Hi, I own Intel Parallel Studio 2016, but I find the compiler way too slow, unreliable and actually providing worse code than Clang, but the SVML (and IPP) is great, so I'm trying to use it via Agner Fog's VectorClass (https://www.agner.org/optimize/). Things like sine and cosine seem working, but pow (__svml_powf4) doesn't. It basically produces a pile of nonsense. __svml_powf8 is the same...
Could there be a problem with calling convention? (considering it has 2 arguments, while sin/cos have just one) Any other ideas?
I tried it in MSVC and Clang, in both this ended up with different results, in both cases wrong ones. I also tried the demo of the newest compiler 2019, same thing... The function definition is simple:
extern "C" { extern __m128 __svml_powf4 (__m128, __m128); };
And linked to svml_dispmt.lib, Windows (but I intend to use it on OSX as well).
