On Tue, 24 Feb 2009, Huiqun Zhou wrote: ZH> Dear list users: ZH> ZH> I happened to test duration times of calculating the system I'm ZH> investigating against number of pools used. There are totally ZH> 36 k points. But the results surprised me quite a lot. ZH> ZH> no pool: 6m21.02s CPU time, 6m45.88s wall time ZH> 2 pools: 7m19.39s CPU time, 7m38.99s wall time ZH> 4 pools: 11m59.09s CPU time, 12m14.66s wall time ZH> 8 pools: 21m28.77s CPU time, 21m38.71s wall time
dear huiqun zhou, please note that benchmark numbers are useless without offering a way to reproduce them. so please provide the the input used for this test. also, this looks almost like you are running multiple copies of a serial binary. please check the top of the pw.x (you _are_ pw.x, right?) output and check whether there are the to be expected differences. finally, how does the job scale with just one pool? ZH> ZH> The machine I'm using is an AMD box with 2 quad core shanghai. ZH> ZH> Is my understanding of usage of pool wrong? impossible to tell. cheers, axel. ZH> ZH> Huiqun Zhou ZH> @Nanjing University, China ZH> _______________________________________________ ZH> Pw_forum mailing list ZH> Pw_forum at pwscf.org ZH> http://www.democritos.it/mailman/listinfo/pw_forum ZH> -- ======================================================================= Axel Kohlmeyer akohlmey at cmm.chem.upenn.edu http://www.cmm.upenn.edu Center for Molecular Modeling -- University of Pennsylvania Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323 tel: 1-215-898-1582, fax: 1-215-573-6233, office-tel: 1-215-898-5425 ======================================================================= If you make something idiot-proof, the universe creates a better idiot.
