search for: altneg

...nted fmul. I certainly agree that this is less than ideal. Adding a new intrinsic would be better. I would like to add two, one would be something that adds the odd elements and subtracts the even elements, maybe: @llvm.addsub.* And another is one which negates only the even elements, maybe @llvm.altneg.* is a good name? (actually, only this one is really necessary, but would we like both for clarity or some other reason?) In combination with permutations, this should allow the encoding of the full range of asymmetric-fma-type operations (some of which can be directly represented on some platform...

On Oct 17, 2011, at 6:40 PM, Hal Finkel wrote: > On Mon, 2011-10-17 at 17:33 -0700, Dan Gohman wrote: >> On Oct 17, 2011, at 3:40 PM, Hal Finkel wrote: >> >>> How should I go about matching floating-point addsub-like vector >>> instructions? My first inclination is to write something which matches >>> build_vector 1.0, -1.0, and then use that in

On Mon, 2011-10-17 at 17:33 -0700, Dan Gohman wrote: > On Oct 17, 2011, at 3:40 PM, Hal Finkel wrote: > > > How should I go about matching floating-point addsub-like vector > > instructions? My first inclination is to write something which matches > > build_vector 1.0, -1.0, and then use that in combination with a match on > > fadd, but that does not seem to work. I