[llvm] [VPlan] Use VPWidenIntrinsicRecipe to support binary and unary operations with EVL-vectorization (PR #114205)

[Alexey Bataev via llvm-commits]

================
@@ -1267,7 +1267,7 @@ define void @test_fadd(ptr nocapture %a, ptr nocapture readonly %b) {
 ; IF-EVL-NEXT:    [[TMP14:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[TMP13]]
 ; IF-EVL-NEXT:    [[TMP15:%.*]] = getelementptr inbounds float, ptr [[TMP14]], i32 0
 ; IF-EVL-NEXT:    [[VP_OP_LOAD:%.*]] = call <vscale x 4 x float> @llvm.vp.load.nxv4f32.p0(ptr align 4 [[TMP15]], <vscale x 4 x i1> splat (i1 true), i32 [[TMP12]])
-; IF-EVL-NEXT:    [[VP_OP:%.*]] = call fast <vscale x 4 x float> @llvm.vp.fadd.nxv4f32(<vscale x 4 x float> [[VP_OP_LOAD]], <vscale x 4 x float> splat (float 3.000000e+00), <vscale x 4 x i1> splat (i1 true), i32 [[TMP12]])
+; IF-EVL-NEXT:    [[VP_OP:%.*]] = call <vscale x 4 x float> @llvm.vp.fadd.nxv4f32(<vscale x 4 x float> [[VP_OP_LOAD]], <vscale x 4 x float> splat (float 3.000000e+00), <vscale x 4 x i1> splat (i1 true), i32 [[TMP12]])
----------------
alexey-bataev wrote:

What about (removed) fast attribute here?

https://github.com/llvm/llvm-project/pull/114205