On Jul 14, 7:40 am, Robert Dodier <[EMAIL PROTECTED]> wrote: > Harald Schilly wrote: > > >http://www.sciviews.org/benchmark/index.html > > I have seen this benchmark, it's outdated and i think totally wrong. > > Aside from being out of date, what's wrong with it?
Well, in my opinion benchmarks only show you what a program did under a certain circumstance, but you cannot automatically derive general information - especially in that case - about that product. e.g. look at R: they run it in WinXP using mingw. I don't know much about compilers, but i've never heared a good thing about it when i comes to performance (esp. at those version numbers). Those programming tests also claim to test certain features, but there is no explanation why and no methodology. The original benchmark referenced on top goes into more details, but isn't used here, just excerpts. Also, they mix effects, like they combine the creation of data and the execution of a tested routine onto that data in one timing function. This should be done separately. I also like to see more than one size per test to get an impression how it scales! So, i think this benchmark is too short and not fair. But just my opinion ;) H --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
