search for: ext_vector_length

...store), or as little as 1 instruction if it’s inlined and the inputs are already in registers. But clang doesn’t vectorize it. Is there a simple way to get it to vectorize portably? I’d like to port to ARM NEON as well as SSE and AVX, and I’d like to be compatible with GCC, so I’d rather not use ext_vector_length if I can avoid it. If I can’t avoid it, I can cobble something together with a big vector_intrinsics.h file or something. You might think that vectorizing a 4-long loop wouldn’t matter, but of course field addition is inlined and called all the time. GCC 4.7.2 vectorizes the above code, but bail...

On Feb 17, 2014, at 6:38 PM, Stephen Checkoway <s at pahtak.org> wrote: > > On Feb 17, 2014, at 8:10 PM, Michael Hamburg <mike at shiftleft.org> wrote: > >> First, addition. I have multiprecision integer objects, and I’d like to add them component-wise (likewise, subtract, negate, mask…). For example: >> >> struct mp { >> int limb[8]; >> }