[computer-go] Software for a supercomputer?
Hi all, I have a possibility to organize a demonstration match with a strong human against a supercomputer at the Alternative Party in Helsinki, Finland on Oct 23-25, http://www.altparty.org/ which is a fair for computer enthusiast. The expected participation is over 1000 people. The computer (Cray CX1) has a Linux operating system, but I might be able to boot it from a USB stick to Windows, too. It consists of nodes that each have a 2 x quad core Xeon E5472 3.0GHz CPU and 32 GB RAM. I heard that with OpenMP it could parallelize within one node, but to get the full power, one would need to use MPI. I am now asking for help. Do you know which program could be used for best performance (or most cores)? Is it easy to set up things running? (I need to decide soon whether I will organize it or not.) Thanks in advance, Tapani Raiko -- Tapani Raiko, tapani.ra...@tkk.fi, +358 50 5225750 http://www.iki.fi/raiko/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
RE: [computer-go] Software for a supercomputer?
My guess would be mogo. Many Faces uses MPI, but today it only scales well to 32 cores, and it only runs on windows. How many cores do you have in your Cray? It would be easy to port to linux (just change the code to start and wait for a thread), but I don't have a linux MPI machine to test on, so it would be difficult to do it this quickly. It you can boot it to Windows, with an MPI library, I can send you something to test with to see if it works. David -Original Message- From: computer-go-boun...@computer-go.org [mailto:computer-go- boun...@computer-go.org] On Behalf Of Tapani Raiko Sent: Friday, October 09, 2009 6:21 AM To: computer-go Subject: [computer-go] Software for a supercomputer? Hi all, I have a possibility to organize a demonstration match with a strong human against a supercomputer at the Alternative Party in Helsinki, Finland on Oct 23-25, http://www.altparty.org/ which is a fair for computer enthusiast. The expected participation is over 1000 people. The computer (Cray CX1) has a Linux operating system, but I might be able to boot it from a USB stick to Windows, too. It consists of nodes that each have a 2 x quad core Xeon E5472 3.0GHz CPU and 32 GB RAM. I heard that with OpenMP it could parallelize within one node, but to get the full power, one would need to use MPI. I am now asking for help. Do you know which program could be used for best performance (or most cores)? Is it easy to set up things running? (I need to decide soon whether I will organize it or not.) Thanks in advance, Tapani Raiko -- Tapani Raiko, tapani.ra...@tkk.fi, +358 50 5225750 http://www.iki.fi/raiko/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] [Fwd: Announcement ICGA Events 2010]
On Tue, Oct 06, 2009 at 01:33:10AM +0900, Hideki Kato wrote: The tournament in Taiwan allows playing through KGS but according to the rules http://ai.csie.ndhu.edu.tw:9898/eng/p_7.htm, it's preferred to participate at least one person from each team. If not, the entry fee will be doubled. Please ask the organizer for detail (Click Contact us on http://ai.csie.ndhu.edu.tw:9898/eng/). Thanks both to you and Jacques for all the information! I will yet see if I come up with something on par with MoGoTW or Zen until the tournament. ;-) In addition to above, UEC Cup will allow remote participants, though the rules are not open yet. The registration will start Oct 9th. See http://jsb.cs.uec.ac.jp/~igo/eng/index.html for detail. Interesting! It now says that Japanese rules will be used; this is a severe problem for me, I wonder how will other cope - can any of the strong MonteCarlo programs play with Japanese rules? Reading further, it says we do not permit to play through the Internet, because it is important for us to meet together in the venue; so maybe remote participation is not possible after all... -- Petr Pasky Baudis A lot of people have my books on their bookshelves. That's the problem, they need to read them. -- Don Knuth ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
RE: [computer-go] [Fwd: Announcement ICGA Events 2010]
Many Faces fully supports Japanese rules. When it plays on KGS it only uses Japanese rules. Internally I use Chinese rules, but I adjust the score at the end of the playouts depending on how many important passes have been played. Getting it right was a little tricky, but it's not a lot of code in the end. There was some discussion in the archives you can probably find. Remi recommended adjusting the komi by one point to account for who passes last, but I don't do that. David It now says that Japanese rules will be used; this is a severe problem for me, I wonder how will other cope - can any of the strong MonteCarlo programs play with Japanese rules? Reading further, it says we do not permit to play through the Internet, because it is important for us to meet together in the venue; so maybe remote participation is not possible after all... -- Petr Pasky Baudis A lot of people have my books on their bookshelves. That's the problem, they need to read them. -- Don Knuth ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
Re: [computer-go] Rating variability on CGOS
David Fotland schrieb: Many Faces also had more trouble against pachi than you would expect from its rating. Perhaps Pachi is generally stronger, but throws away some percentage of games (even against weak players) because of some bug. Seems plausible. But instead of guessing, the standard deviation of the rating should give a good indication of such problems. So why doesn't CGOS provide the standard deviation of ranks? Should be easy enough to calculate and it provides valuable information about the buggyness of a program. In physics, a measured value without standard deviation is useless. For good reasons. David -Original Message- From: computer-go-boun...@computer-go.org [mailto:computer-go- boun...@computer-go.org] On Behalf Of Brian Sheppard Sent: Thursday, October 08, 2009 12:48 PM To: computer-go@computer-go.org Subject: [computer-go] Rating variability on CGOS About two weeks ago I took Pebbles offline for an extensive overhaul of its board representation. At that time Valkyria 3.3.4 had a 9x9 CGOS rating of roughly 2475. When I looked today, I saw Valkyria 3.3.4 rated at roughly 2334, so I wondered what was going on. I found a contributing factor: Valkyria has massively different results against Pachi than against Pebbles. It happens that Pachi started playing a day or two after Pebbles went offline. Pebbles and Pachi are both rated around 2200, but Valkyria shreds Pebbles a lot more often than Pachi: Pachi: 185 / 273 = 67.8% Pebbles: 429 / 503 = 85.3% There are a lot of lessons here... ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/ ___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
[computer-go] Software for a supercomputer?
I am now asking for help. Do you know which program could be used for best performance (or most cores)? Is it easy to set up things running? (I need to decide soon whether I will organize it or not.) My guess would be - either Mogo, or Fuego with IBM's BlueFuego MPI extension. We have some hope that BlueFuego might scale better than Mogo, but have not done any comparison of current systems. Is this 9x9 or 19x19? How many cores? The systems we have mainly tested on are 8 or 10 nodes with 8 shared memory cores in each node and infiniband interconnect. E.g. the system that won the 9x9 game against Zhou 9-Dan was a 10x8 core. Is it easy to set up? If you have MPI running it should be OK. Martin___ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/