I'm not sure on the landing status of SIMD in firefox, sorry. It might already be there, but it might also be behind a flag or such.
Add, subtract, etc. should work, see the simd tests for examples including that. I don't follow PNaCl closely but do talk to the devs, I can raise the topic. But since emscripten follows the JS SIMD API, which was derived from the Dart one, it might or might not be a match to the PNaCl one (which I haven't seen). - Alon On Sat, Mar 29, 2014 at 2:27 PM, Sören Balko <[email protected]> wrote: > Looks good! And hasn't Firefox very recently gained SIMD support? In other > words, wouldn't I benefit from SIMD hardware support when making use of the > intrinsics you added? > > Other than that, do I need to use the intrinsics you introduced or can I > also use plain math operators between vector types and emcc maps these to > the intrinsics? > > Btw, PNaCl will add SIMD support very soon (see here: > https://groups.google.com/forum/m/#!topic/native-client-discuss/CgLtRlGIDIU). > Any chance to co-ordinate the efforts, such that one can compile into > asm.js and PNaCl from the same source? I know, that's asking a lot... ;-) > > Soeren > > Am 30.03.2014 um 03:19 schrieb Alon Zakai <[email protected]>: > > See system/include/emscripten/vector.h, emmintrin.h and xmmintrin.h > > It implements the JS SIMD spec, which is a portable subset of SIMD stuff, > > https://github.com/johnmccutchan/ecmascript_simd > > JS engines don't optimize it yet, but hopefully soon. > > - Alon > > > > On Sat, Mar 29, 2014 at 2:32 AM, Soeren Balko <[email protected]> wrote: > >> I was wondering how SIMD is supported in fastcomp - is it the same as >> gcc's vector types with basic math operations? Or are specific >> platform-specific intrinsics (MMX, SSE, AVX, NEON and friends) supported >> too? The latter is probably not possible as it is inherently not portable. >> And are there any plans for an auto-vectorization feature? >> >> Soeren >> >> -- >> 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 [email protected]. >> For more options, visit https://groups.google.com/d/optout. >> > > -- > You received this message because you are subscribed to a topic in the > Google Groups "emscripten-discuss" group. > To unsubscribe from this topic, visit > https://groups.google.com/d/topic/emscripten-discuss/YypFwcfL16A/unsubscribe > . > To unsubscribe from this group and all its topics, send an email to > [email protected]. > > 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 [email protected]. > 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 [email protected]. For more options, visit https://groups.google.com/d/optout.
