Closed mojinfu closed 5 years ago
I currently does not use more that threads until it has found a first feasible solution for the LNS to kick in. Laurent Perron | Operations Research | lperron@google.com | (33) 1 42 68 53 00
Le ven. 12 juil. 2019 à 03:55, mojinfu notifications@github.com a écrit :
my project is running on an 48-logic-core computer
when I set parameter num_search_workers = 6 , it works perfectly. ( takes 600% cpu ) ''' pyhton self.solver.parameters.search_branching = self.rule.search_branching self.solver.parameters.max_time_in_seconds = 10000 self.solver.parameters.log_search_progress = True self.solver.parameters.num_search_workers = 6 '''
but when I set parameter num_search_workers = 7 or > 7 , it still works with 600% cpu
Are there any additional parameters that need to be set ?
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/google/or-tools/issues/1420?email_source=notifications&email_token=ACUPL3NB5CTXRAOS35KSPPLP7BPIRA5CNFSM4ICHGNY2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4G63HEEA, or mute the thread https://github.com/notifications/unsubscribe-auth/ACUPL3NGKM3NTSDRO64E6S3P7BPIRANCNFSM4ICHGNYQ .
Please note that this is completely changing for the next version. Laurent Perron | Operations Research | lperron@google.com | (33) 1 42 68 53 00
Le ven. 12 juil. 2019 à 05:52, Laurent Perron lperron@google.com a écrit :
I currently does not use more that threads until it has found a first feasible solution for the LNS to kick in. Laurent Perron | Operations Research | lperron@google.com | (33) 1 42 68 53 00
Le ven. 12 juil. 2019 à 03:55, mojinfu notifications@github.com a écrit :
my project is running on an 48-logic-core computer
when I set parameter num_search_workers = 6 , it works perfectly. ( takes 600% cpu ) ''' pyhton self.solver.parameters.search_branching = self.rule.search_branching self.solver.parameters.max_time_in_seconds = 10000 self.solver.parameters.log_search_progress = True self.solver.parameters.num_search_workers = 6 '''
but when I set parameter num_search_workers = 7 or > 7 , it still works with 600% cpu
Are there any additional parameters that need to be set ?
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/google/or-tools/issues/1420?email_source=notifications&email_token=ACUPL3NB5CTXRAOS35KSPPLP7BPIRA5CNFSM4ICHGNY2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4G63HEEA, or mute the thread https://github.com/notifications/unsubscribe-auth/ACUPL3NGKM3NTSDRO64E6S3P7BPIRANCNFSM4ICHGNYQ .
i understand the reason is that “This tool contains only six algorithms for finding solutions,and every algorithm to another is can run on different thread .
but single algorithm can not run on two thread “
is that correct?
Yes and yes. Laurent Perron | Operations Research | lperron@google.com | (33) 1 42 68 53 00
Le ven. 12 juil. 2019 à 09:59, mojinfu notifications@github.com a écrit :
i understand the reason is that “This tool contains only six algorithms for finding solutions,and every algorithm to another is can run on different thread . but single algorithm can not run on two thread “ is that correct?
— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/google/or-tools/issues/1420?email_source=notifications&email_token=ACUPL3MHMC54TF55ANC445TP7CZ6DA5CNFSM4ICHGNY2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZ2J32A#issuecomment-510959080, or mute the thread https://github.com/notifications/unsubscribe-auth/ACUPL3L26N4UKEOLCHUKJALP7CZ6DANCNFSM4ICHGNYQ .
Shared search tree does not work on SAT, and is usually bad in general for other solvers. Laurent Perron | Operations Research | lperron@google.com | (33) 1 42 68 53 00
Le ven. 12 juil. 2019 à 10:06, Laurent Perron lperron@google.com a écrit :
Yes and yes. Laurent Perron | Operations Research | lperron@google.com | (33) 1 42 68 53 00
Le ven. 12 juil. 2019 à 09:59, mojinfu notifications@github.com a écrit :
i understand the reason is that “This tool contains only six algorithms for finding solutions,and every algorithm to another is can run on different thread . but single algorithm can not run on two thread “ is that correct?
— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/google/or-tools/issues/1420?email_source=notifications&email_token=ACUPL3MHMC54TF55ANC445TP7CZ6DA5CNFSM4ICHGNY2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZ2J32A#issuecomment-510959080, or mute the thread https://github.com/notifications/unsubscribe-auth/ACUPL3L26N4UKEOLCHUKJALP7CZ6DANCNFSM4ICHGNYQ .
The next version of ortools , will it be changed ? or it is still can not be ...
We are working very actively on this part. But we will never implement search tree sharing. Laurent Perron | Operations Research | lperron@google.com | (33) 1 42 68 53 00
Le ven. 12 juil. 2019 à 10:44, mojinfu notifications@github.com a écrit :
The next version of ortools , will it be changed ? or it is still can not be ...
— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/google/or-tools/issues/1420?email_source=notifications&email_token=ACUPL3LEMHKQD2F35VE2ZSTP7C7GNA5CNFSM4ICHGNY2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZ2NJYA#issuecomment-510973152, or mute the thread https://github.com/notifications/unsubscribe-auth/ACUPL3OOWPVXPWQGADPQBYLP7C7GNANCNFSM4ICHGNYQ .
my project is running on an 48-logic-core computer
when I set parameter num_search_workers = 6 , it works perfectly. ( takes 600% cpu ) ''' pyhton self.solver.parameters.search_branching = self.rule.search_branching self.solver.parameters.max_time_in_seconds = 10000 self.solver.parameters.log_search_progress = True self.solver.parameters.num_search_workers = 6 '''
but when I set parameter num_search_workers = 7 or > 7 , it still works with 600% cpu
Are there any additional parameters that need to be set ?