using FMA intrinsics

[bronxzv]

intrinsics for FMA AVX instructions are well explained in theC++ Composer XE 2011 documentation

Intrinsics Reference -> Intrinsics for Intel Advanced Vector Extensions -> Intrinsics for Fused Multiply Add Operations

though these intrinsics are missing in "immintrin.h" and all other such headers apparently, thus a simple question :how can we use these intrinsics ? (a compile flag ? other ?)

[TimP]

There's no AVX FMA support in current released Intel compilers, so it's strange to see them described in the documentation. Such compilers haven't even reached beta. There's fma support in beta gcc, I believe.

[bronxzv]

thanks for the information, though it looks rather odd since there is more than 20 pages in the user manual detailing FMA intrinsics