On 1/11/2019 12:17 PM, Derek Buitenhuis wrote: > On 11/01/2019 13:28, Hendrik Leppkes wrote: >> Because the computation accumulates more inaccuarcy then FLT_EPSILON >> allows for. That value is really not of that great use. If you have >> two accurate numbers and do one calculation, it may work, but if you >> do a whole bunch of them, the error accumulates and eventually gets >> bigger then FLT_EPSILON. >> x86_32 floating point is for $reasons a tad bit less accurate then on >> x86_64, for example, resulting in the test failing. We have some other >> float tests that do (or used to) fail sporadically due to inaccuracy >> problems, which sometimes where fixed by similar means - or >> multiplifying FLT_EPSILON to make it bigger. > > OK. > > Two things: > > 1. That should be in the commit messages. > 2. Would some multiple of FLT_EPSILON make more sense?
Michael suggested 1000*FLT_EPSILON but IMO that's too big and may hide errors in future implementations. The value i used is the smallest value i found that didn't fail after several runs. 6.1e-05 for example fails. > > - Derek > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel