So far I blamed clang's low speed on Windows to a poor disc IO implementation, and/or NTFS having bad performance with many small IO tasks. But without having real proof ;)
But I've seen similar behaviour in Subversion, where checking out a repository with many small files was up to 10x slower on Windows than on a Linux VM running in the same Windows instance, doing the same checkout to the same disc, but in the VM's Linux filesystem mapped to a single NTFS file. On Friday, 31 May 2019 19:24:51 UTC+2, Alon Zakai wrote: > > Thanks for the info, interesting. > > I doublechecked and we do build LLVM in release mode, on Windows and > elsewhere. We do build with assertions on all platforms, which is something > we can perhaps remove for speed, but it is useful. > > Another possible issue is that we use Python multiprocess code to compile > libraries (so we use all cores), and I've heard it's slower on windows for > various reasons. Perhaps there is something faster? I opened > https://github.com/emscripten-core/emscripten/issues/8719 > > - Alon > > > > On Fri, May 31, 2019 at 12:39 AM Floh <flo...@gmail.com <javascript:>> > wrote: > >> > About compiling libc which takes so long - is compiling your own code >> slower as well? >> >> Yes, compiling my own code is also slower, but I've seen this since >> forever for the same stuff compiled via clang on OSX/Linux vs Windows. I >> just started to accept it as some weird clang oddity ;) >> >> I haven't installed / activated emscripten SDKs so often in the past on >> Windows though, so I didn't notice this as much as now. >> >> On Friday, 31 May 2019 04:35:01 UTC+2, Alon Zakai wrote: >>> >>> About compiling libc which takes so long - is compiling your own code >>> slower as well? >>> >>> On Thu, May 30, 2019 at 10:54 AM Floh <flo...@gmail.com> wrote: >>> >>>> Would be interesting to see in a fresh install what happens when >>>>> building a hello world on windows, where the output isn't hidden. >>>>> >>>>>> >>>>>> >>>> It's libc which takes so long: >>>> >>>> cache:INFO: generating system library: libc.bc... (this will be cached >>>> in... >>>> >>>> -- >>>> You received this message because you are subscribed to the Google >>>> Groups "emscripten-discuss" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to emscripten-discuss+unsubscr...@googlegroups.com. >>>> To view this discussion on the web visit >>>> https://groups.google.com/d/msgid/emscripten-discuss/c678df7b-5330-4170-8e65-11c35a1d96b2%40googlegroups.com >>>> >>>> <https://groups.google.com/d/msgid/emscripten-discuss/c678df7b-5330-4170-8e65-11c35a1d96b2%40googlegroups.com?utm_medium=email&utm_source=footer> >>>> . >>>> For more options, visit https://groups.google.com/d/optout. >>>> >>> -- >> You received this message because you are subscribed to the Google Groups >> "emscripten-discuss" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to emscripten-discuss+unsubscr...@googlegroups.com <javascript:>. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/emscripten-discuss/e2632bd3-67cf-4364-b6c1-424faaae8be3%40googlegroups.com >> >> <https://groups.google.com/d/msgid/emscripten-discuss/e2632bd3-67cf-4364-b6c1-424faaae8be3%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> For more options, visit https://groups.google.com/d/optout. >> > -- You received this message because you are subscribed to the Google Groups "emscripten-discuss" group. To unsubscribe from this group and stop receiving emails from it, send an email to emscripten-discuss+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/emscripten-discuss/faa85c02-050c-48bd-9aa7-3732ad5675b2%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
