Thank you - my primary target is Windows 10 x64 for now, so if I can get determinism between two computers running win10 that should be enough for my needs. I'll report back should the calculations diverge. Once again thanks for your help, I very much appreciate it.
David On Wednesday, February 26, 2020 at 1:53:46 AM UTC-6, Dmitry Babokin wrote: > > We don't guarantee bit reproducibility for floating point results across > platforms (especially across x86 / ARM) - we never specifically tested for > that. But I would expect that in most of the cases the same results. If you > find the case with different results, please report them, I would be > interested to see them. > > By the way, LLVM doesn't generally provide bit reproducibility for FP > computations. But they are working on the mode, which would enable that. We > are going to reuse it. > > Dmitry > > On Tue, Feb 25, 2020 at 11:14 PM David Nadaski <[email protected] > <javascript:>> wrote: > >> Thanks so much for the quick reply Dmitry. >> So if I'm getting you right, with fast-maths off and disable-fma on, I >> can pretty much run any floating point calculations and the results are >> always guaranteed to be deterministic (the same), even across >> platforms/architectures/etc. as long as I'm using the same version of the >> compiler? >> If so that would be fantastic. >> >> David >> >> On Wednesday, February 26, 2020 at 12:59:44 AM UTC-6, David Nadaski wrote: >>> >>> I've seen a slide from SIGGRAPH 2019 whose page 45 says "to increase >>> floating point precision/determinism, don't use --opt=fast-maths and do use >>> --opt=disable-fma" >>> My question is if doing as the slides suggests makes programs compiled >>> w/ ispc completely deterministic. >>> >> -- >> You received this message because you are subscribed to the Google Groups >> "Intel SPMD Program Compiler Users" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected] <javascript:>. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/ispc-users/1f7cb640-c98b-4c33-aaff-49e5ff5e7683%40googlegroups.com >> >> <https://groups.google.com/d/msgid/ispc-users/1f7cb640-c98b-4c33-aaff-49e5ff5e7683%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> > -- You received this message because you are subscribed to the Google Groups "Intel SPMD Program Compiler Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/ispc-users/8eb71e77-1702-4a9d-b087-68845e7214d9%40googlegroups.com.
