[clang] Implement the `fmod` intrinsic (PR #130320)

[Farzon Lotfi via cfe-commits]

================
@@ -22,56 +22,136 @@
 //
 // RUN: %clang_cc1 -finclude-default-header -x hlsl -triple \
 // RUN:   spirv-unknown-vulkan-compute %s -fnative-half-type \
-// RUN:   -emit-llvm -disable-llvm-passes -o - | FileCheck %s \
+// RUN:   -emit-llvm -o - | FileCheck %s \
 // RUN:   -DFNATTRS="spir_func noundef nofpclass(nan inf)" -DTYPE=half
 
 //
 // ---------- No Native Half support test -----------
 //
 // RUN: %clang_cc1 -finclude-default-header -x hlsl -triple \
-// RUN:   spirv-unknown-vulkan-compute %s -emit-llvm -disable-llvm-passes \
+// RUN:   spirv-unknown-vulkan-compute %s -emit-llvm \
 // RUN:   -o - | FileCheck %s \
 // RUN:   -DFNATTRS="spir_func noundef nofpclass(nan inf)" -DTYPE=float
 
 
 
+// DXCHECK: define [[FNATTRS]] [[TYPE]] @
+// DXCHECK: %div1.i = fdiv reassoc nnan ninf nsz arcp afn [[TYPE]] %4, %5
+// DXCHECK: %fneg.i = fneg reassoc nnan ninf nsz arcp afn [[TYPE]] %7
+// DXCHECK: %cmp.i = fcmp reassoc nnan ninf nsz arcp afn oge [[TYPE]] %6, %fneg.i
+// DXCHECK: %elt.abs.i = call reassoc nnan ninf nsz arcp afn [[TYPE]] @llvm.fabs.[[INT_TYPE]]([[TYPE]] %8)
+// DXCHECK: %hlsl.frac.i = call reassoc nnan ninf nsz arcp afn [[TYPE]] @llvm.dx.frac.[[INT_TYPE]]([[TYPE]] %elt.abs.i)
+// DXCHECK: %fneg2.i = fneg reassoc nnan ninf nsz arcp afn [[TYPE]] %11
+// DXCHECK: %hlsl.select.i = select reassoc nnan ninf nsz arcp afn i1 %loadedv.i, [[TYPE]] %10, [[TYPE]] %fneg2.i
+// DXCHECK: %mul.i = fmul reassoc nnan ninf nsz arcp afn [[TYPE]] %hlsl.select.i, %12
+// DXCHECK: ret [[TYPE]] %mul.i
 // CHECK: define [[FNATTRS]] [[TYPE]] @
-// CHECK: %fmod = frem reassoc nnan ninf nsz arcp afn [[TYPE]]
-// CHECK: ret [[TYPE]] %fmod
+// CHECK: %fmod.i = frem reassoc nnan ninf nsz arcp afn [[TYPE]]
+// CHECK: ret [[TYPE]] %fmod.i
 half test_fmod_half(half p0, half p1) { return fmod(p0, p1); }
 
+// DXCHECK: define [[FNATTRS]] <2 x [[TYPE]]> @
----------------
farzonl wrote:

Why would we duplicate something we are already checking  just to force `*-LABEL`? I don't think we need to be so rigid especially in this case where the `fmod.hlsl` already has a set pattern for how this test should be. You are not going to find a one size fits all approach and its best to keep the current pattern.

https://github.com/llvm/llvm-project/pull/130320