Komodo 9.1 Release

[Frank Quisinsky]

Hi Larry,

very friendly ... thanks Larry!

But let me pay for it, I am not a VIP!
My little work is only just for fun in my free time and I have enough free time at the moment.

I will start 2 times in the year "The Fight for Place 1". The next in maybe 4 months with SF dev. from the same day a newer Komodo version will be available. I like that, because it's a hard fight and users of my page like that too.

At the moment I have a lot to do on my site in adding newer / missed engines. I hope of a new release of "Booot" and "Hakkapeliitta" ... and if users wish "Rybka" and "Houdini" I will add this engines (collect at the moment votes for it).

2 or maybe 4 newer engines for my rating list to test ...
A lot of engine updates not tested so far (EXchess, Protector, Andscacs) and a new "League 2" is building! Updates from Crafty, Hiarcs others are on the way. I believe I need up to end of October and I am up to date the first time in testing TOP-42 engines.

And to cost for elektricity:
One of my i7 4.3Ghz systems produced 18 EUR consumption per month (both running 24 hours each day). I have four of the same i7 systems but only two I am using for my FCP list. I think two system for testing TOP-42 are enough. The other two for test games, opening book other things.

Is in U.S. electricity really cheaper?

Thanks again!

Best
Frank

PS: Be sure, in October 15 your engine have from my site clearly stronger opponents. A new test will be really interesting at this time. And yes, Komodo is at the moment one of my 5 favorits ..., means Spark, Cheng, Junior, Deuterium are available too ... not say it to the others. I hope Komodo will be in October 15 my number 1.

[RJN]

WOW! I just recently purchased Komodo 9.01. This is a very fast improvement, but it also seems like a slap if the face of those of did not pay the one year subscription. It seems like a very clever way to make people feel like they should have purchased the one year option. As a long time customer, and a beta tester ( gave CPU time to help development) I did not expect this. Simple math seems to say, be cheap, get F*****. Funny Don once gave me a free copy of 8.0 when we had purchase troubles, I declined and still made the purchase. Times have changed. IMHO

Sounds like you should have gone with the subscription plan -- Larry is a very reasonable person and i'm sure if you asked nicely and told him how you blew it , he would let you convert the single copy purchase to a subscription plan by paying the difference.

Indeed, but, even though I did state my opinion ; I knew I would purchase the upgraded engine either way.

Yeah, in hindsight should have got the sub. Maybe it can be an option for existing customers to upgrade within a certain time (or one release)? A backdate of the sub service start would be acceptable as a penality

[shrapnel]

This latest Komodo seems to be reaching greater Depths even faster than previous versions and WITHOUT any loss of strength....bad news indeed for Stockfish .
Default Drawscore -10 seems PERFECT !
Here is a game I just won using the latest Komodo....
[pgn][Event "120s/Move"]
[Site "www.come2play.com"]
[Date "2015.06.22"]
[Round "?"]
[White "Stockfish 050615 64 BMI2"]
[Black "Komodo 9.1 64-bit"]
[Result "0-1"]
[ECO "D47"]
[PlyCount "124"]

{16384MB, 1337chess PRO-140615 -1n9c4d0c3 -id7712.ctg, Anil} 1. d4 {0} d5 {B 0}
2. c4 {19} c6 {B 0} 3. Nc3 {23} Nf6 {B 0} 4. Nf3 {20} e6 {B 0} 5. e3 {20} Nbd7
{B 0} 6. Bd3 {24} dxc4 {B 0} 7. Bxc4 {10} b5 {B 0} 8. Bd3 {20} Bb7 {B 0} 9. e4
{19} b4 {B 0} 10. Na4 {20} c5 {B 0} 11. e5 {18} Nd5 {B 0} 12. dxc5 {22} Nxc5 {
B 0} 13. Nxc5 {16} Bxc5 {B 0} 14. O-O {21} h6 {B 0} 15. Nd2 {19} Qc7 {B 0} 16.
Re1 {24} O-O {B 0} 17. Ne4 {31} Bd4 {B 0} 18. Nd6 {20} Bxe5 {B 0} 19. Nxb7 {24}
Bf6 {B 0} 20. Qg4 {41} Qxb7 {B 0} 21. Bxh6 {25} Bxb2 {-0.24/29 69} 22. Qe4 {-0.
25/22 0} Rfe8 {-0.33/30 92} 23. Qh7+ {-0.18/23 1} Kf8 {-0.20/28 17} 24. Bg5 {
-0.26/23 1} f6 {-0.33/30 85} 25. Be3 f5 {-0.13/30 102} 26. Rab1 Nf6 {-0.47/29
76} 27. Qg6 Bc3 {-0.90/28 90} 28. Red1 a5 {-0.81/28 70} 29. Be2 Rac8 {-0.87/27
72} 30. Bf3 Qf7 {-1.02/29 71} 31. Qg5 {-1.01/23 2} Kg8 {-1.08/28 69} 32. Rbc1
a4 {-1.33/28 81} 33. Bd2 {-1.27/26 12} e5 {-1.39/28 56} 34. Bxc3 {-1.35/23 1}
bxc3 {-1.38/26 18} 35. Rc2 e4 {-1.58/27 40} 36. Be2 {-1.51/21 0} Rc5 {-1.53/27
27} 37. Rd4 Rb8 {-1.51/26 22} 38. Qc1 {-1.60/27 27} Qc7 {-1.65/26 28} 39. Bf1
a3 {-2.70/27 36} 40. Qxa3 Rb1 {-3.97/27 16} 41. g3 Qc6 {-5.07/28 16} 42. Qa4
Qxa4 {-5.62/29 13} 43. Rxa4 {-5.75/29 10} Nd5 {-5.86/29 13} 44. Kg2 Nb4 {-6.76/
26 9} 45. Ra8+ Kh7 {-7.53/28 7} 46. Rb8 Nxc2 {-8.78/27 7} 47. Rxb1 {-9.30/29 18
} Nb4 {-9.64/26 9} 48. a4 c2 {-10.98/26 7} 49. Ra1 {-12.87/29 26} c1=Q {-14.50/
27 1} 50. Rxc1 {-250.00/34 15} Rxc1 {-250.00/27 0} 51. Bb5 {-250.00/37 46} Nd3
{-250.00/35 4} 52. h4 Rc2 {-250.00/35 4} 53. Kg1 {-250.00/37 22} Nxf2 {-250.00/
35 5} 54. a5 e3 {-250.00/35 4} 55. Kf1 Ne4 {-250.00/32 4} 56. a6 {-#28/25 21}
Nxg3+ {-#28/25 4} 57. Kg1 {-#23/27 18} Ne2+ {-#15/26 5} 58. Kf1 Nd4 {-#13/25 3}
59. a7 Rf2+ {-#12/24 4} 60. Ke1 Ra2 {-#11/24 4} 61. Kf1 {-#10/27 21} Nxb5 {
-#10/25 4} 62. a8=Q Rxa8 {-#9/25 2} 0-1[/pgn]

[RJN]

Does anyone else see hash size for Komodo 9.1 being 1-2GB smaller when looking at what is actually being reported with task manager?

With 4GB hash, I see 3GB reported as used.
With 8GB hash, I see 6GB reported as used.

(I did not try other sizes yet)

Komodo 9.1 Win7 64 bit

[MikeB]

Indeed, but, even though I did state my opinion ; I knew I would purchase the upgraded engine either way.

hey I'm in the same boat as you - but I don't have misgivings at all. In fact , I have always felt good about supporting those who are making a living from this type of endeavor - it can't be easy. To paraphrase that old saying with a twist, "you're only as good as the last engine you'd developed". The volunteer Stockfish team is doing great things and like wise the professional team of Mark and Larry. I only wish that Don was here to see what he help started.

And Crafty is not dead, Bob's on a path to make Crafty a formidable foe once again. We still have a ways to go , but progress is being made and emails are often exchanged in the wee hours of the night between Bob, Tracey and myself. There's no magic bullet to get one big ELO change ( well unless you decide to clone somebody else's stronger engine) but the changes come in a few ELO here, a few Elo there etc. But both Stockfish and Komodo are far ahead of the rest of the pack , it will take some time to even get close to that level. I would say right now Crafty is in-between Deep Junior and Hiarcs and probably closer to DJ than HIARCS.

[mjlef]

Does anyone else see hash size for Komodo 9.1 being 1-2GB smaller when looking at what is actually being reported with task manager?

With 4GB hash, I see 3GB reported as used.
With 8GB hash, I see 6GB reported as used.

(I did not try other sizes yet)

Komodo 9.1 Win7 64 bit

That is normal. I redesigned the hash system to fit in about 75% of the space it took up before. When you tell the computer to use say 8 GB of memory, Komodo will allocate the multiple of 2 for the biggest hash table it can fit in that space. Due to the way hashing works, each increase in size is a doubling. So anyway, 6 GB is the biggest hash it can allocate in an 8 GB chunk of memory. The next step up would be 12 GB, and so on. Doing this let us fit in more hashed positions in a given amount of memory. Basically you are getting what was 8 GB of memory now compacted to only take up 6 GB.

[RJN]

Does anyone else see hash size for Komodo 9.1 being 1-2GB smaller when looking at what is actually being reported with task manager?

With 4GB hash, I see 3GB reported as used.
With 8GB hash, I see 6GB reported as used.

(I did not try other sizes yet)

Komodo 9.1 Win7 64 bit

That is normal. I redesigned the hash system to fit in about 75% of the space it took up before. When you tell the computer to use say 8 GB of memory, Komodo will allocate the multiple of 2 for the biggest hash table it can fit in that space. Due to the way hashing works, each increase in size is a doubling. So anyway, 6 GB is the biggest hash it can allocate in an 8 GB chunk of memory. The next step up would be 12 GB, and so on. Doing this let us fit in more hashed positions in a given amount of memory. Basically you are getting what was 8 GB of memory now compacted to only take up 6 GB.

Thanks Mark. I assume this is new with 9.1, as I don't recall seeing it with previous Komodo versions.

Besides saving memory, is there any other strength advantage to this, like a speedup?

[mjlef]

Does anyone else see hash size for Komodo 9.1 being 1-2GB smaller when looking at what is actually being reported with task manager?

With 4GB hash, I see 3GB reported as used.
With 8GB hash, I see 6GB reported as used.

(I did not try other sizes yet)

Komodo 9.1 Win7 64 bit

That is normal. I redesigned the hash system to fit in about 75% of the space it took up before. When you tell the computer to use say 8 GB of memory, Komodo will allocate the multiple of 2 for the biggest hash table it can fit in that space. Due to the way hashing works, each increase in size is a doubling. So anyway, 6 GB is the biggest hash it can allocate in an 8 GB chunk of memory. The next step up would be 12 GB, and so on. Doing this let us fit in more hashed positions in a given amount of memory. Basically you are getting what was 8 GB of memory now compacted to only take up 6 GB.

Thanks Mark. I assume this is new with 9.1, as I don't recall seeing it with previous Komodo versions.

Besides saving memory, is there any other strength advantage to this, like a speedup?

Yes, it is about 2% faster, since Hash memory is accessed pretty slowly. On a NUMA machine, maybe a bit more since Hash tables are generally shared between all NUMA nodes.

[lkaufman]

Hi Larry,

very friendly ... thanks Larry!

But let me pay for it, I am not a VIP!
My little work is only just for fun in my free time and I have enough free time at the moment.

I will start 2 times in the year "The Fight for Place 1". The next in maybe 4 months with SF dev. from the same day a newer Komodo version will be available. I like that, because it's a hard fight and users of my page like that too.

At the moment I have a lot to do on my site in adding newer / missed engines. I hope of a new release of "Booot" and "Hakkapeliitta" ... and if users wish "Rybka" and "Houdini" I will add this engines (collect at the moment votes for it).

2 or maybe 4 newer engines for my rating list to test ...
A lot of engine updates not tested so far (EXchess, Protector, Andscacs) and a new "League 2" is building! Updates from Crafty, Hiarcs others are on the way. I believe I need up to end of October and I am up to date the first time in testing TOP-42 engines.

And to cost for elektricity:
One of my i7 4.3Ghz systems produced 18 EUR consumption per month (both running 24 hours each day). I have four of the same i7 systems but only two I am using for my FCP list. I think two system for testing TOP-42 are enough. The other two for test games, opening book other things.

Is in U.S. electricity really cheaper?

Thanks again!

Best
Frank

PS: Be sure, in October 15 your engine have from my site clearly stronger opponents. A new test will be really interesting at this time. And yes, Komodo is at the moment one of my 5 favorits ..., means Spark, Cheng, Junior, Deuterium are available too ... not say it to the others. I hope Komodo will be in October 15 my number 1.

I've heard electricity is something like 1/3 of the price in Germany. I run 60 cores 24 hours a day, and don't pay much attention to how much it adds to my bill; maybe it eats up 1% of Komodo sales or so, even though I have to run an extra air conditioner all summer.

[Vinvin]

20 Elo in 2 months is tremendous !
Will it be an upgrade for the Chessbase version too ? I plan to buy the Chessbase version in a couple of weeks ...