Hi Zach,
can you please tell me how to use ZCT 2 processors?
I am running it under windows 64bit and using the compile of JA.
best
ZCT with 2 processors
Moderator: Ras
-
Werner
- Posts: 3017
- Joined: Wed Mar 08, 2006 10:09 pm
- Location: Germany
- Full name: Werner Schüle
ZCT with 2 processors
Werner
-
Werner
- Posts: 3017
- Joined: Wed Mar 08, 2006 10:09 pm
- Location: Germany
- Full name: Werner Schüle
Re: ZCT with 2 processors - answered
Here is the answer I got via email:
"Hi Werner,
You need an smp build. ZCT will only compile with smp properly using cygwin.
I did include an smp version before, but it was reported that it suffers from the
same problem as Crafty i.e not releasing threads. I'll compile one and send it to you
later so you can test it yourself.
regards,
Jim."
Sorry cannot delete my question now...
"Hi Werner,
You need an smp build. ZCT will only compile with smp properly using cygwin.
I did include an smp version before, but it was reported that it suffers from the
same problem as Crafty i.e not releasing threads. I'll compile one and send it to you
later so you can test it yourself.
regards,
Jim."
Sorry cannot delete my question now...
Werner
-
Zach Wegner
- Posts: 1922
- Joined: Thu Mar 09, 2006 12:51 am
- Location: Earth
Re: ZCT with 2 processors
Hello Werner,
The problem is a bit deeper than the solution that Jim gives. For one, the last released sources I don't think will work with SMP. The version that ran in the WCRCC hasn't been updated yet on CVS. I'd also like to fix a small bug that was popping up during the tournament that I haven't had the time to fix yet.
But the real problem is that SMP doesn't work properly on Windows. I haven't compiled it myself there, so I don't know exactly what the issues are, but it has to do with shared memory. Teemu Pudas has sent me a version that he modified to use threads on Windows, but it was based on one of the older versions. I need to integrate his changes into the latest sources before its ready for a release. All this work might take a few days.
I think I'm going to convert ZCT to threads, as it should be easier to deal with. It will also be easier to fix the problem that Jim mentioned of eating up processor time while idle.
Thanks,
Zach
The problem is a bit deeper than the solution that Jim gives. For one, the last released sources I don't think will work with SMP. The version that ran in the WCRCC hasn't been updated yet on CVS. I'd also like to fix a small bug that was popping up during the tournament that I haven't had the time to fix yet.
But the real problem is that SMP doesn't work properly on Windows. I haven't compiled it myself there, so I don't know exactly what the issues are, but it has to do with shared memory. Teemu Pudas has sent me a version that he modified to use threads on Windows, but it was based on one of the older versions. I need to integrate his changes into the latest sources before its ready for a release. All this work might take a few days.
I think I'm going to convert ZCT to threads, as it should be easier to deal with. It will also be easier to fix the problem that Jim mentioned of eating up processor time while idle.
Thanks,
Zach
-
Werner
- Posts: 3017
- Joined: Wed Mar 08, 2006 10:09 pm
- Location: Germany
- Full name: Werner Schüle
Re: ZCT: I have a wish
Hi Zach,
thanks for the answer. I will wait for your new release. For this release I have a wish:
please make an ini file for the parameters or an uci entry:
e.g. path to egbbs (as now they must be in the same folder as the engine, one for 32bit and one for 64bit with different egbb.dll....
best
thanks for the answer. I will wait for your new release. For this release I have a wish:
please make an ini file for the parameters or an uci entry:
e.g. path to egbbs (as now they must be in the same folder as the engine, one for 32bit and one for 64bit with different egbb.dll....
best
Werner
-
Dr.Wael Deeb
- Posts: 9773
- Joined: Wed Mar 08, 2006 8:44 pm
- Location: Amman,Jordan
Re: ZCT with 2 processors
Thanks Zach,I will eagerly wait for your MP versionZach Wegner wrote:Hello Werner,
The problem is a bit deeper than the solution that Jim gives. For one, the last released sources I don't think will work with SMP. The version that ran in the WCRCC hasn't been updated yet on CVS. I'd also like to fix a small bug that was popping up during the tournament that I haven't had the time to fix yet.
But the real problem is that SMP doesn't work properly on Windows. I haven't compiled it myself there, so I don't know exactly what the issues are, but it has to do with shared memory. Teemu Pudas has sent me a version that he modified to use threads on Windows, but it was based on one of the older versions. I need to integrate his changes into the latest sources before its ready for a release. All this work might take a few days.
I think I'm going to convert ZCT to threads, as it should be easier to deal with. It will also be easier to fix the problem that Jim mentioned of eating up processor time while idle.
Thanks,
Zach
Good luck....
_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….
-
Zach Wegner
- Posts: 1922
- Joined: Thu Mar 09, 2006 12:51 am
- Location: Earth
Re: ZCT: I have a wish
Werner,
I was considering removing EGBBs as well. I'd prefer not to have a huge chunk of my engine be written by someone else. I will leave them in for now, and if I decide keep them at some point I will modify their use to be more inline with my philosophies (I.e. it will keep searching after it gets a hit so it will return an exact mate score, not an approximation that will usually make progress).
But your suggestion is real easy, that should take about 5 minutes. I think I read that suggestion somewhere earlier, maybe from you, and just forgot to do it.
I was considering removing EGBBs as well. I'd prefer not to have a huge chunk of my engine be written by someone else. I will leave them in for now, and if I decide keep them at some point I will modify their use to be more inline with my philosophies (I.e. it will keep searching after it gets a hit so it will return an exact mate score, not an approximation that will usually make progress).
But your suggestion is real easy, that should take about 5 minutes. I think I read that suggestion somewhere earlier, maybe from you, and just forgot to do it.