================
@@ -0,0 +1,102 @@
+; RUN: opt -S -dxil-op-lower %s | FileCheck %s
+
+target triple = "dxil-pc-shadermodel6.6-compute"
+
+declare void @scalar_user(float)
+declare void @vector_user(<4 x float>)
+
+define void @loadfloats() {
+ ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding
+ ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32
217, %dx.types.Handle [[BIND]]
+ %buffer = call target("dx.TypedBuffer", <4 x float>, 0, 0, 0)
+ @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f32_0_0_0(
+ i32 0, i32 0, i32 1, i32 0, i1 false)
+
+ ; The temporary casts should all have been cleaned up
+ ; CHECK-NOT: %dx.cast_handle
+
+ ; CHECK: [[DATA0:%.*]] = call %dx.types.ResRet.f32 @dx.op.bufferLoad.f32(i32
68, %dx.types.Handle [[HANDLE]], i32 0, i32 undef)
+ %data0 = call <4 x float> @llvm.dx.typedBufferLoad(
+ target("dx.TypedBuffer", <4 x float>, 0, 0, 0) %buffer, i32 0)
+
+ ; The extract order depends on the users, so don't enforce that here.
+ ; CHECK-DAG: extractvalue %dx.types.ResRet.f32 [[DATA0]], 0
+ %data0_0 = extractelement <4 x float> %data0, i32 0
+ ; CHECK-DAG: extractvalue %dx.types.ResRet.f32 [[DATA0]], 2
+ %data0_2 = extractelement <4 x float> %data0, i32 2
+
+ ; If all of the uses are extracts, we skip creating a vector
+ ; CHECK-NOT: insertelement
+ call void @scalar_user(float %data0_0)
+ call void @scalar_user(float %data0_2)
+
+ ; CHECK: [[DATA4:%.*]] = call %dx.types.ResRet.f32 @dx.op.bufferLoad.f32(i32
68, %dx.types.Handle [[HANDLE]], i32 4, i32 undef)
+ %data4 = call <4 x float> @llvm.dx.typedBufferLoad(
+ target("dx.TypedBuffer", <4 x float>, 0, 0, 0) %buffer, i32 4)
+
+ ; CHECK: extractvalue %dx.types.ResRet.f32 [[DATA4]], 0
+ ; CHECK: extractvalue %dx.types.ResRet.f32 [[DATA4]], 1
+ ; CHECK: extractvalue %dx.types.ResRet.f32 [[DATA4]], 2
+ ; CHECK: extractvalue %dx.types.ResRet.f32 [[DATA4]], 3
+ ; CHECK: insertelement <4 x float> undef
+ ; CHECK: insertelement <4 x float>
+ ; CHECK: insertelement <4 x float>
+ ; CHECK: insertelement <4 x float>
+ call void @vector_user(<4 x float> %data4)
+
+ ; CHECK: [[DATA12:%.*]] = call %dx.types.ResRet.f32
@dx.op.bufferLoad.f32(i32 68, %dx.types.Handle [[HANDLE]], i32 12, i32 undef)
+ %data12 = call <4 x float> @llvm.dx.typedBufferLoad(
----------------
bogner wrote:
In the current implementation it always returns a vec4. There are some
tradeoffs to this, which I discuss in llvm/wg-hlsl#54. Long story short, this
is probably going to change.
https://github.com/llvm/llvm-project/pull/104252
_______________________________________________
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
