ok... Majortom.... tks.. 
Welcome to Chess2U - The Ultimate Computer Chess Forum!
If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post and use all thefeatures of the Chess2U forums. To start viewing messages, select the forum that you want to visit from the selection below.
Message [Page 5 of 9]
Gauntlet against Critter 1.6a (TC 60+0.05):
- Code:
Program Elo + - Games Score Av.Op. Draws
1 Critter 1.6a 64-bit : 3008 6 6 5788 52.3 % 2992 60.2 %
2 Mars 1.5 AVX x64 : 3006 16 16 738 49.7 % 3008 60.8 %
3 Mars 1 AVX x64 : 2994 15 16 723 47.9 % 3008 62.7 %
4 Mars 1.7 beta x64 : 2990 8 8 2585 47.4 % 3008 60.0 %
5 Mars 1.7 bet2 x64 : 2988 10 10 1742 47.1 % 3008 59.4 %
hi, we are still waiting for the exe. generic, ... we are rooting for Mars 2.0 .... be able to overcome the 1.5 rsrsr ... you could put the item: current movie ... big hug ..
@mcdiniz wrote:hi, we are still waiting for the exe. generic, ... we are rooting for Mars 2.0 .... be able to overcome the 1.5 rsrsr ... you could put the item: current movie ... big hug ..
Hi, msdiniz! Special verion without PGO (Intel 14.0 compiler) Mars 1.5 SSE4.2 no PGO
I suppose that Intel's SSE means "for Intel SSE CPUs only". But anyway try the new one.
This said Chess_Dragon from Immortal forum:
Thank you very much, Trap, for all the versions!!
I read that popcnt accounted for most of the speed increase with newer CPU's, and mine looks very fast.
- Code:
--AMD Ph.II X6 @ 3.3 GHz
Mars 1.5 non-SSE x64
compiled by Trap
POPCNT: yes
6 CPUs
using 6 threads
go movetime 100000
-Windows 7 Home
info multipv 1 time 100012 nodes 996517752 nps 9963000 score cp 15 depth 23 ...
bestmove g1f3 ponder g8f6
He has the same CPU and average speed ~10 000 kn/s. I think Mars non-SSE is good enough for AMD.
Vladimir wrote:@majortom wrote:Mars 1.7 beta AVX x64
Is there an elo improvement?
No, Mars 1.5 still the best version.
Is this version based on Ivan or Fire? I will say early on it appears to have made real progress and is stronger than Mars 1.0 or 1.1 currently a 30-50 point gain I expect if it holds but I hate to say that based on a handful of games of 6500 but so far it appears to be able to play with the very top engines and at times is down right BRUTAL! 
I would not be surprised if Mars 1.5 cracks my top ten when I stopped the tournament (to add Komodo 7a and another engine) after 46 games it was ahead of these impressive engines: Equinox2.02,Komodo CCT,Deep Matrix, Ivanhoe B50Q,Bouquet 1.7, Samsung 1.0, Rybka 4.0, Critter 1.6a, Fire 3.0, Firenzinia 2.4.2, DeepSaros eXp, Ryanfish 1.0 and ten or so other strong engines but I think the point is made that this is a very powerful engine!!
Is it stronger than Mars 1? if so, I'll give it a try.@majortom wrote:Mars 1.8 12 versions
- Code:
Program Elo + - Games Score Av.Op. Draws
Mars 1 x64 : 3168 10 10 2100 47.2 % 3187 56.4 %
Mars 1.1 x64 : 3160 8 8 3300 42.7 % 3211 51.9 %
http://www.chess2u.com/t6886p45-engineswar-rating-list#57022
Vladimir wrote:Is it stronger than Mars 1? if so, I'll give it a try.@majortom wrote:Mars 1.8 12 versions
I'm sure that Mars 1.8 is stronger.
Will you design a version that will support egtb and does your latest version support SSE4.2 instructions? Thanks.
What engine is Mars 1.91 based off of? I would like to run an engine test.
What engine is Mars 1.91 based off of? I would like to run an engine test.
________________
VocalTechnique wrote:Will you design a version that will support egtb and does your latest version support SSE4.2 instructions? Thanks.
What engine is Mars 1.91 based off of? I would like to run an engine test.
I will design a version with egtb if Mars will be much stronger than now.
You can find the latest version that supports SSE4.2 instructions here (Mars_1.91_SSE4.2_x64.exe or Mars_1.91_SSE4.2_x86.exe - but these versions are compiled for Intel only).
Message [Page 5 of 9]
Related Topics
|
|
|