Martijn van Beurden wrote:> op 27-06-14 14:11, Erik de Castro Lopo schreef: > > There has been at least one person on this list who is able to do > > perfromance tests on FLAC. I would appreciate it if someone could > > run such a performance test before and after the above commit. > > Are you specifically referring to the ARM platform or to any > platform?The main CPU architectures; x86_64 and ix86 Erik -- ---------------------------------------------------------------------- Erik de Castro Lopo http://www.mega-nerd.com/
op 28-06-14 00:23, Erik de Castro Lopo schreef:> Martijn van Beurden wrote: >> Are you specifically referring to the ARM platform or to any >> platform? > The main CPU architectures; x86_64 and ix86 > > ErikOkay, I ran the usual scripts, and the results can be found here: http://www.audiograaf.nl/misc_stuff/long-set-of-samples-check-commit-40e817.pdf As far as I can see, this is within measurement error.
op 28-06-14 07:56, Martijn van Beurden schreef:> Okay, I ran the usual scripts, and the results can be found > here: > http://www.audiograaf.nl/misc_stuff/long-set-of-samples-check-commit-40e817.pdf > > As far as I can see, this is within measurement error.I forgot to mention: this script was ran on a Kubuntu 14.04 (x86-64) system, with FLAC compiled by GCC 4.8.2.
Martijn van Beurden wrote:> Okay, I ran the usual scripts, and the results can be found > here: > http://www.audiograaf.nl/misc_stuff/long-set-of-samples-check-commit-40e817.pdf > > As far as I can see, this is within measurement error.Cool, thats what I would expect. Currently working on fixing the cast-align issue in md5.c. This is a lot more difficult and requires that I first write some unit tests for the md5 stuff before I even change anything. Erik -- ---------------------------------------------------------------------- Erik de Castro Lopo http://www.mega-nerd.com/