Help with Komodo testing.

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 2:27 pm

Help with Komodo testing.

Post by Don » Tue Apr 27, 2010 9:58 pm

I'm looking for help with my testing of Komodo. I can use either 64 bit windows or 64 bit linux machines.

I have a distributed tester that only requires you to run a client executable that I will provide. The tester does the rest. Any help would be greatly appreciated and would help us improve Komodo.

The way this is set up you can connect or disconnect at any time - helping as much or as little as you wish. There is web site where the games and result are monitored.

When the client is started, you give it the number cpu cores you which to utilize - it should probably be between 1 and 2X the number of physical cores in your machine. The tester estimates the speed of your computer (given the number of cores you wish to utilize) and adjusts the time controls accordingly.

Please PM me if you are willing to help, or email me at drd@mit.edu and I will give you further instructions - also provide a simple "client" name to describe your test instance, 1 for each machine you wish to utilize. I like names with your initials and simple hardware description such as "drd_i7" or something like that.

tomgdrums
Posts: 736
Joined: Wed Dec 02, 2009 8:48 am

Re: Help with Komodo testing.

Post by tomgdrums » Wed Apr 28, 2010 4:53 am

Don wrote:I'm looking for help with my testing of Komodo. I can use either 64 bit windows or 64 bit linux machines.

I have a distributed tester that only requires you to run a client executable that I will provide. The tester does the rest. Any help would be greatly appreciated and would help us improve Komodo.

The way this is set up you can connect or disconnect at any time - helping as much or as little as you wish. There is web site where the games and result are monitored.

When the client is started, you give it the number cpu cores you which to utilize - it should probably be between 1 and 2X the number of physical cores in your machine. The tester estimates the speed of your computer (given the number of cores you wish to utilize) and adjusts the time controls accordingly.

Please PM me if you are willing to help, or email me at drd@mit.edu and I will give you further instructions - also provide a simple "client" name to describe your test instance, 1 for each machine you wish to utilize. I like names with your initials and simple hardware description such as "drd_i7" or something like that.
Hi Don!

When you guys release the MP version of Komodo for sale will it be all good in the Shredder GUI? The free versions have had some kinks in them especially with regards to analysis mode. And will there be a limit strength feature?

Thanks!

User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 2:27 pm

Re: Help with Komodo testing.

Post by Don » Wed Apr 28, 2010 1:22 pm

tomgdrums wrote:
Don wrote:I'm looking for help with my testing of Komodo. I can use either 64 bit windows or 64 bit linux machines.

I have a distributed tester that only requires you to run a client executable that I will provide. The tester does the rest. Any help would be greatly appreciated and would help us improve Komodo.

The way this is set up you can connect or disconnect at any time - helping as much or as little as you wish. There is web site where the games and result are monitored.

When the client is started, you give it the number cpu cores you which to utilize - it should probably be between 1 and 2X the number of physical cores in your machine. The tester estimates the speed of your computer (given the number of cores you wish to utilize) and adjusts the time controls accordingly.

Please PM me if you are willing to help, or email me at drd@mit.edu and I will give you further instructions - also provide a simple "client" name to describe your test instance, 1 for each machine you wish to utilize. I like names with your initials and simple hardware description such as "drd_i7" or something like that.
Hi Don!

When you guys release the MP version of Komodo for sale will it be all good in the Shredder GUI? The free versions have had some kinks in them especially with regards to analysis mode. And will there be a limit strength feature?

Thanks!
Hi Tom,

If there are some glitches I would definitely want to fix it. Since I don't have the shredder GUI I cannot be sure of what the issue is, but please PM me with your email address and you can test Komodo on Shredder GUI for me when I get ready to resolve this.

Don

User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 2:27 pm

Re: Help with Komodo testing.

Post by Don » Wed Apr 28, 2010 7:10 pm

I just want to mention that the distributed tester is working quite well! I want to really think all the testers who are letting me use their spare CPU cycles.

Also, if anyone else want to join in, let me know. You are free to test as little or as much as you please. You can of course also scale your test down to just 1 core for instance and still have CPU cycles for your own stuff.

I run the testing on my own desktop machine while I'm doing other things and have no trouble. In fact I'm currently running the server, 2 cores allocated to the testing, and compiling chess programs and doing other stuff on a sow core 2 duo older machine. I'm not sure Windows can handle all of this as gracefully, but I don't even notice it on my desktop.

Don wrote:I'm looking for help with my testing of Komodo. I can use either 64 bit windows or 64 bit linux machines.

I have a distributed tester that only requires you to run a client executable that I will provide. The tester does the rest. Any help would be greatly appreciated and would help us improve Komodo.

The way this is set up you can connect or disconnect at any time - helping as much or as little as you wish. There is web site where the games and result are monitored.

When the client is started, you give it the number cpu cores you which to utilize - it should probably be between 1 and 2X the number of physical cores in your machine. The tester estimates the speed of your computer (given the number of cores you wish to utilize) and adjusts the time controls accordingly.

Please PM me if you are willing to help, or email me at drd@mit.edu and I will give you further instructions - also provide a simple "client" name to describe your test instance, 1 for each machine you wish to utilize. I like names with your initials and simple hardware description such as "drd_i7" or something like that.

User avatar
Dr.Wael Deeb
Posts: 9635
Joined: Wed Mar 08, 2006 7:44 pm
Location: Amman,Jordan

Re: Help with Komodo testing.

Post by Dr.Wael Deeb » Wed Apr 28, 2010 7:52 pm

Don wrote:I just want to mention that the distributed tester is working quite well! I want to really think all the testers who are letting me use their spare CPU cycles.

Also, if anyone else want to join in, let me know. You are free to test as little or as much as you please. You can of course also scale your test down to just 1 core for instance and still have CPU cycles for your own stuff.

I run the testing on my own desktop machine while I'm doing other things and have no trouble. In fact I'm currently running the server, 2 cores allocated to the testing, and compiling chess programs and doing other stuff on a sow core 2 duo older machine. I'm not sure Windows can handle all of this as gracefully, but I don't even notice it on my desktop.

Don wrote:I'm looking for help with my testing of Komodo. I can use either 64 bit windows or 64 bit linux machines.

I have a distributed tester that only requires you to run a client executable that I will provide. The tester does the rest. Any help would be greatly appreciated and would help us improve Komodo.

The way this is set up you can connect or disconnect at any time - helping as much or as little as you wish. There is web site where the games and result are monitored.

When the client is started, you give it the number cpu cores you which to utilize - it should probably be between 1 and 2X the number of physical cores in your machine. The tester estimates the speed of your computer (given the number of cores you wish to utilize) and adjusts the time controls accordingly.

Please PM me if you are willing to help, or email me at drd@mit.edu and I will give you further instructions - also provide a simple "client" name to describe your test instance, 1 for each machine you wish to utilize. I like names with your initials and simple hardware description such as "drd_i7" or something like that.
Are you insulting Windows Don or what :!: :?:

:lol:

:wink:
_No one can hit as hard as life.But it ain’t about how hard you can hit.It’s about how hard you can get hit and keep moving forward.How much you can take and keep moving forward….

User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 2:27 pm

Re: Help with Komodo testing.

Post by Don » Wed Apr 28, 2010 8:14 pm

Dr.Wael Deeb wrote:
Don wrote:I just want to mention that the distributed tester is working quite well! I want to really think all the testers who are letting me use their spare CPU cycles.

Also, if anyone else want to join in, let me know. You are free to test as little or as much as you please. You can of course also scale your test down to just 1 core for instance and still have CPU cycles for your own stuff.

I run the testing on my own desktop machine while I'm doing other things and have no trouble. In fact I'm currently running the server, 2 cores allocated to the testing, and compiling chess programs and doing other stuff on a sow core 2 duo older machine. I'm not sure Windows can handle all of this as gracefully, but I don't even notice it on my desktop.

Don wrote:I'm looking for help with my testing of Komodo. I can use either 64 bit windows or 64 bit linux machines.

I have a distributed tester that only requires you to run a client executable that I will provide. The tester does the rest. Any help would be greatly appreciated and would help us improve Komodo.

The way this is set up you can connect or disconnect at any time - helping as much or as little as you wish. There is web site where the games and result are monitored.

When the client is started, you give it the number cpu cores you which to utilize - it should probably be between 1 and 2X the number of physical cores in your machine. The tester estimates the speed of your computer (given the number of cores you wish to utilize) and adjusts the time controls accordingly.

Please PM me if you are willing to help, or email me at drd@mit.edu and I will give you further instructions - also provide a simple "client" name to describe your test instance, 1 for each machine you wish to utilize. I like names with your initials and simple hardware description such as "drd_i7" or something like that.
Are you insulting Windows Don or what :!: :?:

:lol:

:wink:
Of course not! I'm basing my comment on my knowledge of Unix and how robust it is and my lack of knowledge of how well Windows handles this.

I know that Windows does work pretty well for most things if you reboot it a few times every day :-) The only OS I know of where you have to reboot just to install ANY software. Have they fixed that yet?

Just kidding, I think Windows is great - I'm just putting it down. ;-)

Frank Quisinsky
Posts: 4852
Joined: Wed Nov 18, 2009 6:16 pm
Location: Trier, Germany
Contact:

Re: Help with Komodo testing.

Post by Frank Quisinsky » Wed Apr 28, 2010 8:41 pm

Hi Don,

Stockfish Interview:
http://www.schach-welt.de/spezial/compu ... -team.html

Here question 20:

20.
Frank Quisinsky
Zappa or Komodo profit from 64-Bit around 65 ELO, Rybka and Onno around 50 ELO, Glaurung and Stockfish with around 30 ELO. What is the reason for it? Maybe Stockfish will get 30 ELO points more with improvements of 64-Bit support? Is this possible?

Tord Romstad
I think this is very unlikely. It is difficult to be sure without knowing the source code of the other programs discussed, but I think Stockfish is just as well optimized for 64-bit CPUs as Zappa, Komodo, Rybka and Onno. It is more likely that the explanation for Stockfish's smaller Elo gain on 64-bit CPUs is the reverse: Stockfish is better optimized for 32-bit CPUs than most other bitboard engines. Many of the important low-level parts of Stockfish exist in two versions: A version optimized for 64-bit CPUs, and a version optimized for 32-bit CPUs. Thanks to this, Stockfish's speed hit on 32-bit hardware, while still substantial, isn't quite as big as that of most other bitboard programs.

---

I know that you develops on Linux systems and your interest is 64-Bit only. But perhaps its possible to make the 32-Bit stronger, I don't know, you are the expert.

Thinking on it:
The most chess player using 32-Bit OS systems (not all are computer chess freaks and have install a normaly 32-Bit OS, thinking on club players).

My biggest wish for the up comming Komodo version is to improve the 32-Bit version of Komodo.

Unfortunately, many work in the garden and not so many time for testing. My right-hand is sprained and I wrote with the left-hand only. For all things I need a lot of time so far, garden work too.

But I am happy if I can help with test work in the future. I like Komodo and it would be an event for me.

Have fun and thanks for your work and time.
Good luck with the upcomming version.

Best
Frank
I like computer chess!

User avatar
Don
Posts: 5106
Joined: Tue Apr 29, 2008 2:27 pm

Re: Help with Komodo testing.

Post by Don » Wed Apr 28, 2010 8:57 pm

Frank Quisinsky wrote:Hi Don,

Stockfish Interview:
http://www.schach-welt.de/spezial/compu ... -team.html

Here question 20:

20.
Frank Quisinsky
Zappa or Komodo profit from 64-Bit around 65 ELO, Rybka and Onno around 50 ELO, Glaurung and Stockfish with around 30 ELO. What is the reason for it? Maybe Stockfish will get 30 ELO points more with improvements of 64-Bit support? Is this possible?

Tord Romstad
I think this is very unlikely. It is difficult to be sure without knowing the source code of the other programs discussed, but I think Stockfish is just as well optimized for 64-bit CPUs as Zappa, Komodo, Rybka and Onno. It is more likely that the explanation for Stockfish's smaller Elo gain on 64-bit CPUs is the reverse: Stockfish is better optimized for 32-bit CPUs than most other bitboard engines. Many of the important low-level parts of Stockfish exist in two versions: A version optimized for 64-bit CPUs, and a version optimized for 32-bit CPUs. Thanks to this, Stockfish's speed hit on 32-bit hardware, while still substantial, isn't quite as big as that of most other bitboard programs.

---

I know that you develops on Linux systems and your interest is 64-Bit only. But perhaps its possible to make the 32-Bit stronger, I don't know, you are the expert.
I think Tord is exactly correct. I have spend very little effort optimizing for the 32 bit version.

I could probably look at stockish code and see what they do for 32 bit programs and make some improvement in this area. I actually did make a small improvement some time ago with doch - I improved one of the low level routines and that gave a good boost for the 32 bit stuff with little effort. But I could probably do more of this.

Thinking on it:
The most chess player using 32-Bit OS systems (not all are computer chess freaks and have install a normaly 32-Bit OS, thinking on club players).

My biggest wish for the up comming Komodo version is to improve the 32-Bit version of Komodo.
However support for 32 bit windows is dwindling and I think we are about to see a rather abrupt move to 64 bit everywhere. That means that within a year or two, probably most computing systems will be 64 bit as this is what comes with Windows 7 and all new computers. Yes, you can buy and install the 32 bit version but this won't be supported by new hardware from what I understand and there is no good reason to run 32 bit windows 7 other than installing it on an old computer.

Nevertheless, I will at least take a look at this and see if I can make an improvement - but it's not worth a major investment in time for me - since time is in short supply these days and I consider 32 bit OS's a kind of sinking ship. It just might take 2 or 3 years for it to fully sink!


Unfortunately, many work in the garden and not so many time for testing. My right-hand is sprained and I wrote with the left-hand only. For all things I need a lot of time so far, garden work too.

But I am happy if I can help with test work in the future. I like Komodo and it would be an event for me.

Have fun and thanks for your work and time.
Good luck with the upcomming version.

Best
Frank

Frank Quisinsky
Posts: 4852
Joined: Wed Nov 18, 2009 6:16 pm
Location: Trier, Germany
Contact:

Re: Help with Komodo testing.

Post by Frank Quisinsky » Wed Apr 28, 2010 9:57 pm

Hi Don,

wow, thanks for your fast message.

Yes, same opinion to 64-Bit.
End of the year I start in SWCR (rating list of the German chess magazin SCHACHWELT) with x64 engines. I think a good time to do that.

But it would be great if Komodo is a little bit more optimize for 32-Bit. Very hard to see that Komodo is with 64-Bit around 75 ELO stronger.

I think on Netbooks OS with 32-Bit can be a longer time actual. For sure with new PCs / Notebooks its need around max. 1 year and all comes with x64.

Have different good contacts to Intel Germany. Intel try it since end of the last year in first co-coperation with Acer and at the moment with others. With other words, Intel try to power x64 with a lot of energy ... with more energy as Microsoft.

Best
Frank
I like computer chess!

Post Reply