136366 d4 Nf6 Nf3 d5 c4 Bf5 cxd5 Nxd5 Qb3 b6
00:00:00.5 0.26 10 262190 d4 Nf6 Nf3 Nc6 Nc3 e6 h4 d5 Bf4 Bd6
00:00:00.9 0.21 11 483110 d4 Nf6 Nf3 d5 c4 e6 cxd5 Bb4+ Bd2

Bxd2+ Nbxd2 exd5 Ne5 O-O
00:00:01.5 0.20 12 894742 d4 Nf6 Nf3 d5 c4 e6 cxd5 Bb4+ Bd2

Bxd2+ Nbxd2 exd5 Ne5 O-O Qc2 Nc6
00:00:03.0 0.12 13 1793634 d4 Nf6 Nf3 d5 Nc3 e6 Bf4 Bd6 Bxd6

cxd6 e3 O-O Bb5 Nc6 Bxc6 bxc6
00:00:06.8 0.30 13 4082377 e4 d5 exd5 Nf6 Bb5+ Bd7 Bxd7+ Qxd7

Nf3 Nxd5 O-O Nc6 Nc3 Nxc3 dxc3 e5 Bg5 Qxd1 Raxd1
00:00:13.1 0.32 14 7944528 e4 d5 exd5 Nf6 Nf3 Nxd5 Bc4 e6 Nc3

Be7 O-O O-O d4 Nc6 Re1 Nxc3 bxc3
00:00:36.4 0.34 15 22336156 e4 d5 e5 c5 Nf3 Nc6 Bb5 Bd7

Bxc6 Bxc6 O-O Nh6 d3 Nf5 Nc3 Qb6 b3
00:01:19.8 0.19 16 48951117 e4 e5 Nf3 Nc6 Bb5 Bc5 Nc3 Nf6

O-O O-O d3 d6 Be3 Bxe3 fxe3 Bd7 Bxc6 Bxc6
00:01:30.1 0.19 16 55107367 e4 e5 Nf3 Nc6 Bb5 Bc5 Nc3 Nf6

O-O O-O d3 d6 Be3 Bxe3 fxe3 Bd7 Bxc6 Bxc6

Toga II 3.1.2SE (New ICC Optimized build)

00:00:00.1 0.06 1 2 Na3
00:00:00.1 0.34 1 3 Nc3
00:00:00.1 0.40 1 13 d4
00:00:00.1 0.20 2 45 d4 d5
00:00:00.1 0.34 3 212 d4 d5 Nf3
00:00:00.1 0.20 4 815 d4 d5 Nf3 Nf6
00:00:00.1 0.30 5 3549 d4 d5 Nf3 Nf6 Nc3
00:00:00.1 0.20 6 8925 d4 d5 Nf3 Nf6 Nc3 Nc6
00:00:00.1 0.20 7 24889 d4 d5 Nf3 Nf6 Nc3 Nc6 Bf4
00:00:00.2 0.20 8 51095 d4 d5 Nf3 Nf6 Nc3 Nc6 Bf4 Bf5
00:00:00.3 0.15 9 136366 d4 Nf6 Nf3 d5 c4 Bf5 cxd5 Nxd5 Qb3 b6
00:00:00.5 0.26 10 262190 d4 Nf6 Nf3 Nc6 Nc3 e6 h4 d5 Bf4 Bd6
00:00:00.9 0.21 11 483110 d4 Nf6 Nf3 d5 c4 e6 cxd5 Bb4+ Bd2

Bxd2+ Nbxd2 exd5 Ne5 O-O
00:00:01.5 0.20 12 894742 d4 Nf6 Nf3 d5 c4 e6 cxd5 Bb4+ Bd2

Bxd2+ Nbxd2 exd5 Ne5 O-O Qc2 Nc6
00:00:03.0 0.12 13 1793634 d4 Nf6 Nf3 d5 Nc3 e6 Bf4 Bd6 Bxd6

cxd6 e3 O-O Bb5 Nc6 Bxc6 bxc6
00:00:06.8 0.30 13 4082377 e4 d5 exd5 Nf6 Bb5+ Bd7 Bxd7+ Qxd7

Nf3 Nxd5 O-O Nc6 Nc3 Nxc3 dxc3 e5 Bg5 Qxd1 Raxd1
00:00:13.0 0.32 14 7944528 e4 d5 exd5 Nf6 Nf3 Nxd5 Bc4 e6 Nc3

Be7 O-O O-O d4 Nc6 Re1 Nxc3 bxc3
00:00:36.2 0.34 15 22336156 e4 d5 e5 c5 Nf3 Nc6 Bb5 Bd7

Bxc6 Bxc6 O-O Nh6 d3 Nf5 Nc3 Qb6 b3
00:01:18.8 0.19 16 48951117 e4 e5 Nf3 Nc6 Bb5 Bc5 Nc3 Nf6

O-O O-O d3 d6 Be3 Bxe3 fxe3 Bd7 Bxc6 Bxc6
00:01:28.8 0.19 16 55107367 e4 e5 Nf3 Nc6 Bb5 Bc5 Nc3 Nf6

O-O O-O d3 d6 Be3 Bxe3 fxe3 Bd7 Bxc6 Bxc6

Toga II 3.1.2SE JA

00:00:00.0 0.06 1 2 Na3
00:00:00.0 0.34 1 3 Nc3
00:00:00.0 0.40 1 13 d4
00:00:00.0 0.20 2 45 d4 d5
00:00:00.0 0.34 3 212 d4 d5 Nf3
00:00:00.0 0.20 4 815 d4 d5 Nf3 Nf6
00:00:00.0 0.30 5 3549 d4 d5 Nf3 Nf6 Nc3
00:00:00.0 0.20 6 8925 d4 d5 Nf3 Nf6 Nc3 Nc6
00:00:00.1 0.20 7 24889 d4 d5 Nf3 Nf6 Nc3 Nc6 Bf4
00:00:00.1 0.20 8 51095 d4 d5 Nf3 Nf6 Nc3 Nc6 Bf4 Bf5
00:00:00.2 0.15 9 136366 d4 Nf6 Nf3 d5 c4 Bf5 cxd5 Nxd5 Qb3 b6
00:00:00.5 0.26 10 262190 d4 Nf6 Nf3 Nc6 Nc3 e6 h4 d5 Bf4 Bd6
00:00:00.8 0.21 11 483110 d4 Nf6 Nf3 d5 c4 e6 cxd5 Bb4+ Bd2

Bxd2+ Nbxd2 exd5 Ne5 O-O
00:00:01.5 0.20 12 894742 d4 Nf6 Nf3 d5 c4 e6 cxd5 Bb4+ Bd2

Bxd2+ Nbxd2 exd5 Ne5 O-O Qc2 Nc6
00:00:02.9 0.12 13 1793634 d4 Nf6 Nf3 d5 Nc3 e6 Bf4 Bd6 Bxd6

cxd6 e3 O-O Bb5 Nc6 Bxc6 bxc6
00:00:06.6 0.30 13 4082377 e4 d5 exd5 Nf6 Bb5+ Bd7 Bxd7+ Qxd7

Nf3 Nxd5 O-O Nc6 Nc3 Nxc3 dxc3 e5 Bg5 Qxd1 Raxd1
00:00:12.7 0.32 14 7944528 e4 d5 exd5 Nf6 Nf3 Nxd5 Bc4 e6 Nc3

Be7 O-O O-O d4 Nc6 Re1 Nxc3 bxc3
00:00:35.3 0.34 15 22336156 e4 d5 e5 c5 Nf3 Nc6 Bb5 Bd7

Bxc6 Bxc6 O-O Nh6 d3 Nf5 Nc3 Qb6 b3
00:01:17.4 0.19 16 48951117 e4 e5 Nf3 Nc6 Bb5 Bc5 Nc3 Nf6

O-O O-O d3 d6 Be3 Bxe3 fxe3 Bd7 Bxc6 Bxc6
00:01:27.4 0.19 16 55107367 e4 e5 Nf3 Nc6 Bb5 Bc5 Nc3 Nf6

O-O O-O d3 d6 Be3 Bxe3 fxe3 Bd7 Bxc6 Bxc6

...and it seems Jim Ablett's compile is a little faster so that's the one I'll be using

[Denis P. Mendoza]

Thanks Ulysses. My mentor (JA) really is the best when it comes to speed. No doubt about it!

[Jim Ablett]

Thanks Ulysses. My mentor (JA) really is the best when it comes to speed. No doubt about it!

Hi Denis,

I think there is nothing between us now. My Toga build is slightly faster
because I had your previous builds to test/measure against. This was very
important because it enabled me to just try various switch combinations
till I found something faster. So the trick is not to compile first if you want to be the fastest

regards,
Jim.

[Denis P. Mendoza]

Thanks Ulysses. My mentor (JA) really is the best when it comes to speed. No doubt about it!

Hi Denis,

I think there is nothing between us now. My Toga build is slightly faster
because I had your previous builds to test/measure against. This was very
important because it enabled me to just try various switch combinations
till I found something faster. So the trick is not to compile first if you want to be the fastest

regards,
Jim.

Hello Jim,

I'll remember that.. .... now it's my turn .
Anyway, your compiles were also my benchmarks to at least reach that level. There's always room for improvement!
[/b]

[Tony Thomas]

Jim do you think Bryans 2D if he were to come back and start actively compiling again?

[Jim Ablett]

Jim do you think Bryans 2D if he were to come back and start actively compiling again?

Latest Intel compilers produce such fast code automatically, that improving on that yourself (through manual inlining) is now almost impossible.
Manual inlining was the method Bryan used to gain extra speed in his compiles. i.e forcing the compiler to inline functions where you tell it to.

Jim.

[Tony Thomas]

For some reason I only wrote about Bryans 2D part, and didnt say anything about kicking it. Weird, I could have sworn that I typed it. I meant to ask " do you think you can kick Bryans 2D if he were to come back?".. I guess both of you will be kicking yourself, because such a competition would be dead even, right?

[Jim Ablett]

For some reason I only wrote about Bryans 2D part, and didnt say anything about kicking it. Weird, I could have sworn that I typed it. I meant to ask " do you think you can kick Bryans 2D if he were to come back?".. I guess both of you will be kicking yourself, because such a competition would be dead even, right?

I'm sure Bryan would skillfully find ways to squeeze out extra nps if he needed to.
That's why he's still the best speed compiler in my opinion, whether he's active now or not

Jim.