Chess2u
Would you like to react to this message? Create an account in a few clicks or log in to continue.

Chess2uLog in

Mars

descriptionMars EmptyMars

more_horiz
New Ippolit:

Mars 1 x64
Mars 1 x86 : there are 2 versions - usual and PGO.

Last edited by majortom on Mon Aug 19, 2013 5:48 am; edited 2 times in total

descriptionMars EmptyRe: Mars

more_horiz
@majortom wrote:
New Ippolit:

Mars 1 x64

Is there an elo improvement?

descriptionMars EmptyRe: Mars

more_horiz
Congrats majortom.. and thanks for the new engine. thumb up

descriptionMars EmptyRe: Mars

more_horiz

Code:

Games Completed = 1000 of 1000 (Avg game length = 15.861 sec)
Settings = RR/32MB/2000ms+100ms/M 700000cp for 7 moves, D 150 moves/EPD:E:\LittleBlitzer\openings.epd(4000)
Time = 4462 sec elapsed, 0 sec remaining

 1.  Mars 1 x64               577.5/1000 406-251-343   (L: m=251 t=0 i=0 a=0) (D: r=217 i=34 f=56 s=0 a=36) (tpm=114.2 d=13.05 nps=1390772)
 2.  Stockfish 090813 x64     422.5/1000 251-406-343   (L: m=406 t=0 i=0 a=0) (D: r=217 i=34 f=56 s=0 a=36) (tpm=114.4 d=15.83 nps=1054487)

descriptionMars EmptyRe: Mars

more_horiz

Code:

Games Completed = 500 of 500 (Avg game length = 33.773 sec)
Settings = RR/32MB/4000ms+200ms/M 700000cp for 7 moves, D 150 moves/EPD:E:\LittleBlitzer\openings.epd(4000)
Time = 4447 sec elapsed, 0 sec remaining

 1.  Mars 1 x64               264.0/500 172-144-184   (L: m=144 t=0 i=0 a=0) (D: r=112 i=28 f=22 s=5 a=17) (tpm=223.6 d=14.33 nps=1389083)
 2.  Komodo 5.1r2 64-bit      236.0/500 144-172-184   (L: m=172 t=0 i=0 a=0) (D: r=112 i=28 f=22 s=5 a=17) (tpm=227.2 d=12.95 nps=811687)

descriptionMars EmptyRe: Mars

more_horiz

Code:

Games Completed = 500 of 500 (Avg game length = 33.968 sec)
Settings = RR/32MB/4000ms+200ms/M 700000cp for 7 moves, D 150 moves/EPD:E:\LittleBlitzer\openings.epd(4000)
Time = 4480 sec elapsed, 0 sec remaining

 1.  Houdini 3 Pro x64        318.0/500 229-93-178   (L: m=57 t=36 i=0 a=0) (D: r=77 i=51 f=25 s=5 a=20) (tpm=221.0 d=13.97 nps=1351823)
 2.  Mars 1 x64               182.0/500 93-229-178   (L: m=229 t=0 i=0 a=0) (D: r=77 i=51 f=25 s=5 a=20) (tpm=229.0 d=13.88 nps=1328708)

descriptionMars EmptyRe: Mars

more_horiz
Andres wrote:
Congrats majortom.. and thanks for the new engine. thumb up
sorry to disappoint guys, but unfortunately it's neither "new Ippolit", nor is it a new engine

it appears to be Fire 2.2 (or perhaps Firenzina) recompiled with many UCI options removed
...for ex: type "default" or "random" at the console prompt and the engine will produce Fire's fire.cfg

Last edited by kranium on Sun Aug 11, 2013 12:17 pm; edited 1 time in total

descriptionMars EmptyRe: Mars

more_horiz
Thanks Norm.

BTW; Are there any new projects you're working on?

descriptionMars EmptyRe: Mars

more_horiz
Vladimir wrote:
Thanks Norm.

BTW; Are there any new projects you're working on?

nothing new at the moment Vlad...
(beautiful weather here, have been playing a lot of golf!)

descriptionMars EmptyRe: Mars

more_horiz
@kranium wrote:
(beautiful weather here, have been playing a lot of golf!)

yea, like this? razz bigsmile

descriptionMars EmptyRe: Mars

more_horiz
ouch i think he busted his knee!

smile 
well unlike him i can actually hit the ball...
(i usually don't consume a 5th of Vodka before playing!)

descriptionMars EmptyRe: Mars

more_horiz
Mars 1 AVX x64 30 sec per game 1-core tests:

Code:

Critter 1.6a 64-bit           : 1000 (+228,=491,-281), 47.3 %
Komodo 5.1r2 64-bit           : 1000 (+278,=393,-329), 47.4 %

descriptionMars EmptyRe: Mars

more_horiz
Vladimir wrote:
@majortom wrote:
New Ippolit:

Mars 1 x64
Is there an elo improvement?

I guess, Mars 1 is about 20 ELO stronger than Fire 2.2 xTreme Tr4

descriptionMars EmptyRe: Mars

more_horiz
Bob wrote:

Code:

Games Completed = 500 of 500 (Avg game length = 33.968 sec)
Settings = RR/32MB/4000ms+200ms/M 700000cp for 7 moves, D 150 moves/EPD:E:\LittleBlitzer\openings.epd(4000)
Time = 4480 sec elapsed, 0 sec remaining

 1.  Houdini 3 Pro x64        318.0/500 229-93-178   (L: m=57 t=36 i=0 a=0) (D: r=77 i=51 f=25 s=5 a=20) (tpm=221.0 d=13.97 nps=1351823)
 2.  Mars 1 x64               182.0/500 93-229-178   (L: m=229 t=0 i=0 a=0) (D: r=77 i=51 f=25 s=5 a=20) (tpm=229.0 d=13.88 nps=1328708)

Bob, thank you for the tests!

descriptionMars EmptyRe: Mars

more_horiz
Mars 1 AVX x64 30 sec per game 1-core tests:

Code:

Stockfish 110813 64 SSE4.2    : 1000 (+318,=407,-275), 52.1 %
Deep Rybka 4.1 SSE42 x64      : 1000 (+368,=409,-223), 57.2 %

Games

descriptionMars EmptyRe: Mars

more_horiz
Mars 1 AVX x64 30 sec per game 1-core tests:

Code:

Houdini 3 Pro x64             : 1000 (+145,=379,-476), 33.5 %
Houdini 2.0c Pro x64          : 1000 (+231,=415,-354), 43.9 %

descriptionMars EmptyRe: Mars

more_horiz
@majortom wrote:
Vladimir wrote:
@majortom wrote:
New Ippolit:

Mars 1 x64

Is there an elo improvement?

I guess, Mars 1 is about 20 ELO stronger than Fire 2.2 xTreme Tr4

Is it stronger than RobboLito 0.21?

descriptionMars EmptyRe: Mars

more_horiz
Vladimir wrote:
@majortom wrote:
Vladimir wrote:
@majortom wrote:
New Ippolit:

Mars 1 x64
Is there an elo improvement?
I guess, Mars 1 is about 20 ELO stronger than Fire 2.2 xTreme Tr4
Is it stronger than RobboLito 0.21?
Yes, Mars 1 is about 20 ELO stronger than RobboLito 0.21.
Now it is testing at http://ls-ratinglist.beepworld.de/
1650 games played: 3050 LS-ELO
The final results next monday.

descriptionMars EmptyRe: Mars

more_horiz
Hmmm....Ok Mars, I will put your strength to the test. cool


Mars 1 x64 Gauntlet
https://www.chess2u.com/t7470-mars-1-x64-gauntlet

descriptionMars EmptyRe: Mars

more_horiz
Mars x32bit pls?

descriptionMars EmptyRe: Mars

more_horiz
I will try out x86 version today or tomorrow when Mars 1.1 will be released. I hope it will be 10 ELO above the predecessor.

descriptionMars EmptyRe: Mars

more_horiz
@majortom wrote:
I will try out x86 version today or tomorrow when Mars 1.1 will be released. I hope it will be 10 ELO above the predecessor.

chok 

The test of Mars 1 for the LS-ratinglist will take still 5 days and there is a new version to come?!?
I would really appreciate, that a new version of one engine had a little longer term...A good test is a long test and so my tests for the LS-ratinglist are taking around 6 days on one PC...
Its not so nice, if a new engine-version is coming, before the test of the older version is finished...And a longer term between two versions would allow more progress.
If there is a new release this week, I will abort the Mars 1 testrun and start a new testrun with Mars 1.1, because I dont want to waste time and money for an outdated version. But then we will dont know, if there is a progress from Mars 1 to 1.1.

By the way: Congrats to your Gull 2.1 AVX compile. Scores very good. After 1750 games 3035 LS-Elo (+17 to Gull 2.1). Perhaps you shot Gull 2.1 up into the LS-top10 tournament (and into the top10 of the strongest engines on the planet !) - but this is no sure, yet. Still a lot of games to play. I hope there will be no new version of that engine from you this week, too?!? smile 

Best - Stefan

descriptionMars EmptyRe: Mars

more_horiz
You are right, Stefan, I need wait some time before new release. I will out Mars 1 32-bit in two days, and new version will come next week. Thank you for the tests thumb up

descriptionMars EmptyRe: Mars

more_horiz
My Logo For Chess2u Friends! thumb up

Mars
Mars Mars10

descriptionMars EmptyRe: Mars

more_horiz
Nicolas wrote:
My Logo For Chess2u Friends! thumb up

Mars
Mars Mars10

Nice... a new strong engine bigsmile bigsmile bigsmile

descriptionMars EmptyRe: Mars

more_horiz
Permissions in this forum:
You cannot reply to topics in this forum