tag:blogger.com,1999:blog-12199273.post6948072804124280859..comments2008-03-02T15:53:17.464-07:00Comments on antonywilliams.com: Windows vs Linux Benchmarks with MatlabSyXbiThttp://www.blogger.com/profile/00362291142603955773noreply@blogger.comBlogger4125tag:blogger.com,1999:blog-12199273.post-5765237908635162632008-03-02T15:53:00.000-07:002008-03-02T15:53:00.000-07:00It is not controlled if you're comparing operating...It is not controlled if you're comparing operating system performance, but if you want to see whether your Matlab code will work faster on Linux or Vista, it seems to be a fair test.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-12199273.post-11620426312030412382008-02-07T10:53:00.000-07:002008-02-07T10:53:00.000-07:00You left out the possibility that the native MATLA...You left out the possibility that the native MATLAB implementations were different. This isn't really a controlled test at all.Mattnoreply@blogger.comtag:blogger.com,1999:blog-12199273.post-74748597645935712902008-01-29T23:28:00.000-07:002008-01-29T23:28:00.000-07:00i often just use "tic;for i=1:1000;rand(1000);end;...i often just use "tic;for i=1:1000;rand(1000);end;toc" (26 seconds on a core2,48 on a mid range p4) - its easy to type, although it doesn't test memory bandwidth, as the rand(1000) is only 1mb in size.andy_t_roohttp://www.blogger.com/profile/02029998081457691232noreply@blogger.comtag:blogger.com,1999:blog-12199273.post-19953341931092079972008-01-29T23:11:00.000-07:002008-01-29T23:11:00.000-07:00I would think that a "door" would have better benc...I would think that a "door" would have better benchmarks than "window" ha i kill http://popblogculture.blogspot.com/Anonymousnoreply@blogger.com