[ dev-595 ] [ GNU/Linux 2.6.31-20-generic #58-Ubuntu SMP four-processors ] [ compiled from source with --no-gui --no-apps --openmp ] i noticed phaser splits the rotation onto 2 threads - just wondering why that might be, since i used nproc=4 on the command line.... oh, hmmm... i seem to have just got this at the packing function : |python: malloc.c:4591: _int_malloc: Assertion `(unsigned long)(size)
= (unsigned long)(nb)' failed.
-Bryan
On Tue, Nov 30, 2010 at 1:58 PM, Bryan Lepore
[ dev-595 ] [ GNU/Linux 2.6.31-20-generic #58-Ubuntu SMP four-processors ] [ compiled from source with --no-gui --no-apps --openmp ]
i noticed phaser splits the rotation onto 2 threads - just wondering why that might be, since i used nproc=4 on the command line....
Which program are you running, phenix.phaser or AutoMR? As far as I know AutoMR does not enable Phaser's parallelism, despite the presence of the nproc keyword - I think this is only there to be passed on to AutoBuild when running on the command line. -Nat
On Tue, Nov 30, 2010 at 6:24 PM, Nathaniel Echols
On Tue, Nov 30, 2010 at 1:58 PM, Bryan Lepore
wrote: [ dev-595 ] [ GNU/Linux 2.6.31-20-generic #58-Ubuntu SMP four-processors ] [ compiled from source with --no-gui --no-apps --openmp ]
i noticed phaser splits the rotation onto 2 threads - just wondering why that might be, since i used nproc=4 on the command line.... Which program are you running, phenix.phaser or AutoMR?
phenix.phaser - it is clearly "Spreading calculation onto 2 threads". just wondering. -Bryan
On Tue, Nov 30, 2010 at 3:29 PM, Bryan Lepore
On Tue, Nov 30, 2010 at 6:24 PM, Nathaniel Echols
wrote: On Tue, Nov 30, 2010 at 1:58 PM, Bryan Lepore
wrote: [ dev-595 ] [ GNU/Linux 2.6.31-20-generic #58-Ubuntu SMP four-processors ] [ compiled from source with --no-gui --no-apps --openmp ]
i noticed phaser splits the rotation onto 2 threads - just wondering why that might be, since i used nproc=4 on the command line.... Which program are you running, phenix.phaser or AutoMR?
phenix.phaser - it is clearly "Spreading calculation onto 2 threads".
Phaser doesn't have an nproc keyword - so it is probably using some default setting. In CCP4 syntax, the directive you want is this: JOBS 4 You could also try setting the environment variable OMP_NUM_THREADS - but I'm not sure if that will have any effect on Phaser. -Nat
participants (2)
-
Bryan Lepore
-
Nathaniel Echols