[llvm] [LV] Teach the vectorizer to cost and vectorize llvm.sincos intrinsics (PR #123210)

[Benjamin Maxwell via llvm-commits]

================
@@ -1,17 +0,0 @@
-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
-; RUN: opt %s -passes="function(scalarizer)" -S | FileCheck %s
-
-; Test to make sure that struct return intrinsics that are not `isTriviallyScalarizable` do not get scalarized.
-
-define <4 x float> @test_(<4 x float> %Val) {
-; CHECK-LABEL: define <4 x float> @test_(
-; CHECK-SAME: <4 x float> [[VAL:%.*]]) {
-; CHECK-NEXT:    [[R:%.*]] = call { <4 x float>, <4 x float> } @llvm.sincos.v4f32(<4 x float> [[VAL]])
-; CHECK-NEXT:    [[EL:%.*]] = extractvalue { <4 x float>, <4 x float> } [[R]], 0
-; CHECK-NEXT:    ret <4 x float> [[EL]]
-;
-  %r = call { <4 x float>, <4 x float> } @llvm.sincos.v4f32(<4 x float>  %Val)
-  %el = extractvalue { <4 x float>, <4 x float> } %r, 0
----------------
MacDue wrote:

Note: I replaced this with an equivalent `llvm.vector.deinterleave2` to test (which matches the intent of this test -- checking non-trivially-scalarizable struct-ret intrinsics are not scalarized). 

https://github.com/llvm/llvm-project/pull/123210