> Le 22 mars 2023 à 21:01, Oleg Nesterov <o...@redhat.com> a écrit : > > On 03/22, Stéphane Letz wrote: >> >> Those are purely experimental and internal. Like the gUseMemmove stuff BTW. > > OK. Thanks for the info. > > Yet I think an option for gUseMemmove makes sense. I didn't try to check, but > my gut feeling tells me gUseMemmove = true might speedup the code ;) Even if > not, it can make the generated code look more readable. Nevermind.
Well look at the generated code. I did, looking at LLVM IR, and some of the delay lines related loops got simply rewritten using LLVM memmove intrinsic (https://llvm.org/docs/LangRef.html#llvm-memmove-intrinsic) Stéphane _______________________________________________ Faudiostream-users mailing list Faudiostream-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/faudiostream-users
