Merge branch 'main' into users/meinersbur/flang_runtime_split-headersusers/meinersbur/flang_runtime_split-headers

19 files changed, 751 insertions, 111 deletions

diff --git a/llvm/test/CodeGen/DirectX/BufferLoad.ll b/llvm/test/CodeGen/DirectX/BufferLoad.ll
index 24d65fe1648c..7f1291bf4a5c 100644
--- a/llvm/test/CodeGen/DirectX/BufferLoad.ll
+++ b/llvm/test/CodeGen/DirectX/BufferLoad.ll
@@ -10,14 +10,14 @@ define void @loadv4f32() {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %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(
+      @llvm.dx.resource.handlefrombinding.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-NOT: %dx.resource.casthandle
 
   ; 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(
+  %data0 = call <4 x float> @llvm.dx.resource.load.typedbuffer(
       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.
@@ -34,7 +34,7 @@ define void @loadv4f32() {
   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(
+  %data4 = call <4 x float> @llvm.dx.resource.load.typedbuffer(
       target("dx.TypedBuffer", <4 x float>, 0, 0, 0) %buffer, i32 4)
 
   ; CHECK: extractvalue %dx.types.ResRet.f32 [[DATA4]], 0
@@ -48,7 +48,7 @@ define void @loadv4f32() {
   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(
+  %data12 = call <4 x float> @llvm.dx.resource.load.typedbuffer(
       target("dx.TypedBuffer", <4 x float>, 0, 0, 0) %buffer, i32 12)
 
   ; CHECK: [[DATA12_3:%.*]] = extractvalue %dx.types.ResRet.f32 [[DATA12]], 3
@@ -66,11 +66,11 @@ define void @index_dynamic(i32 %bufindex, i32 %elemindex) {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %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(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_0_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
 
   ; CHECK: [[LOAD:%.*]] = call %dx.types.ResRet.f32 @dx.op.bufferLoad.f32(i32 68, %dx.types.Handle [[HANDLE]], i32 %bufindex, i32 undef)
-  %load = call <4 x float> @llvm.dx.typedBufferLoad(
+  %load = call <4 x float> @llvm.dx.resource.load.typedbuffer(
       target("dx.TypedBuffer", <4 x float>, 0, 0, 0) %buffer, i32 %bufindex)
 
   ; CHECK: [[ALLOCA:%.*]] = alloca [4 x float]
@@ -101,11 +101,11 @@ define void @loadf32() {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
   %buffer = call target("dx.TypedBuffer", float, 0, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_f32_0_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_f32_0_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
 
   ; CHECK: [[DATA0:%.*]] = call %dx.types.ResRet.f32 @dx.op.bufferLoad.f32(i32 68, %dx.types.Handle [[HANDLE]], i32 0, i32 undef)
-  %data0 = call float @llvm.dx.typedBufferLoad(
+  %data0 = call float @llvm.dx.resource.load.typedbuffer(
       target("dx.TypedBuffer", float, 0, 0, 0) %buffer, i32 0)
 
   ; CHECK: [[VAL0:%.*]] = extractvalue %dx.types.ResRet.f32 [[DATA0]], 0
@@ -119,11 +119,11 @@ define void @loadv2f32() {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
   %buffer = call target("dx.TypedBuffer", <2 x float>, 0, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v2f32_0_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v2f32_0_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
 
   ; CHECK: [[DATA0:%.*]] = call %dx.types.ResRet.f32 @dx.op.bufferLoad.f32(i32 68, %dx.types.Handle [[HANDLE]], i32 0, i32 undef)
-  %data0 = call <2 x float> @llvm.dx.typedBufferLoad(
+  %data0 = call <2 x float> @llvm.dx.resource.load.typedbuffer(
       target("dx.TypedBuffer", <2 x float>, 0, 0, 0) %buffer, i32 0)
 
   ret void
@@ -133,11 +133,11 @@ define void @loadv4f32_checkbit() {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %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(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_0_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
 
   ; 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>, i1} @llvm.dx.typedBufferLoad.checkbit.f32(
+  %data0 = call {<4 x float>, i1} @llvm.dx.resource.loadchecked.typedbuffer.f32(
       target("dx.TypedBuffer", <4 x float>, 0, 0, 0) %buffer, i32 0)
 
   ; CHECK: [[STATUS:%.*]] = extractvalue %dx.types.ResRet.f32 [[DATA0]], 4
@@ -154,11 +154,11 @@ define void @loadv4i32() {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
   %buffer = call target("dx.TypedBuffer", <4 x i32>, 0, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4i32_0_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4i32_0_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
 
   ; CHECK: [[DATA0:%.*]] = call %dx.types.ResRet.i32 @dx.op.bufferLoad.i32(i32 68, %dx.types.Handle [[HANDLE]], i32 0, i32 undef)
-  %data0 = call <4 x i32> @llvm.dx.typedBufferLoad(
+  %data0 = call <4 x i32> @llvm.dx.resource.load.typedbuffer(
       target("dx.TypedBuffer", <4 x i32>, 0, 0, 0) %buffer, i32 0)
 
   ret void
@@ -168,11 +168,11 @@ define void @loadv4f16() {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
   %buffer = call target("dx.TypedBuffer", <4 x half>, 0, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f16_0_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f16_0_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
 
   ; CHECK: [[DATA0:%.*]] = call %dx.types.ResRet.f16 @dx.op.bufferLoad.f16(i32 68, %dx.types.Handle [[HANDLE]], i32 0, i32 undef)
-  %data0 = call <4 x half> @llvm.dx.typedBufferLoad(
+  %data0 = call <4 x half> @llvm.dx.resource.load.typedbuffer(
       target("dx.TypedBuffer", <4 x half>, 0, 0, 0) %buffer, i32 0)
 
   ret void
@@ -182,11 +182,11 @@ define void @loadv4i16() {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
   %buffer = call target("dx.TypedBuffer", <4 x i16>, 0, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4i16_0_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4i16_0_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
 
   ; CHECK: [[DATA0:%.*]] = call %dx.types.ResRet.i16 @dx.op.bufferLoad.i16(i32 68, %dx.types.Handle [[HANDLE]], i32 0, i32 undef)
-  %data0 = call <4 x i16> @llvm.dx.typedBufferLoad(
+  %data0 = call <4 x i16> @llvm.dx.resource.load.typedbuffer(
       target("dx.TypedBuffer", <4 x i16>, 0, 0, 0) %buffer, i32 0)
 
   ret void
diff --git a/llvm/test/CodeGen/DirectX/BufferStore-errors.ll b/llvm/test/CodeGen/DirectX/BufferStore-errors.ll
index ffdcb037b896..6e529973bd60 100644
--- a/llvm/test/CodeGen/DirectX/BufferStore-errors.ll
+++ b/llvm/test/CodeGen/DirectX/BufferStore-errors.ll
@@ -8,10 +8,10 @@ target triple = "dxil-pc-shadermodel6.6-compute"
 ; CHECK-SAME: typedBufferStore data must be a vector of 4 elements
 define void @storetoomany(<5 x float> %data, i32 %index) "hlsl.export" {
   %buffer = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f32_1_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_1_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
 
-  call void @llvm.dx.typedBufferStore.tdx.TypedBuffer_v4f32_1_0_0t.v5f32(
+  call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4f32_1_0_0t.v5f32(
       target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer,
       i32 %index, <5 x float> %data)
 
@@ -23,15 +23,15 @@ define void @storetoomany(<5 x float> %data, i32 %index) "hlsl.export" {
 ; CHECK-SAME: typedBufferStore data must be a vector of 4 elements
 define void @storetoofew(<3 x i32> %data, i32 %index) "hlsl.export" {
   %buffer = call target("dx.TypedBuffer", <4 x i32>, 1, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4i32_1_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4i32_1_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
 
-  call void @llvm.dx.typedBufferStore.tdx.TypedBuffer_v4i32_1_0_0t.v3i32(
+  call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4i32_1_0_0t.v3i32(
       target("dx.TypedBuffer", <4 x i32>, 1, 0, 0) %buffer,
       i32 %index, <3 x i32> %data)
 
   ret void
 }
 
-declare void @llvm.dx.typedBufferStore.tdx.TypedBuffer_v4f32_1_0_0t.v5f32(target("dx.TypedBuffer", <4 x float>, 1, 0, 0), i32, <5 x float>)
-declare void @llvm.dx.typedBufferStore.tdx.TypedBuffer_v4i32_1_0_0t.v3i32(target("dx.TypedBuffer", <4 x i32>, 1, 0, 0), i32, <3 x i32>)
+declare void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4f32_1_0_0t.v5f32(target("dx.TypedBuffer", <4 x float>, 1, 0, 0), i32, <5 x float>)
+declare void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4i32_1_0_0t.v3i32(target("dx.TypedBuffer", <4 x i32>, 1, 0, 0), i32, <3 x i32>)
diff --git a/llvm/test/CodeGen/DirectX/BufferStore.ll b/llvm/test/CodeGen/DirectX/BufferStore.ll
index 81cc5fd328e0..381df6a63962 100644
--- a/llvm/test/CodeGen/DirectX/BufferStore.ll
+++ b/llvm/test/CodeGen/DirectX/BufferStore.ll
@@ -7,18 +7,18 @@ define void @storefloat(<4 x float> %data, i32 %index) {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
   %buffer = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f32_1_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_1_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-NOT: %dx.resource.casthandle
 
   ; CHECK: [[DATA0_0:%.*]] = extractelement <4 x float> %data, i32 0
   ; CHECK: [[DATA0_1:%.*]] = extractelement <4 x float> %data, i32 1
   ; CHECK: [[DATA0_2:%.*]] = extractelement <4 x float> %data, i32 2
   ; CHECK: [[DATA0_3:%.*]] = extractelement <4 x float> %data, i32 3
   ; CHECK: call void @dx.op.bufferStore.f32(i32 69, %dx.types.Handle [[HANDLE]], i32 %index, i32 undef, float [[DATA0_0]], float [[DATA0_1]], float [[DATA0_2]], float [[DATA0_3]], i8 15)
-  call void @llvm.dx.typedBufferStore(
+  call void @llvm.dx.resource.store.typedbuffer(
       target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer,
       i32 %index, <4 x float> %data)
 
@@ -30,7 +30,7 @@ define void @storeint(<4 x i32> %data, i32 %index) {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
   %buffer = call target("dx.TypedBuffer", <4 x i32>, 1, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4i32_1_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4i32_1_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
 
   ; CHECK: [[DATA0_0:%.*]] = extractelement <4 x i32> %data, i32 0
@@ -38,7 +38,7 @@ define void @storeint(<4 x i32> %data, i32 %index) {
   ; CHECK: [[DATA0_2:%.*]] = extractelement <4 x i32> %data, i32 2
   ; CHECK: [[DATA0_3:%.*]] = extractelement <4 x i32> %data, i32 3
   ; CHECK: call void @dx.op.bufferStore.i32(i32 69, %dx.types.Handle [[HANDLE]], i32 %index, i32 undef, i32 [[DATA0_0]], i32 [[DATA0_1]], i32 [[DATA0_2]], i32 [[DATA0_3]], i8 15)
-  call void @llvm.dx.typedBufferStore(
+  call void @llvm.dx.resource.store.typedbuffer(
       target("dx.TypedBuffer", <4 x i32>, 1, 0, 0) %buffer,
       i32 %index, <4 x i32> %data)
 
@@ -50,18 +50,18 @@ define void @storehalf(<4 x half> %data, i32 %index) {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
   %buffer = call target("dx.TypedBuffer", <4 x half>, 1, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f16_1_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f16_1_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-NOT: %dx.resource.casthandle
 
   ; CHECK: [[DATA0_0:%.*]] = extractelement <4 x half> %data, i32 0
   ; CHECK: [[DATA0_1:%.*]] = extractelement <4 x half> %data, i32 1
   ; CHECK: [[DATA0_2:%.*]] = extractelement <4 x half> %data, i32 2
   ; CHECK: [[DATA0_3:%.*]] = extractelement <4 x half> %data, i32 3
   ; CHECK: call void @dx.op.bufferStore.f16(i32 69, %dx.types.Handle [[HANDLE]], i32 %index, i32 undef, half [[DATA0_0]], half [[DATA0_1]], half [[DATA0_2]], half [[DATA0_3]], i8 15)
-  call void @llvm.dx.typedBufferStore(
+  call void @llvm.dx.resource.store.typedbuffer(
       target("dx.TypedBuffer", <4 x half>, 1, 0, 0) %buffer,
       i32 %index, <4 x half> %data)
 
@@ -73,18 +73,18 @@ define void @storei16(<4 x i16> %data, i32 %index) {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
   %buffer = call target("dx.TypedBuffer", <4 x i16>, 1, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4i16_1_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4i16_1_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-NOT: %dx.resource.casthandle
 
   ; CHECK: [[DATA0_0:%.*]] = extractelement <4 x i16> %data, i32 0
   ; CHECK: [[DATA0_1:%.*]] = extractelement <4 x i16> %data, i32 1
   ; CHECK: [[DATA0_2:%.*]] = extractelement <4 x i16> %data, i32 2
   ; CHECK: [[DATA0_3:%.*]] = extractelement <4 x i16> %data, i32 3
   ; CHECK: call void @dx.op.bufferStore.i16(i32 69, %dx.types.Handle [[HANDLE]], i32 %index, i32 undef, i16 [[DATA0_0]], i16 [[DATA0_1]], i16 [[DATA0_2]], i16 [[DATA0_3]], i8 15)
-  call void @llvm.dx.typedBufferStore(
+  call void @llvm.dx.resource.store.typedbuffer(
       target("dx.TypedBuffer", <4 x i16>, 1, 0, 0) %buffer,
       i32 %index, <4 x i16> %data)
 
@@ -96,7 +96,7 @@ define void @store_scalarized_floats(float %data0, float %data1, float %data2, f
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217,
   ; CHECK: [[HANDLE:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
   %buffer = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f32_1_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_1_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
 
   ; We shouldn't end up with any inserts/extracts.
@@ -108,7 +108,7 @@ define void @store_scalarized_floats(float %data0, float %data1, float %data2, f
   %vec.upto1 = insertelement <4 x float> %vec.upto0, float %data1, i64 1
   %vec.upto2 = insertelement <4 x float> %vec.upto1, float %data2, i64 2
   %vec = insertelement <4 x float> %vec.upto2, float %data3, i64 3
-  call void @llvm.dx.typedBufferStore(
+  call void @llvm.dx.resource.store.typedbuffer(
       target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer,
       i32 %index, <4 x float> %vec)
 
diff --git a/llvm/test/CodeGen/DirectX/ContainerData/PSVResources.ll b/llvm/test/CodeGen/DirectX/ContainerData/PSVResources.ll
index 2bd7a2e8df12..ce67812c3988 100644
--- a/llvm/test/CodeGen/DirectX/ContainerData/PSVResources.ll
+++ b/llvm/test/CodeGen/DirectX/ContainerData/PSVResources.ll
@@ -15,7 +15,7 @@ define void @main() #0 {
 ; CHECK:          Flags:
 ; CHECK:            UsedByAtomic64:  false
   %srv0 = call target("dx.RawBuffer", i8, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.RawBuffer_i8_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.RawBuffer_i8_0_0t(
           i32 1, i32 8, i32 1, i32 0, i1 false)
 
   ; struct S { float4 a; uint4 b; };
@@ -28,7 +28,7 @@ define void @main() #0 {
 ; CHECK:          Flags:
 ; CHECK:            UsedByAtomic64:  false
   %srv1 = call target("dx.RawBuffer", {<4 x float>, <4 x i32>}, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.RawBuffer_sl_v4f32v4i32s_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.RawBuffer_sl_v4f32v4i32s_0_0t(
           i32 4, i32 2, i32 1, i32 0, i1 false)
 
   ; Buffer<uint4> Buf[24] : register(t3, space5)
@@ -40,7 +40,7 @@ define void @main() #0 {
 ; CHECK:          Flags:
 ; CHECK:            UsedByAtomic64:  false
   %srv2 = call target("dx.TypedBuffer", <4 x i32>, 0, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_i32_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_i32_0_0t(
           i32 5, i32 3, i32 24, i32 0, i1 false)
 
   ; RWBuffer<int> Buf : register(u7, space2)
@@ -52,7 +52,7 @@ define void @main() #0 {
 ; CHECK:          Flags:
 ; CHECK:            UsedByAtomic64:  false
   %uav0 = call target("dx.TypedBuffer", i32, 1, 0, 1)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_i32_1_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_i32_1_0t(
           i32 2, i32 7, i32 1, i32 0, i1 false)
 
   ; RWBuffer<float4> Buf : register(u5, space3)
@@ -64,7 +64,7 @@ define void @main() #0 {
 ; CHECK:          Flags:
 ; CHECK:            UsedByAtomic64:  false
   %uav1 = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0)
-              @llvm.dx.handle.fromBinding.tdx.TypedBuffer_f32_1_0(
+              @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_f32_1_0(
                   i32 3, i32 5, i32 1, i32 0, i1 false)
 
   ; RWBuffer<float4> BufferArray[10] : register(u0, space4)
@@ -77,11 +77,11 @@ define void @main() #0 {
 ; CHECK:            UsedByAtomic64:  false
   ; RWBuffer<float4> Buf = BufferArray[0]
   %uav2_1 = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0)
-              @llvm.dx.handle.fromBinding.tdx.TypedBuffer_f32_1_0(
+              @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_f32_1_0(
                   i32 4, i32 0, i32 10, i32 0, i1 false)
   ; RWBuffer<float4> Buf = BufferArray[5]
   %uav2_2 = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0)
-              @llvm.dx.handle.fromBinding.tdx.TypedBuffer_f32_1_0(
+              @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_f32_1_0(
                   i32 4, i32 0, i32 10, i32 5, i1 false)
   ret void
 }
diff --git a/llvm/test/CodeGen/DirectX/CreateHandle.ll b/llvm/test/CodeGen/DirectX/CreateHandle.ll
index 234d4e035bf1..80daa879f0f8 100644
--- a/llvm/test/CodeGen/DirectX/CreateHandle.ll
+++ b/llvm/test/CodeGen/DirectX/CreateHandle.ll
@@ -1,4 +1,4 @@
-; RUN: opt -S -passes=dxil-op-lower,dxil-translate-metadata %s | FileCheck %s
+; RUN: opt -S -passes=dxil-translate-metadata,dxil-op-lower %s | FileCheck %s
 ; RUN: opt -S -passes=dxil-pretty-printer %s 2>&1 >/dev/null | FileCheck --check-prefix=CHECK-PRETTY %s
 
 ; CHECK-PRETTY:       Type  Format         Dim      ID      HLSL Bind     Count
@@ -17,14 +17,14 @@ declare i32 @some_val();
 define void @test_buffers() {
   ; RWBuffer<float4> Buf : register(u5, space3)
   %typed0 = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0)
-              @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f32_1_0_0(
+              @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_1_0_0(
                   i32 3, i32 5, i32 1, i32 0, i1 false)
   ; CHECK: call %dx.types.Handle @dx.op.createHandle(i32 57, i8 1, i32 1, i32 5, i1 false)
   ; CHECK-NOT: @llvm.dx.cast.handle
 
   ; RWBuffer<int> Buf : register(u7, space2)
   %typed1 = call target("dx.TypedBuffer", i32, 1, 0, 1)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_i32_1_0_1t(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_i32_1_0_1t(
           i32 2, i32 7, i32 1, i32 0, i1 false)
   ; CHECK: call %dx.types.Handle @dx.op.createHandle(i32 57, i8 1, i32 0, i32 7, i1 false)
 
@@ -32,20 +32,20 @@ define void @test_buffers() {
   ; Buffer<uint4> typed2 = Buf[4]
   ; Note that the index below is 3 + 4 = 7
   %typed2 = call target("dx.TypedBuffer", <4 x i32>, 0, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_i32_0_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_i32_0_0_0t(
           i32 5, i32 3, i32 24, i32 4, i1 false)
   ; CHECK: call %dx.types.Handle @dx.op.createHandle(i32 57, i8 0, i32 3, i32 7, i1 false)
 
   ; struct S { float4 a; uint4 b; };
   ; StructuredBuffer<S> Buf : register(t2, space4)
   %struct0 = call target("dx.RawBuffer", {<4 x float>, <4 x i32>}, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.RawBuffer_sl_v4f32v4i32s_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.RawBuffer_sl_v4f32v4i32s_0_0t(
           i32 4, i32 2, i32 1, i32 0, i1 true)
   ; CHECK: call %dx.types.Handle @dx.op.createHandle(i32 57, i8 0, i32 2, i32 2, i1 true)
 
   ; ByteAddressBuffer Buf : register(t8, space1)
   %byteaddr0 = call target("dx.RawBuffer", i8, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.RawBuffer_i8_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.RawBuffer_i8_0_0t(
           i32 1, i32 8, i32 1, i32 0, i1 false)
   ; CHECK: call %dx.types.Handle @dx.op.createHandle(i32 57, i8 0, i32 1, i32 8, i1 false)
 
@@ -53,7 +53,7 @@ define void @test_buffers() {
   ; Buffer<float4> typed3 = Buf[ix]
   %typed3_ix = call i32 @some_val()
   %typed3 = call target("dx.TypedBuffer", <4 x float>, 0, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f32_0_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_0_0_0t(
           i32 0, i32 7, i32 -1, i32 %typed3_ix, i1 false)
   ; CHECK: %[[IX:.*]] = add i32 %typed3_ix, 7
   ; CHECK: call %dx.types.Handle @dx.op.createHandle(i32 57, i8 0, i32 0, i32 %[[IX]], i1 false)
diff --git a/llvm/test/CodeGen/DirectX/CreateHandleFromBinding.ll b/llvm/test/CodeGen/DirectX/CreateHandleFromBinding.ll
index aa143dfa8211..bf11bfa143c9 100644
--- a/llvm/test/CodeGen/DirectX/CreateHandleFromBinding.ll
+++ b/llvm/test/CodeGen/DirectX/CreateHandleFromBinding.ll
@@ -1,4 +1,4 @@
-; RUN: opt -S -passes=dxil-op-lower,dxil-translate-metadata %s | FileCheck %s
+; RUN: opt -S -passes=dxil-translate-metadata,dxil-op-lower %s | FileCheck %s
 ; RUN: opt -S -passes=dxil-pretty-printer %s 2>&1 >/dev/null | FileCheck --check-prefix=CHECK-PRETTY %s
 
 ; CHECK-PRETTY:       Type  Format         Dim      ID      HLSL Bind     Count
@@ -17,14 +17,14 @@ declare i32 @some_val();
 define void @test_bindings() {
   ; RWBuffer<float4> Buf : register(u5, space3)
   %typed0 = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0)
-              @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f32_1_0_0(
+              @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_1_0_0(
                   i32 3, i32 5, i32 1, i32 0, i1 false)
   ; CHECK: [[BUF0:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217, %dx.types.ResBind { i32 5, i32 5, i32 3, i8 1 }, i32 5, i1 false)
   ; CHECK: call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BUF0]], %dx.types.ResourceProperties { i32 4106, i32 1033 })
 
   ; RWBuffer<int> Buf : register(u7, space2)
   %typed1 = call target("dx.TypedBuffer", i32, 1, 0, 1)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_i32_1_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_i32_1_0_0t(
           i32 2, i32 7, i32 1, i32 0, i1 false)
   ; CHECK: [[BUF1:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217, %dx.types.ResBind { i32 7, i32 7, i32 2, i8 1 }, i32 7, i1 false)
   ; CHECK: call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BUF1]], %dx.types.ResourceProperties { i32 4106, i32 260 })
@@ -33,7 +33,7 @@ define void @test_bindings() {
   ; Buffer<uint4> typed2 = Buf[4]
   ; Note that the index below is 3 + 4 = 7
   %typed2 = call target("dx.TypedBuffer", <4 x i32>, 0, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_i32_0_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_i32_0_0_0t(
           i32 5, i32 3, i32 24, i32 4, i1 false)
   ; CHECK: [[BUF2:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217, %dx.types.ResBind { i32 3, i32 26, i32 5, i8 0 }, i32 7, i1 false)
   ; CHECK: call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BUF2]], %dx.types.ResourceProperties { i32 10, i32 1029 })
@@ -41,14 +41,14 @@ define void @test_bindings() {
   ; struct S { float4 a; uint4 b; };
   ; StructuredBuffer<S> Buf : register(t2, space4)
   %struct0 = call target("dx.RawBuffer", {<4 x float>, <4 x i32>}, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.RawBuffer_sl_v4f32v4i32s_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.RawBuffer_sl_v4f32v4i32s_0_0t(
           i32 4, i32 2, i32 1, i32 0, i1 true)
   ; CHECK: [[BUF3:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217, %dx.types.ResBind { i32 2, i32 2, i32 4, i8 0 }, i32 2, i1 true)
   ; CHECK: = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BUF3]], %dx.types.ResourceProperties { i32 1036, i32 32 })
 
   ; ByteAddressBuffer Buf : register(t8, space1)
   %byteaddr0 = call target("dx.RawBuffer", i8, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.RawBuffer_i8_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.RawBuffer_i8_0_0t(
           i32 1, i32 8, i32 1, i32 0, i1 false)
   ; CHECK: [[BUF4:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217, %dx.types.ResBind { i32 8, i32 8, i32 1, i8 0 }, i32 8, i1 false)
   ; CHECK: call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BUF4]], %dx.types.ResourceProperties { i32 11, i32 0 })
@@ -57,7 +57,7 @@ define void @test_bindings() {
   ; Buffer<float4> typed3 = Buf[ix]
   %typed3_ix = call i32 @some_val()
   %typed3 = call target("dx.TypedBuffer", <4 x float>, 0, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f32_0_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_0_0_0t(
           i32 0, i32 7, i32 -1, i32 %typed3_ix, i1 false)
   ; CHECK: %[[IX:.*]] = add i32 %typed3_ix, 7
   ; CHECK: [[BUF5:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217, %dx.types.ResBind { i32 7, i32 -1, i32 0, i8 0 }, i32 %[[IX]], i1 false)
diff --git a/llvm/test/CodeGen/DirectX/Metadata/resource-symbols.ll b/llvm/test/CodeGen/DirectX/Metadata/resource-symbols.ll
new file mode 100644
index 000000000000..6ebe4b5eb23a
--- /dev/null
+++ b/llvm/test/CodeGen/DirectX/Metadata/resource-symbols.ll
@@ -0,0 +1,48 @@
+; RUN: opt -S -passes=dxil-translate-metadata %s | FileCheck %s
+
+target triple = "dxil-pc-shadermodel6.6-compute"
+
+%struct.S = type { <4 x float>, <4 x i32> }
+
+define void @test() {
+  ; Buffer<float4>
+  %float4 = call target("dx.TypedBuffer", <4 x float>, 0, 0, 0)
+      @llvm.dx.resource.handlefrombinding(i32 0, i32 0, i32 1, i32 0, i1 false)
+  ; CHECK: %TypedBuffer = type { <4 x float> }
+
+  ; Buffer<int>
+  %int = call target("dx.TypedBuffer", i32, 0, 0, 1)
+      @llvm.dx.resource.handlefrombinding(i32 0, i32 1, i32 1, i32 0, i1 false)
+  ; CHECK: %TypedBuffer.0 = type { i32 }
+
+  ; Buffer<uint3>
+  %uint3 = call target("dx.TypedBuffer", <3 x i32>, 0, 0, 0)
+      @llvm.dx.resource.handlefrombinding(i32 0, i32 2, i32 1, i32 0, i1 false)
+  ; CHECK: %TypedBuffer.1 = type { <3 x i32> }
+
+  ; StructuredBuffer<S>
+  %struct0 = call target("dx.RawBuffer", %struct.S, 0, 0)
+      @llvm.dx.resource.handlefrombinding(i32 0, i32 10, i32 1, i32 0, i1 true)
+  ; CHECK: %StructuredBuffer = type { %struct.S }
+
+  ; ByteAddressBuffer
+  %byteaddr = call target("dx.RawBuffer", i8, 0, 0)
+      @llvm.dx.resource.handlefrombinding(i32 0, i32 20, i32 1, i32 0, i1 false)
+  ; CHECK: %ByteAddressBuffer = type { i32 }
+
+  ret void
+}
+
+; CHECK:      @[[T0:.*]] = external constant %TypedBuffer
+; CHECK-NEXT: @[[T1:.*]] = external constant %TypedBuffer.0
+; CHECK-NEXT: @[[T2:.*]] = external constant %TypedBuffer.1
+; CHECK-NEXT: @[[S0:.*]] = external constant %StructuredBuffer
+; CHECK-NEXT: @[[B0:.*]] = external constant %ByteAddressBuffer
+
+; CHECK: !{i32 0, ptr @[[T0]], !""
+; CHECK: !{i32 1, ptr @[[T1]], !""
+; CHECK: !{i32 2, ptr @[[T2]], !""
+; CHECK: !{i32 3, ptr @[[S0]], !""
+; CHECK: !{i32 4, ptr @[[B0]], !""
+
+attributes #0 = { nocallback nofree nosync nounwind willreturn memory(none) }
diff --git a/llvm/test/CodeGen/DirectX/ResourceAccess/load_typedbuffer.ll b/llvm/test/CodeGen/DirectX/ResourceAccess/load_typedbuffer.ll
new file mode 100644
index 000000000000..9b7e7fd04f60
--- /dev/null
+++ b/llvm/test/CodeGen/DirectX/ResourceAccess/load_typedbuffer.ll
@@ -0,0 +1,35 @@
+; RUN: opt -S -dxil-resource-access %s | FileCheck %s
+
+target triple = "dxil-pc-shadermodel6.6-compute"
+
+declare void @use_float4(<4 x float>)
+declare void @use_float(float)
+
+; CHECK-LABEL: define void @load_float4
+define void @load_float4(i32 %index, i32 %elemindex) {
+  %buffer = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0)
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_1_0_0(
+          i32 0, i32 0, i32 1, i32 0, i1 false)
+
+  ; CHECK-NOT: @llvm.dx.resource.getpointer
+  %ptr = call ptr @llvm.dx.resource.getpointer(
+      target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer, i32 %index)
+
+  ; CHECK: %[[VALUE:.*]] = call <4 x float> @llvm.dx.resource.load.typedbuffer.v4f32.tdx.TypedBuffer_v4f32_1_0_0t(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer, i32 %index)
+  %vec_data = load <4 x float>, ptr %ptr
+  call void @use_float4(<4 x float> %vec_data)
+
+  ; CHECK: %[[VALUE:.*]] = call <4 x float> @llvm.dx.resource.load.typedbuffer.v4f32.tdx.TypedBuffer_v4f32_1_0_0t(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer, i32 %index)
+  ; CHECK: extractelement <4 x float> %[[VALUE]], i32 1
+  %y_ptr = getelementptr inbounds <4 x float>, ptr %ptr, i32 0, i32 1
+  %y_data = load float, ptr %y_ptr
+  call void @use_float(float %y_data)
+
+  ; CHECK: %[[VALUE:.*]] = call <4 x float> @llvm.dx.resource.load.typedbuffer.v4f32.tdx.TypedBuffer_v4f32_1_0_0t(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer, i32 %index)
+  ; CHECK: extractelement <4 x float> %[[VALUE]], i32 %elemindex
+  %dynamic = getelementptr inbounds <4 x float>, ptr %ptr, i32 0, i32 %elemindex
+  %dyndata = load float, ptr %dynamic
+  call void @use_float(float %dyndata)
+
+  ret void
+}
diff --git a/llvm/test/CodeGen/DirectX/ResourceAccess/store_typedbuffer.ll b/llvm/test/CodeGen/DirectX/ResourceAccess/store_typedbuffer.ll
new file mode 100644
index 000000000000..17606408cadf
--- /dev/null
+++ b/llvm/test/CodeGen/DirectX/ResourceAccess/store_typedbuffer.ll
@@ -0,0 +1,103 @@
+; RUN: opt -S -dxil-resource-access %s | FileCheck %s
+
+target triple = "dxil-pc-shadermodel6.6-compute"
+
+; CHECK-LABEL: define void @store_float4
+define void @store_float4(<4 x float> %data, i32 %index, i32 %elemindex) {
+  %buffer = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0)
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_1_0_0(
+          i32 0, i32 0, i32 1, i32 0, i1 false)
+
+  ; CHECK-NOT: @llvm.dx.resource.getpointer
+  %ptr = call ptr @llvm.dx.resource.getpointer(
+      target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer, i32 %index)
+
+  ; Store the whole value
+  ; CHECK: call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4f32_1_0_0t.v4f32(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer, i32 %index, <4 x float> %data)
+  store <4 x float> %data, ptr %ptr
+
+  ; Store just the .x component
+  %scalar = extractelement <4 x float> %data, i32 0
+  ; CHECK: %[[LOAD:.*]] = call <4 x float> @llvm.dx.resource.load.typedbuffer.v4f32.tdx.TypedBuffer_v4f32_1_0_0t(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer, i32 %index)
+  ; CHECK: %[[INSERT:.*]] = insertelement <4 x float> %[[LOAD]], float %scalar, i32 0
+  ; CHECK: call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4f32_1_0_0t.v4f32(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer, i32 %index, <4 x float> %[[INSERT]])
+  store float %scalar, ptr %ptr
+
+  ; Store just the .y component
+  ; CHECK: %[[LOAD:.*]] = call <4 x float> @llvm.dx.resource.load.typedbuffer.v4f32.tdx.TypedBuffer_v4f32_1_0_0t(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer, i32 %index)
+  ; CHECK: %[[INSERT:.*]] = insertelement <4 x float> %[[LOAD]], float %scalar, i32 1
+  ; CHECK: call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4f32_1_0_0t.v4f32(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer, i32 %index, <4 x float> %[[INSERT]])
+  %y_ptr = getelementptr inbounds i8, ptr %ptr, i32 4
+  store float %scalar, ptr %y_ptr
+
+  ; Store to one of the elements dynamically
+  ; CHECK: %[[LOAD:.*]] = call <4 x float> @llvm.dx.resource.load.typedbuffer.v4f32.tdx.TypedBuffer_v4f32_1_0_0t(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer, i32 %index)
+  ; CHECK: %[[INSERT:.*]] = insertelement <4 x float> %[[LOAD]], float %scalar, i32 %elemindex
+  ; CHECK: call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4f32_1_0_0t.v4f32(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %buffer, i32 %index, <4 x float> %[[INSERT]])
+  %dynamic = getelementptr inbounds <4 x float>, ptr %ptr, i32 0, i32 %elemindex
+  store float %scalar, ptr %dynamic
+
+  ret void
+}
+
+; CHECK-LABEL: define void @store_half4
+define void @store_half4(<4 x half> %data, i32 %index) {
+  %buffer = call target("dx.TypedBuffer", <4 x half>, 1, 0, 0)
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f16_1_0_0(
+          i32 0, i32 0, i32 1, i32 0, i1 false)
+
+  ; CHECK-NOT: @llvm.dx.resource.getpointer
+  %ptr = call ptr @llvm.dx.resource.getpointer(
+      target("dx.TypedBuffer", <4 x half>, 1, 0, 0) %buffer, i32 %index)
+
+  ; Store the whole value
+  ; CHECK: call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4f16_1_0_0t.v4f16(target("dx.TypedBuffer", <4 x half>, 1, 0, 0) %buffer, i32 %index, <4 x half> %data)
+  store <4 x half> %data, ptr %ptr
+
+  ; Store just the .x component
+  %scalar = extractelement <4 x half> %data, i32 0
+  ; CHECK: %[[LOAD:.*]] = call <4 x half> @llvm.dx.resource.load.typedbuffer.v4f16.tdx.TypedBuffer_v4f16_1_0_0t(target("dx.TypedBuffer", <4 x half>, 1, 0, 0) %buffer, i32 %index)
+  ; CHECK: %[[INSERT:.*]] = insertelement <4 x half> %[[LOAD]], half %scalar, i32 0
+  ; CHECK: call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4f16_1_0_0t.v4f16(target("dx.TypedBuffer", <4 x half>, 1, 0, 0) %buffer, i32 %index, <4 x half> %[[INSERT]])
+  store half %scalar, ptr %ptr
+
+  ; Store just the .y component
+  ; CHECK: %[[LOAD:.*]] = call <4 x half> @llvm.dx.resource.load.typedbuffer.v4f16.tdx.TypedBuffer_v4f16_1_0_0t(target("dx.TypedBuffer", <4 x half>, 1, 0, 0) %buffer, i32 %index)
+  ; CHECK: %[[INSERT:.*]] = insertelement <4 x half> %[[LOAD]], half %scalar, i32 1
+  ; CHECK: call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4f16_1_0_0t.v4f16(target("dx.TypedBuffer", <4 x half>, 1, 0, 0) %buffer, i32 %index, <4 x half> %[[INSERT]])
+  %y_ptr = getelementptr inbounds i8, ptr %ptr, i32 2
+  store half %scalar, ptr %y_ptr
+
+  ret void
+}
+
+; CHECK-LABEL: define void @store_double2
+define void @store_double2(<2 x double> %data, i32 %index) {
+  %buffer = call target("dx.TypedBuffer", <2 x double>, 1, 0, 0)
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v2f64_1_0_0(
+          i32 0, i32 0, i32 1, i32 0, i1 false)
+
+  ; CHECK-NOT: @llvm.dx.resource.getpointer
+  %ptr = call ptr @llvm.dx.resource.getpointer(
+      target("dx.TypedBuffer", <2 x double>, 1, 0, 0) %buffer, i32 %index)
+
+  ; Store the whole value
+  ; CHECK: call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v2f64_1_0_0t.v2f64(target("dx.TypedBuffer", <2 x double>, 1, 0, 0) %buffer, i32 %index, <2 x double> %data)
+  store <2 x double> %data, ptr %ptr
+
+  ; Store just the .x component
+  %scalar = extractelement <2 x double> %data, i32 0
+  ; CHECK: %[[LOAD:.*]] = call <2 x double> @llvm.dx.resource.load.typedbuffer.v2f64.tdx.TypedBuffer_v2f64_1_0_0t(target("dx.TypedBuffer", <2 x double>, 1, 0, 0) %buffer, i32 %index)
+  ; CHECK: %[[INSERT:.*]] = insertelement <2 x double> %[[LOAD]], double %scalar, i32 0
+  ; CHECK: call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v2f64_1_0_0t.v2f64(target("dx.TypedBuffer", <2 x double>, 1, 0, 0) %buffer, i32 %index, <2 x double> %[[INSERT]])
+  store double %scalar, ptr %ptr
+
+  ; Store just the .y component
+  ; CHECK: %[[LOAD:.*]] = call <2 x double> @llvm.dx.resource.load.typedbuffer.v2f64.tdx.TypedBuffer_v2f64_1_0_0t(target("dx.TypedBuffer", <2 x double>, 1, 0, 0) %buffer, i32 %index)
+  ; CHECK: %[[INSERT:.*]] = insertelement <2 x double> %[[LOAD]], double %scalar, i32 1
+  ; CHECK: call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v2f64_1_0_0t.v2f64(target("dx.TypedBuffer", <2 x double>, 1, 0, 0) %buffer, i32 %index, <2 x double> %[[INSERT]])
+  %y_ptr = getelementptr inbounds i8, ptr %ptr, i32 8
+  store double %scalar, ptr %y_ptr
+
+  ret void
+}
diff --git a/llvm/test/CodeGen/DirectX/ResourceGlobalElimination.ll b/llvm/test/CodeGen/DirectX/ResourceGlobalElimination.ll
index c0fbc3d9150b..c837b36a19e1 100644
--- a/llvm/test/CodeGen/DirectX/ResourceGlobalElimination.ll
+++ b/llvm/test/CodeGen/DirectX/ResourceGlobalElimination.ll
@@ -21,27 +21,27 @@ define void @main() local_unnamed_addr #0 {
 entry:
   ; DXOP: %In_h.i1 = call %dx.types.Handle @dx.op.createHandle
   ; DXOP: %Out_h.i2 = call %dx.types.Handle @dx.op.createHandle
-  %In_h.i = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0) @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f32_1_0_0t(i32 0, i32 0, i32 1, i32 0, i1 false)
+  %In_h.i = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0) @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_1_0_0t(i32 0, i32 0, i32 1, i32 0, i1 false)
   store target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %In_h.i, ptr @In, align 4
-  %Out_h.i = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0) @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f32_1_0_0t(i32 4, i32 1, i32 1, i32 0, i1 false)
+  %Out_h.i = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0) @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_1_0_0t(i32 4, i32 1, i32 1, i32 0, i1 false)
   store target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %Out_h.i, ptr @Out, align 4
   ; CSE: call i32 @llvm.dx.flattened.thread.id.in.group()
   %0 = call i32 @llvm.dx.flattened.thread.id.in.group()
   ; CHECK-NOT: load {{.*}} ptr @In
   %1 = load target("dx.TypedBuffer", <4 x float>, 1, 0, 0), ptr @In, align 4
-  ; CSE: call noundef <4 x float> @llvm.dx.typedBufferLoad.v4f32.tdx.TypedBuffer_v4f32_1_0_0t
-  %2 = call noundef <4 x float> @llvm.dx.typedBufferLoad.v4f32.tdx.TypedBuffer_v4f32_1_0_0t(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %1, i32 %0)
+  ; CSE: call noundef <4 x float> @llvm.dx.resource.load.typedbuffer.v4f32.tdx.TypedBuffer_v4f32_1_0_0t
+  %2 = call noundef <4 x float> @llvm.dx.resource.load.typedbuffer.v4f32.tdx.TypedBuffer_v4f32_1_0_0t(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %1, i32 %0)
   ; CHECK-NOT: load {{.*}} ptr @In
   %3 = load target("dx.TypedBuffer", <4 x float>, 1, 0, 0), ptr @In, align 4
-  %4 = call noundef <4 x float> @llvm.dx.typedBufferLoad.v4f32.tdx.TypedBuffer_v4f32_1_0_0t(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %3, i32 %0)
+  %4 = call noundef <4 x float> @llvm.dx.resource.load.typedbuffer.v4f32.tdx.TypedBuffer_v4f32_1_0_0t(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %3, i32 %0)
   %add.i = fadd <4 x float> %2, %4
-  call void @llvm.dx.typedBufferStore.tdx.TypedBuffer_v4f32_1_0_0t.v4f32(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %Out_h.i, i32 %0, <4 x float> %add.i)
+  call void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4f32_1_0_0t.v4f32(target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %Out_h.i, i32 %0, <4 x float> %add.i)
   ; CHECK: ret void
   ret void
 }
 
-; CSE-DAG: declare <4 x float> @llvm.dx.typedBufferLoad.v4f32.tdx.TypedBuffer_v4f32_1_0_0t(target("dx.TypedBuffer", <4 x float>, 1, 0, 0), i32) [[ROAttr:#[0-9]+]]
-; CSE-DAG: declare void @llvm.dx.typedBufferStore.tdx.TypedBuffer_v4f32_1_0_0t.v4f32(target("dx.TypedBuffer", <4 x float>, 1, 0, 0), i32, <4 x float>) [[WOAttr:#[0-9]+]]
+; CSE-DAG: declare <4 x float> @llvm.dx.resource.load.typedbuffer.v4f32.tdx.TypedBuffer_v4f32_1_0_0t(target("dx.TypedBuffer", <4 x float>, 1, 0, 0), i32) [[ROAttr:#[0-9]+]]
+; CSE-DAG: declare void @llvm.dx.resource.store.typedbuffer.tdx.TypedBuffer_v4f32_1_0_0t.v4f32(target("dx.TypedBuffer", <4 x float>, 1, 0, 0), i32, <4 x float>) [[WOAttr:#[0-9]+]]
 
 attributes #0 = { convergent noinline norecurse "frame-pointer"="all" "hlsl.numthreads"="8,1,1" "hlsl.shader"="compute" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
 
diff --git a/llvm/test/CodeGen/DirectX/ShaderFlags/typed-uav-load-additional-formats.ll b/llvm/test/CodeGen/DirectX/ShaderFlags/typed-uav-load-additional-formats.ll
new file mode 100644
index 000000000000..26223359dfdf
--- /dev/null
+++ b/llvm/test/CodeGen/DirectX/ShaderFlags/typed-uav-load-additional-formats.ll
@@ -0,0 +1,44 @@
+; RUN: opt -S --passes="print-dx-shader-flags" 2>&1 %s | FileCheck %s
+; RUN: llc %s --filetype=obj -o - | obj2yaml | FileCheck %s --check-prefix=CHECK-OBJ
+
+target triple = "dxil-pc-shadermodel6.7-library"
+
+; CHECK-OBJ: - Name: SFI0
+; CHECK-OBJ:   Flags:
+; CHECK-OBJ:     TypedUAVLoadAdditionalFormats: true
+
+; CHECK:      Combined Shader Flags for Module
+; CHECK-NEXT: Shader Flags Value: 0x00002000
+
+; CHECK: Note: shader requires additional functionality:
+; CHECK:       Typed UAV Load Additional Formats
+
+; CHECK: Function multicomponent : 0x00002000
+define <4 x float> @multicomponent() #0 {
+  %res = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0)
+      @llvm.dx.resource.handlefrombinding(i32 0, i32 0, i32 1, i32 0, i1 false)
+  %val = call <4 x float> @llvm.dx.resource.load.typedbuffer(
+      target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %res, i32 0)
+  ret <4 x float> %val
+}
+
+; CHECK: Function onecomponent : 0x00000000
+define float @onecomponent() #0 {
+  %res = call target("dx.TypedBuffer", float, 1, 0, 0)
+      @llvm.dx.resource.handlefrombinding(i32 0, i32 0, i32 1, i32 0, i1 false)
+  %val = call float @llvm.dx.resource.load.typedbuffer(
+      target("dx.TypedBuffer", float, 1, 0, 0) %res, i32 0)
+  ret float %val
+}
+
+; CHECK: Function noload : 0x00000000
+define void @noload(<4 x float> %val) #0 {
+  %res = call target("dx.TypedBuffer", <4 x float>, 1, 0, 0)
+      @llvm.dx.resource.handlefrombinding(i32 0, i32 0, i32 1, i32 0, i1 false)
+  call void @llvm.dx.resource.store.typedbuffer(
+      target("dx.TypedBuffer", <4 x float>, 1, 0, 0) %res, i32 0,
+      <4 x float> %val)
+  ret void
+}
+
+attributes #0 = { convergent norecurse nounwind "hlsl.export"}
diff --git a/llvm/test/CodeGen/DirectX/WaveActiveAllTrue.ll b/llvm/test/CodeGen/DirectX/WaveActiveAllTrue.ll
new file mode 100644
index 000000000000..7e84f33579a6
--- /dev/null
+++ b/llvm/test/CodeGen/DirectX/WaveActiveAllTrue.ll
@@ -0,0 +1,10 @@
+; RUN: opt -S -dxil-op-lower -mtriple=dxil-pc-shadermodel6.3-compute %s | FileCheck %s
+
+define noundef i1 @wave_all_simple(i1 noundef %p1) {
+entry:
+; CHECK: call i1 @dx.op.waveAllTrue(i32 114, i1 %p1)
+  %ret = call i1 @llvm.dx.wave.all(i1 %p1)
+  ret i1 %ret
+}
+
+declare i1 @llvm.dx.wave.all(i1)
diff --git a/llvm/test/CodeGen/DirectX/bufferUpdateCounter.ll b/llvm/test/CodeGen/DirectX/bufferUpdateCounter.ll
index 3f2610649cba..57a47d0a39a7 100644
--- a/llvm/test/CodeGen/DirectX/bufferUpdateCounter.ll
+++ b/llvm/test/CodeGen/DirectX/bufferUpdateCounter.ll
@@ -7,12 +7,12 @@ target triple = "dxil-pc-shadermodel6.6-compute"
 define void @update_counter_decrement_vector() {
  ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217, 
   %buffer = call target("dx.TypedBuffer", <4 x float>, 0, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f32_0_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_0_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
 
  ; CHECK-NEXT: [[BUFFANOT:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
  ; CHECK-NEXT: [[REG:%.*]] = call i32 @dx.op.bufferUpdateCounter(i32 70, %dx.types.Handle [[BUFFANOT]], i8 -1)
-  %1 = call i32 @llvm.dx.bufferUpdateCounter(target("dx.TypedBuffer", <4 x float>, 0, 0, 0) %buffer, i8 -1)
+  %1 = call i32 @llvm.dx.resource.updatecounter(target("dx.TypedBuffer", <4 x float>, 0, 0, 0) %buffer, i8 -1)
   ret void
 }
 
@@ -20,11 +20,11 @@ define void @update_counter_decrement_vector() {
 define void @update_counter_increment_vector() {
   ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217, 
   %buffer = call target("dx.TypedBuffer", <4 x float>, 0, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.TypedBuffer_v4f32_0_0_0(
+      @llvm.dx.resource.handlefrombinding.tdx.TypedBuffer_v4f32_0_0_0(
           i32 0, i32 0, i32 1, i32 0, i1 false)
   ; CHECK-NEXT: [[BUFFANOT:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
   ; CHECK-NEXT: [[REG:%.*]] = call i32 @dx.op.bufferUpdateCounter(i32 70, %dx.types.Handle [[BUFFANOT]], i8 1)
-  %1 = call i32 @llvm.dx.bufferUpdateCounter(target("dx.TypedBuffer", <4 x float>, 0, 0, 0) %buffer, i8 1)
+  %1 = call i32 @llvm.dx.resource.updatecounter(target("dx.TypedBuffer", <4 x float>, 0, 0, 0) %buffer, i8 1)
   ret void
 }
 
@@ -32,10 +32,10 @@ define void @update_counter_increment_vector() {
 define void @update_counter_decrement_scalar() {
     ; CHECK: [[BIND:%.*]] = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217, 
   %buffer = call target("dx.RawBuffer", i8, 0, 0)
-      @llvm.dx.handle.fromBinding.tdx.RawBuffer_i8_0_0t(
+      @llvm.dx.resource.handlefrombinding.tdx.RawBuffer_i8_0_0t(
           i32 1, i32 8, i32 1, i32 0, i1 false)
   ; CHECK-NEXT: [[BUFFANOT:%.*]] = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle [[BIND]]
   ; CHECK-NEXT: [[REG:%.*]] = call i32 @dx.op.bufferUpdateCounter(i32 70, %dx.types.Handle [[BUFFANOT]], i8 -1)
-  %1 = call i32 @llvm.dx.bufferUpdateCounter(target("dx.RawBuffer", i8, 0, 0) %buffer, i8 -1)
+  %1 = call i32 @llvm.dx.resource.updatecounter(target("dx.RawBuffer", i8, 0, 0) %buffer, i8 -1)
   ret void
 }
diff --git a/llvm/test/CodeGen/DirectX/flatten-bug-117273.ll b/llvm/test/CodeGen/DirectX/flatten-bug-117273.ll
new file mode 100644
index 000000000000..3ae5832ce832
--- /dev/null
+++ b/llvm/test/CodeGen/DirectX/flatten-bug-117273.ll
@@ -0,0 +1,23 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -S -passes='dxil-flatten-arrays,dxil-op-lower' -mtriple=dxil-pc-shadermodel6.3-library %s | FileCheck %s
+
+
+@ZerroInitArr = internal constant [2 x [3 x float]] [[3 x float] zeroinitializer, [3 x float] [float 1.000000e+00, float 1.000000e+00, float 1.000000e+00]], align 16
+
+
+define internal void @main() {
+; CHECK-LABEL: define internal void @main() {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr [24 x float], ptr @ZerroInitArr.1dim, i32 1
+; CHECK-NEXT:    [[DOTI0:%.*]] = load float, ptr [[TMP0]], align 16
+; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr [24 x float], ptr @ZerroInitArr.1dim, i32 2
+; CHECK-NEXT:    [[DOTI03:%.*]] = load float, ptr [[TMP1]], align 16
+; CHECK-NEXT:    ret void
+;
+entry:
+  %0 = getelementptr [8 x [3 x float]], ptr @ZerroInitArr, i32 0, i32 1
+  %.i0 = load float, ptr %0, align 16
+  %1 = getelementptr [8 x [3 x float]], ptr @ZerroInitArr, i32 0, i32 2
+  %.i03 = load float, ptr %1, align 16
+  ret void
+}
diff --git a/llvm/test/CodeGen/DirectX/llc-pipeline.ll b/llvm/test/CodeGen/DirectX/llc-pipeline.ll
index 40fa30778a15..b07155724941 100644
--- a/llvm/test/CodeGen/DirectX/llc-pipeline.ll
+++ b/llvm/test/CodeGen/DirectX/llc-pipeline.ll
@@ -1,26 +1,40 @@
-; RUN: llc -mtriple=dxil-pc-shadermodel6.3-library -debug-pass=Structure < %s -o /dev/null 2>&1 | \
-; RUN:     grep -v "Verify generated machine code" | FileCheck %s
+; RUN: llc -filetype=asm -mtriple=dxil-pc-shadermodel6.3-library -debug-pass=Structure < %s -o /dev/null 2>&1 | grep -v "Verify generated machine code" | FileCheck %s --check-prefixes=CHECK,CHECK-ASM
+; RUN: llc -filetype=obj -mtriple=dxil-pc-shadermodel6.3-library -debug-pass=Structure < %s -o /dev/null 2>&1 | grep -v "Verify generated machine code" | FileCheck %s --check-prefixes=CHECK,CHECK-OBJ
 
 ; REQUIRES: asserts
 
 ; CHECK-LABEL: Pass Arguments:
 ; CHECK-NEXT: Target Library Information
+; CHECK-NEXT: DXIL Resource Type Analysis
 ; CHECK-NEXT: Target Transform Information
+
+; CHECK-OBJ-NEXT: Machine Module Information
+; CHECK-OBJ-NEXT: Machine Branch Probability Analysis
+; CHECK-OBJ-NEXT: Create Garbage Collector Module Metadata
+
 ; CHECK-NEXT: ModulePass Manager
 ; CHECK-NEXT:   DXIL Finalize Linkage
 ; CHECK-NEXT:   DXIL Intrinsic Expansion
 ; CHECK-NEXT:   DXIL Data Scalarization
 ; CHECK-NEXT:   DXIL Array Flattener
 ; CHECK-NEXT:   FunctionPass Manager
+; CHECK-NEXT:     DXIL Resource Access
 ; CHECK-NEXT:     Dominator Tree Construction
 ; CHECK-NEXT:     Scalarize vector operations
-; CHECK-NEXT:   DXIL Resource analysis
-; CHECK-NEXT:   DXIL Op Lowering
+; CHECK-NEXT:   DXIL Resource Binding Analysis
 ; CHECK-NEXT:   DXIL resource Information
 ; CHECK-NEXT:   DXIL Shader Flag Analysis
 ; CHECK-NEXT:   DXIL Module Metadata analysis
 ; CHECK-NEXT:   DXIL Translate Metadata
+; CHECK-NEXT:   DXIL Op Lowering
 ; CHECK-NEXT:   DXIL Prepare Module
-; CHECK-NEXT:   DXIL Metadata Pretty Printer
-; CHECK-NEXT:   Print Module IR
- 
+
+; CHECK-ASM-NEXT: DXIL Metadata Pretty Printer
+; CHECK-ASM-NEXT: Print Module IR
+
+; CHECK-OBJ-NEXT: DXIL Embedder
+; CHECK-OBJ-NEXT: DXContainer Global Emitter
+; CHECK-OBJ-NEXT: FunctionPass Manager
+; CHECK-OBJ-NEXT:   Lazy Machine Block Frequency Analysis
+; CHECK-OBJ-NEXT:   Machine Optimization Remark Emitter
+; CHECK-OBJ-NEXT:   DXIL Assembly Printer
diff --git a/llvm/test/CodeGen/DirectX/llc-vector-load-scalarize.ll b/llvm/test/CodeGen/DirectX/llc-vector-load-scalarize.ll
index 597252038365..4e522c6ef5da 100644
--- a/llvm/test/CodeGen/DirectX/llc-vector-load-scalarize.ll
+++ b/llvm/test/CodeGen/DirectX/llc-vector-load-scalarize.ll
@@ -21,7 +21,6 @@
 ; CHECK-NOT: @groushared2dArrayofVectors
 ; CHECK-NOT: @groushared2dArrayofVectors.scalarized
 
-
 define <4 x i32> @load_array_vec_test() #0 {
 ; CHECK-LABEL: define <4 x i32> @load_array_vec_test(
 ; CHECK-SAME: ) #[[ATTR0:[0-9]+]] {
@@ -33,18 +32,13 @@ define <4 x i32> @load_array_vec_test() #0 {
 ; CHECK-NEXT:    [[TMP6:%.*]] = load i32, ptr addrspace(3) [[TMP5]], align 4
 ; CHECK-NEXT:    [[TMP7:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 3) to ptr addrspace(3)
 ; CHECK-NEXT:    [[TMP8:%.*]] = load i32, ptr addrspace(3) [[TMP7]], align 4
-; CHECK-NEXT:    [[TMP9:%.*]] = bitcast ptr addrspace(3) @arrayofVecData.scalarized.1dim to ptr addrspace(3)
-; CHECK-NEXT:    [[TMP10:%.*]] = getelementptr [2 x [3 x float]], ptr addrspace(3) [[TMP9]], i32 0, i32 1
-; CHECK-NEXT:    [[TMP11:%.*]] = bitcast ptr addrspace(3) [[TMP10]] to ptr addrspace(3)
+; CHECK-NEXT:    [[TMP11:%.*]] = bitcast ptr addrspace(3) getelementptr inbounds ([6 x float], ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 1) to ptr addrspace(3)
 ; CHECK-NEXT:    [[TMP12:%.*]] = load i32, ptr addrspace(3) [[TMP11]], align 4
-; CHECK-NEXT:    [[TMP13:%.*]] = bitcast ptr addrspace(3) [[TMP10]] to ptr addrspace(3)
-; CHECK-NEXT:    [[DOTI12:%.*]] = getelementptr i32, ptr addrspace(3) [[TMP13]], i32 1
+; CHECK-NEXT:    [[DOTI12:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([6 x float], ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 1), i32 1) to ptr addrspace(3)
 ; CHECK-NEXT:    [[DOTI13:%.*]] = load i32, ptr addrspace(3) [[DOTI12]], align 4
-; CHECK-NEXT:    [[TMP14:%.*]] = bitcast ptr addrspace(3) [[TMP10]] to ptr addrspace(3)
-; CHECK-NEXT:    [[DOTI24:%.*]] = getelementptr i32, ptr addrspace(3) [[TMP14]], i32 2
+; CHECK-NEXT:    [[DOTI24:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([6 x float], ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 1), i32 2) to ptr addrspace(3)
 ; CHECK-NEXT:    [[DOTI25:%.*]] = load i32, ptr addrspace(3) [[DOTI24]], align 4
-; CHECK-NEXT:    [[TMP15:%.*]] = bitcast ptr addrspace(3) [[TMP10]] to ptr addrspace(3)
-; CHECK-NEXT:    [[DOTI36:%.*]] = getelementptr i32, ptr addrspace(3) [[TMP15]], i32 3
+; CHECK-NEXT:    [[DOTI36:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([6 x float], ptr addrspace(3) @arrayofVecData.scalarized.1dim, i32 1), i32 3) to ptr addrspace(3)
 ; CHECK-NEXT:    [[DOTI37:%.*]] = load i32, ptr addrspace(3) [[DOTI36]], align 4
 ; CHECK-NEXT:    [[DOTI08:%.*]] = add i32 [[TMP2]], [[TMP12]]
 ; CHECK-NEXT:    [[DOTI19:%.*]] = add i32 [[TMP4]], [[DOTI13]]
@@ -87,7 +81,7 @@ define <4 x i32> @load_vec_test() #0 {
 define <4 x i32> @load_static_array_of_vec_test(i32 %index) #0 {
 ; CHECK-LABEL: define <4 x i32> @load_static_array_of_vec_test(
 ; CHECK-SAME: i32 [[INDEX:%.*]]) #[[ATTR0]] {
-; CHECK-NEXT:    [[DOTFLAT:%.*]] = getelementptr [12 x i32], ptr @staticArrayOfVecData.scalarized.1dim, i32 [[INDEX]]
+; CHECK-NEXT:    [[DOTFLAT:%.*]] = getelementptr inbounds [12 x i32], ptr @staticArrayOfVecData.scalarized.1dim, i32 [[INDEX]]
 ; CHECK-NEXT:    [[TMP1:%.*]] = bitcast ptr [[DOTFLAT]] to ptr
 ; CHECK-NEXT:    [[TMP2:%.*]] = load i32, ptr [[TMP1]], align 4
 ; CHECK-NEXT:    [[TMP3:%.*]] = bitcast ptr [[DOTFLAT]] to ptr
@@ -121,18 +115,13 @@ define <4 x i32> @multid_load_test() #0 {
 ; CHECK-NEXT:    [[TMP6:%.*]] = load i32, ptr addrspace(3) [[TMP5]], align 4
 ; CHECK-NEXT:    [[TMP7:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, i32 3) to ptr addrspace(3)
 ; CHECK-NEXT:    [[TMP8:%.*]] = load i32, ptr addrspace(3) [[TMP7]], align 4
-; CHECK-NEXT:    [[TMP9:%.*]] = bitcast ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim to ptr addrspace(3)
-; CHECK-NEXT:    [[TMP10:%.*]] = getelementptr [3 x [3 x [4 x i32]]], ptr addrspace(3) [[TMP9]], i32 0, i32 1, i32 1
-; CHECK-NEXT:    [[TMP11:%.*]] = bitcast ptr addrspace(3) [[TMP10]] to ptr addrspace(3)
+; CHECK-NEXT:    [[TMP11:%.*]] = bitcast ptr addrspace(3) getelementptr inbounds ([36 x i32], ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, i32 1) to ptr addrspace(3)
 ; CHECK-NEXT:    [[TMP12:%.*]] = load i32, ptr addrspace(3) [[TMP11]], align 4
-; CHECK-NEXT:    [[TMP13:%.*]] = bitcast ptr addrspace(3) [[TMP10]] to ptr addrspace(3)
-; CHECK-NEXT:    [[DOTI12:%.*]] = getelementptr i32, ptr addrspace(3) [[TMP13]], i32 1
+; CHECK-NEXT:    [[DOTI12:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([36 x i32], ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, i32 1), i32 1) to ptr addrspace(3)
 ; CHECK-NEXT:    [[DOTI13:%.*]] = load i32, ptr addrspace(3) [[DOTI12]], align 4
-; CHECK-NEXT:    [[TMP14:%.*]] = bitcast ptr addrspace(3) [[TMP10]] to ptr addrspace(3)
-; CHECK-NEXT:    [[DOTI24:%.*]] = getelementptr i32, ptr addrspace(3) [[TMP14]], i32 2
+; CHECK-NEXT:    [[DOTI24:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([36 x i32], ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, i32 1), i32 2) to ptr addrspace(3)
 ; CHECK-NEXT:    [[DOTI25:%.*]] = load i32, ptr addrspace(3) [[DOTI24]], align 4
-; CHECK-NEXT:    [[TMP15:%.*]] = bitcast ptr addrspace(3) [[TMP10]] to ptr addrspace(3)
-; CHECK-NEXT:    [[DOTI36:%.*]] = getelementptr i32, ptr addrspace(3) [[TMP15]], i32 3
+; CHECK-NEXT:    [[DOTI36:%.*]] = bitcast ptr addrspace(3) getelementptr (i32, ptr addrspace(3) getelementptr inbounds ([36 x i32], ptr addrspace(3) @groushared2dArrayofVectors.scalarized.1dim, i32 1), i32 3) to ptr addrspace(3)
 ; CHECK-NEXT:    [[DOTI37:%.*]] = load i32, ptr addrspace(3) [[DOTI36]], align 4
 ; CHECK-NEXT:    [[DOTI08:%.*]] = add i32 [[TMP2]], [[TMP12]]
 ; CHECK-NEXT:    [[DOTI19:%.*]] = add i32 [[TMP4]], [[DOTI13]]
diff --git a/llvm/test/CodeGen/DirectX/scalar-bug-117273.ll b/llvm/test/CodeGen/DirectX/scalar-bug-117273.ll
new file mode 100644
index 000000000000..25dc2c36b4e1
--- /dev/null
+++ b/llvm/test/CodeGen/DirectX/scalar-bug-117273.ll
@@ -0,0 +1,25 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -S -passes='dxil-data-scalarization,dxil-flatten-arrays,function(scalarizer<load-store>),dxil-op-lower' -mtriple=dxil-pc-shadermodel6.3-library %s | FileCheck %s
+
+
+@StaticArr = internal constant [8 x <3 x float>] [<3 x float> zeroinitializer, <3 x float> splat (float 5.000000e-01), <3 x float> <float 1.000000e+00, float 5.000000e-01, float 5.000000e-01>, <3 x float> <float 5.000000e-01, float 1.000000e+00, float 5.000000e-01>, <3 x float> <float 5.000000e-01, float 5.000000e-01, float 1.000000e+00>, <3 x float> <float 5.000000e-01, float 1.000000e+00, float 1.000000e+00>, <3 x float> <float 1.000000e+00, float 5.000000e-01, float 1.000000e+00>, <3 x float> <float 1.000000e+00, float 1.000000e+00, float 5.000000e-01>], align 16
+
+; Function Attrs: alwaysinline convergent mustprogress norecurse nounwind
+define internal void @main() #1 {
+; CHECK-LABEL: define internal void @main() {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[DOTI0:%.*]] = load float, ptr getelementptr inbounds ([24 x float], ptr @StaticArr.scalarized.1dim, i32 1), align 16
+; CHECK-NEXT:    [[DOTI1:%.*]] = load float, ptr getelementptr (float, ptr getelementptr inbounds ([24 x float], ptr @StaticArr.scalarized.1dim, i32 1), i32 1), align 4
+; CHECK-NEXT:    [[DOTI2:%.*]] = load float, ptr getelementptr (float, ptr getelementptr inbounds ([24 x float], ptr @StaticArr.scalarized.1dim, i32 1), i32 2), align 8
+; CHECK-NEXT:    [[DOTI01:%.*]] = load float, ptr getelementptr inbounds ([24 x float], ptr @StaticArr.scalarized.1dim, i32 2), align 16
+; CHECK-NEXT:    [[DOTI12:%.*]] = load float, ptr getelementptr (float, ptr getelementptr inbounds ([24 x float], ptr @StaticArr.scalarized.1dim, i32 2), i32 1), align 4
+; CHECK-NEXT:    [[DOTI23:%.*]] = load float, ptr getelementptr (float, ptr getelementptr inbounds ([24 x float], ptr @StaticArr.scalarized.1dim, i32 2), i32 2), align 8
+; CHECK-NEXT:    ret void
+;
+entry:
+  %arrayidx = getelementptr inbounds [8 x <3 x float>], ptr @StaticArr, i32 0, i32 1
+  %2 = load <3 x float>, ptr %arrayidx, align 16
+  %arrayidx2 = getelementptr inbounds [8 x <3 x float>], ptr @StaticArr, i32 0, i32 2
+  %3 = load <3 x float>, ptr %arrayidx2, align 16
+  ret void
+}
diff --git a/llvm/test/CodeGen/DirectX/scalar-load.ll b/llvm/test/CodeGen/DirectX/scalar-load.ll
index a32db8b8e399..ed1e9109b7b1 100644
--- a/llvm/test/CodeGen/DirectX/scalar-load.ll
+++ b/llvm/test/CodeGen/DirectX/scalar-load.ll
@@ -2,10 +2,10 @@
 
 ; Make sure we can load groupshared, static vectors and arrays of vectors
 
-@"arrayofVecData" = local_unnamed_addr addrspace(3) global [2 x <3 x float>] zeroinitializer, align 16
-@"vecData" = external addrspace(3) global <4 x i32>, align 4
+@arrayofVecData = local_unnamed_addr addrspace(3) global [2 x <3 x float>] zeroinitializer, align 16
+@vecData = external addrspace(3) global <4 x i32>, align 4
 @staticArrayOfVecData = internal global [3 x <4 x i32>] [<4 x i32> <i32 1, i32 2, i32 3, i32 4>, <4 x i32> <i32 5, i32 6, i32 7, i32 8>, <4 x i32> <i32 9, i32 10, i32 11, i32 12>], align 4
-@"groushared2dArrayofVectors" = local_unnamed_addr addrspace(3) global [3 x [ 3 x <4 x i32>]] zeroinitializer, align 16
+@groushared2dArrayofVectors = local_unnamed_addr addrspace(3) global [3 x [ 3 x <4 x i32>]] zeroinitializer, align 16
 
 ; CHECK: @arrayofVecData.scalarized = local_unnamed_addr addrspace(3) global [2 x [3 x float]] zeroinitializer, align 16
 ; CHECK: @vecData.scalarized = external addrspace(3) global [4 x i32], align 4
@@ -19,12 +19,12 @@
 
 
 ; CHECK-LABEL: load_array_vec_test
-define <4 x i32> @load_array_vec_test() #0 {
-  ; CHECK-COUNT-8: load i32, ptr addrspace(3) {{(.*@arrayofVecData.scalarized.*|%.*)}}, align 4
-  %1 = load <4 x i32>, <4 x i32> addrspace(3)* getelementptr inbounds ([2 x <4 x i32>], [2 x <4 x i32>] addrspace(3)* @"arrayofVecData", i32 0, i32 0), align 4
-  %2 = load <4 x i32>, <4 x i32> addrspace(3)* getelementptr inbounds ([2 x <4 x i32>], [2 x <4 x i32>] addrspace(3)* @"arrayofVecData", i32 0, i32 1), align 4
-  %3 = add <4 x i32> %1, %2
-  ret <4 x i32> %3
+define <3 x float> @load_array_vec_test() #0 {
+  ; CHECK-COUNT-6: load float, ptr addrspace(3) {{(.*@arrayofVecData.scalarized.*|%.*)}}, align 4
+  %1 = load <3 x float>, <3 x float> addrspace(3)* getelementptr inbounds ([2 x <3 x float>], [2 x <3 x float>] addrspace(3)* @"arrayofVecData", i32 0, i32 0), align 4
+  %2 = load <3 x float>, <3 x float> addrspace(3)* getelementptr inbounds ([2 x <3 x float>], [2 x <3 x float>] addrspace(3)* @"arrayofVecData", i32 0, i32 1), align 4
+  %3 = fadd <3 x float> %1, %2
+  ret <3 x float> %3
 }
 
 ; CHECK-LABEL: load_vec_test
@@ -36,8 +36,14 @@ define <4 x i32> @load_vec_test() #0 {
 
 ; CHECK-LABEL: load_static_array_of_vec_test
 define <4 x i32> @load_static_array_of_vec_test(i32 %index) #0 {
-  ; CHECK: getelementptr [3 x [4 x i32]], ptr @staticArrayOfVecData.scalarized, i32 0, i32 %index
-  ; CHECK-COUNT-4: load i32, ptr {{.*}}, align 4
+  ; CHECK: getelementptr inbounds [3 x [4 x i32]], ptr @staticArrayOfVecData.scalarized, i32 0, i32 %index
+  ; CHECK: load i32, ptr {{.*}}, align 4
+  ; CHECK: getelementptr i32, ptr {{.*}}, i32 1
+  ; CHECK: load i32, ptr {{.*}}, align 4
+  ; CHECK: getelementptr i32, ptr {{.*}}, i32 2
+  ; CHECK: load i32, ptr {{.*}}, align 4
+  ; CHECK: getelementptr i32, ptr {{.*}}, i32 3
+  ; CHECK: load i32, ptr {{.*}}, align 4
   %3 = getelementptr inbounds [3 x <4 x i32>], [3 x <4 x i32>]* @staticArrayOfVecData, i32 0, i32 %index
   %4 = load <4 x i32>, <4 x i32>* %3, align 4
   ret <4 x i32> %4
diff --git a/llvm/test/CodeGen/DirectX/vector_reduce_add.ll b/llvm/test/CodeGen/DirectX/vector_reduce_add.ll
new file mode 100644
index 000000000000..d4ee16a24cb4
--- /dev/null
+++ b/llvm/test/CodeGen/DirectX/vector_reduce_add.ll
@@ -0,0 +1,343 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -S  -dxil-intrinsic-expansion  < %s | FileCheck %s
+
+; Make sure dxil operation function calls for lvm.vector.reduce.fadd and lvm.vector.reduce.add are generate.
+
+define noundef half @test_length_half2(<2 x half> noundef %p0) {
+; CHECK-LABEL: define noundef half @test_length_half2(
+; CHECK-SAME: <2 x half> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <2 x half> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x half> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = fadd half [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    ret half [[TMP2]]
+;
+entry:
+  %rdx.fadd = call half @llvm.vector.reduce.fadd.v2f16(half 0xH0000, <2 x half> %p0)
+  ret half %rdx.fadd
+}
+
+define noundef half @test_length_half2_start1(<2 x half> noundef %p0) {
+; CHECK-LABEL: define noundef half @test_length_half2_start1(
+; CHECK-SAME: <2 x half> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <2 x half> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = fadd half [[TMP0]], 0xH0001
+; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x half> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP3:%.*]] = fadd half [[TMP1]], [[TMP2]]
+; CHECK-NEXT:    ret half [[TMP3]]
+;
+entry:
+  %rdx.fadd = call half @llvm.vector.reduce.fadd.v2f16(half 0xH0001, <2 x half> %p0)
+  ret half %rdx.fadd
+}
+
+define noundef half @test_length_half3(<3 x half> noundef %p0) {
+; CHECK-LABEL: define noundef half @test_length_half3(
+; CHECK-SAME: <3 x half> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <3 x half> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <3 x half> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = fadd half [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <3 x half> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP4:%.*]] = fadd half [[TMP2]], [[TMP3]]
+; CHECK-NEXT:    ret half [[TMP4]]
+;
+entry:
+  %rdx.fadd = call half @llvm.vector.reduce.fadd.v3f16(half 0xH0000, <3 x half> %p0)
+  ret half %rdx.fadd
+}
+
+define noundef half @test_length_half4(<4 x half> noundef %p0) {
+; CHECK-LABEL: define noundef half @test_length_half4(
+; CHECK-SAME: <4 x half> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <4 x half> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x half> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = fadd half [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x half> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP4:%.*]] = fadd half [[TMP2]], [[TMP3]]
+; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <4 x half> [[P0]], i64 3
+; CHECK-NEXT:    [[TMP6:%.*]] = fadd half [[TMP4]], [[TMP5]]
+; CHECK-NEXT:    ret half [[TMP6]]
+;
+entry:
+  %rdx.fadd = call half @llvm.vector.reduce.fadd.v4f16(half 0xH0000, <4 x half> %p0)
+  ret half %rdx.fadd
+}
+
+define noundef float @test_length_float2(<2 x float> noundef %p0) {
+; CHECK-LABEL: define noundef float @test_length_float2(
+; CHECK-SAME: <2 x float> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <2 x float> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x float> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = fadd float [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    ret float [[TMP2]]
+;
+entry:
+  %rdx.fadd = call float @llvm.vector.reduce.fadd.v2f32(float 0.000000e+00, <2 x float> %p0)
+  ret float %rdx.fadd
+}
+
+define noundef float @test_length_float3(<3 x float> noundef %p0) {
+; CHECK-LABEL: define noundef float @test_length_float3(
+; CHECK-SAME: <3 x float> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <3 x float> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <3 x float> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = fadd float [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <3 x float> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP4:%.*]] = fadd float [[TMP2]], [[TMP3]]
+; CHECK-NEXT:    ret float [[TMP4]]
+;
+entry:
+  %rdx.fadd = call float @llvm.vector.reduce.fadd.v3f32(float 0.000000e+00, <3 x float> %p0)
+  ret float %rdx.fadd
+}
+
+define noundef float @test_length_float3_start1(<3 x float> noundef %p0) {
+; CHECK-LABEL: define noundef float @test_length_float3_start1(
+; CHECK-SAME: <3 x float> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <3 x float> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = fadd float [[TMP0]], 1.000000e+00
+; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <3 x float> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP3:%.*]] = fadd float [[TMP1]], [[TMP2]]
+; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <3 x float> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP5:%.*]] = fadd float [[TMP3]], [[TMP4]]
+; CHECK-NEXT:    ret float [[TMP5]]
+;
+entry:
+  %rdx.fadd = call float @llvm.vector.reduce.fadd.v3f32(float 1.000000e+00, <3 x float> %p0)
+  ret float %rdx.fadd
+}
+
+define noundef float @test_length_float4(<4 x float> noundef %p0) {
+; CHECK-LABEL: define noundef float @test_length_float4(
+; CHECK-SAME: <4 x float> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <4 x float> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = fadd float [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x float> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP4:%.*]] = fadd float [[TMP2]], [[TMP3]]
+; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <4 x float> [[P0]], i64 3
+; CHECK-NEXT:    [[TMP6:%.*]] = fadd float [[TMP4]], [[TMP5]]
+; CHECK-NEXT:    ret float [[TMP6]]
+;
+entry:
+  %rdx.fadd = call float @llvm.vector.reduce.fadd.v4f32(float 0.000000e+00, <4 x float> %p0)
+  ret float %rdx.fadd
+}
+
+define noundef double @test_length_double2(<2 x double> noundef %p0) {
+; CHECK-LABEL: define noundef double @test_length_double2(
+; CHECK-SAME: <2 x double> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <2 x double> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = fadd double [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    ret double [[TMP2]]
+;
+entry:
+  %rdx.fadd = call double @llvm.vector.reduce.fadd.v2f64(double 0.000000e+00, <2 x double> %p0)
+  ret double %rdx.fadd
+}
+
+define noundef double @test_length_double3(<3 x double> noundef %p0) {
+; CHECK-LABEL: define noundef double @test_length_double3(
+; CHECK-SAME: <3 x double> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <3 x double> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <3 x double> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = fadd double [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <3 x double> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP4:%.*]] = fadd double [[TMP2]], [[TMP3]]
+; CHECK-NEXT:    ret double [[TMP4]]
+;
+entry:
+  %rdx.fadd = call double @llvm.vector.reduce.fadd.v3f64(double 0.000000e+00, <3 x double> %p0)
+  ret double %rdx.fadd
+}
+
+define noundef double @test_length_double4(<4 x double> noundef %p0) {
+; CHECK-LABEL: define noundef double @test_length_double4(
+; CHECK-SAME: <4 x double> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <4 x double> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x double> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = fadd double [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x double> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP4:%.*]] = fadd double [[TMP2]], [[TMP3]]
+; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <4 x double> [[P0]], i64 3
+; CHECK-NEXT:    [[TMP6:%.*]] = fadd double [[TMP4]], [[TMP5]]
+; CHECK-NEXT:    ret double [[TMP6]]
+;
+entry:
+  %rdx.fadd = call double @llvm.vector.reduce.fadd.v4f64(double 0.000000e+00, <4 x double> %p0)
+  ret double %rdx.fadd
+}
+
+define noundef double @test_length_double4_start1(<4 x double> noundef %p0) {
+; CHECK-LABEL: define noundef double @test_length_double4_start1(
+; CHECK-SAME: <4 x double> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <4 x double> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = fadd double [[TMP0]], 1.000000e+00
+; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x double> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP3:%.*]] = fadd double [[TMP1]], [[TMP2]]
+; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <4 x double> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP5:%.*]] = fadd double [[TMP3]], [[TMP4]]
+; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <4 x double> [[P0]], i64 3
+; CHECK-NEXT:    [[TMP7:%.*]] = fadd double [[TMP5]], [[TMP6]]
+; CHECK-NEXT:    ret double [[TMP7]]
+;
+entry:
+  %rdx.fadd = call double @llvm.vector.reduce.fadd.v4f64(double 1.000000e+00, <4 x double> %p0)
+  ret double %rdx.fadd
+}
+
+define noundef i16 @test_length_short2(<2 x i16> noundef %p0) {
+; CHECK-LABEL: define noundef i16 @test_length_short2(
+; CHECK-SAME: <2 x i16> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <2 x i16> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x i16> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = add i16 [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    ret i16 [[TMP2]]
+;
+entry:
+  %rdx.add = call i16 @llvm.vector.reduce.add.v2i16(<2 x i16> %p0)
+  ret i16 %rdx.add
+}
+
+define noundef i16 @test_length_short3(<3 x i16> noundef %p0) {
+; CHECK-LABEL: define noundef i16 @test_length_short3(
+; CHECK-SAME: <3 x i16> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <3 x i16> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <3 x i16> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = add i16 [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <3 x i16> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP4:%.*]] = add i16 [[TMP2]], [[TMP3]]
+; CHECK-NEXT:    ret i16 [[TMP4]]
+;
+entry:
+  %rdx.fadd = call i16 @llvm.vector.reduce.add.v3i16(<3 x i16> %p0)
+  ret i16 %rdx.fadd
+}
+
+define noundef i16 @test_length_short4(<4 x i16> noundef %p0) {
+; CHECK-LABEL: define noundef i16 @test_length_short4(
+; CHECK-SAME: <4 x i16> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <4 x i16> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x i16> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = add i16 [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x i16> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP4:%.*]] = add i16 [[TMP2]], [[TMP3]]
+; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <4 x i16> [[P0]], i64 3
+; CHECK-NEXT:    [[TMP6:%.*]] = add i16 [[TMP4]], [[TMP5]]
+; CHECK-NEXT:    ret i16 [[TMP6]]
+;
+entry:
+  %rdx.fadd = call i16 @llvm.vector.reduce.add.v4i16(<4 x i16> %p0)
+  ret i16 %rdx.fadd
+}
+
+define noundef i32 @test_length_int2(<2 x i32> noundef %p0) {
+; CHECK-LABEL: define noundef i32 @test_length_int2(
+; CHECK-SAME: <2 x i32> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <2 x i32> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x i32> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = add i32 [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    ret i32 [[TMP2]]
+;
+entry:
+  %rdx.add = call i32 @llvm.vector.reduce.add.v2i32(<2 x i32> %p0)
+  ret i32 %rdx.add
+}
+
+define noundef i32 @test_length_int3(<3 x i32> noundef %p0) {
+; CHECK-LABEL: define noundef i32 @test_length_int3(
+; CHECK-SAME: <3 x i32> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <3 x i32> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <3 x i32> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = add i32 [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <3 x i32> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP4:%.*]] = add i32 [[TMP2]], [[TMP3]]
+; CHECK-NEXT:    ret i32 [[TMP4]]
+;
+entry:
+  %rdx.fadd = call i32 @llvm.vector.reduce.add.v3i32(<3 x i32> %p0)
+  ret i32 %rdx.fadd
+}
+
+define noundef i32 @test_length_int4(<4 x i32> noundef %p0) {
+; CHECK-LABEL: define noundef i32 @test_length_int4(
+; CHECK-SAME: <4 x i32> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <4 x i32> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x i32> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = add i32 [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x i32> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP4:%.*]] = add i32 [[TMP2]], [[TMP3]]
+; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <4 x i32> [[P0]], i64 3
+; CHECK-NEXT:    [[TMP6:%.*]] = add i32 [[TMP4]], [[TMP5]]
+; CHECK-NEXT:    ret i32 [[TMP6]]
+;
+entry:
+  %rdx.fadd = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> %p0)
+  ret i32 %rdx.fadd
+}
+
+define noundef i64 @test_length_int64_2(<2 x i64> noundef %p0) {
+; CHECK-LABEL: define noundef i64 @test_length_int64_2(
+; CHECK-SAME: <2 x i64> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <2 x i64> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x i64> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    ret i64 [[TMP2]]
+;
+entry:
+  %rdx.add = call i64 @llvm.vector.reduce.add.v2i64(<2 x i64> %p0)
+  ret i64 %rdx.add
+}
+
+define noundef i64 @test_length_int64_3(<3 x i64> noundef %p0) {
+; CHECK-LABEL: define noundef i64 @test_length_int64_3(
+; CHECK-SAME: <3 x i64> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <3 x i64> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <3 x i64> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <3 x i64> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[TMP2]], [[TMP3]]
+; CHECK-NEXT:    ret i64 [[TMP4]]
+;
+entry:
+  %rdx.fadd = call i64 @llvm.vector.reduce.add.v3i64(<3 x i64> %p0)
+  ret i64 %rdx.fadd
+}
+
+define noundef i64 @test_length_int64_4(<4 x i64> noundef %p0) {
+; CHECK-LABEL: define noundef i64 @test_length_int64_4(
+; CHECK-SAME: <4 x i64> noundef [[P0:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[TMP0:%.*]] = extractelement <4 x i64> [[P0]], i64 0
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x i64> [[P0]], i64 1
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[TMP0]], [[TMP1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x i64> [[P0]], i64 2
+; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[TMP2]], [[TMP3]]
+; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <4 x i64> [[P0]], i64 3
+; CHECK-NEXT:    [[TMP6:%.*]] = add i64 [[TMP4]], [[TMP5]]
+; CHECK-NEXT:    ret i64 [[TMP6]]
+;
+entry:
+  %rdx.fadd = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> %p0)
+  ret i64 %rdx.fadd
+}