diff --git a/lib/Target/AMDGPU/SIISelLowering.cpp b/lib/Target/AMDGPU/SIISelLowering.cpp
index 972a929c943..faecf3c1da9 100644
--- a/lib/Target/AMDGPU/SIISelLowering.cpp
+++ b/lib/Target/AMDGPU/SIISelLowering.cpp
@@ -43,9 +43,6 @@ SITargetLowering::SITargetLowering(TargetMachine &TM,
   addRegisterClass(MVT::i1, &AMDGPU::VReg_1RegClass);
   addRegisterClass(MVT::i64, &AMDGPU::SReg_64RegClass);
 
-  addRegisterClass(MVT::v32i8, &AMDGPU::SReg_256RegClass);
-  addRegisterClass(MVT::v64i8, &AMDGPU::SReg_512RegClass);
-
   addRegisterClass(MVT::i32, &AMDGPU::SReg_32RegClass);
   addRegisterClass(MVT::f32, &AMDGPU::VGPR_32RegClass);
 
diff --git a/lib/Target/AMDGPU/SIInstrInfo.td b/lib/Target/AMDGPU/SIInstrInfo.td
index 8735277149a..7de293f012b 100644
--- a/lib/Target/AMDGPU/SIInstrInfo.td
+++ b/lib/Target/AMDGPU/SIInstrInfo.td
@@ -120,7 +120,7 @@ def SIload_input : SDNode<"AMDGPUISD::LOAD_INPUT",
 >;
 
 class SDSample<string opcode> : SDNode <opcode,
-  SDTypeProfile<1, 4, [SDTCisVT<0, v4f32>, SDTCisVT<2, v32i8>,
+  SDTypeProfile<1, 4, [SDTCisVT<0, v4f32>, SDTCisVT<2, v8i32>,
                        SDTCisVT<3, v4i32>, SDTCisVT<4, i32>]>
 >;
 
diff --git a/lib/Target/AMDGPU/SIInstructions.td b/lib/Target/AMDGPU/SIInstructions.td
index 7b4785bc819..9d7e6ca3fc8 100644
--- a/lib/Target/AMDGPU/SIInstructions.td
+++ b/lib/Target/AMDGPU/SIInstructions.td
@@ -2102,7 +2102,6 @@ let AddedComplexity = 100 in {
 defm : SMRD_Pattern <"S_LOAD_DWORD", i32>;
 defm : SMRD_Pattern <"S_LOAD_DWORDX2", v2i32>;
 defm : SMRD_Pattern <"S_LOAD_DWORDX4", v4i32>;
-defm : SMRD_Pattern <"S_LOAD_DWORDX8", v32i8>;
 defm : SMRD_Pattern <"S_LOAD_DWORDX8", v8i32>;
 defm : SMRD_Pattern <"S_LOAD_DWORDX16", v16i32>;
 
@@ -2343,34 +2342,34 @@ defm : ImagePatterns<int_SI_image_load_mip, "IMAGE_LOAD_MIP">;
 
 /* SIsample for simple 1D texture lookup */
 def : Pat <
-  (SIsample i32:$addr, v32i8:$rsrc, v4i32:$sampler, imm),
+  (SIsample i32:$addr, v8i32:$rsrc, v4i32:$sampler, imm),
   (IMAGE_SAMPLE_V4_V1 0xf, 0, 0, 0, 0, 0, 0, 0, $addr, $rsrc, $sampler)
 >;
 
 class SamplePattern<SDNode name, MIMG opcode, ValueType vt> : Pat <
-    (name vt:$addr, v32i8:$rsrc, v4i32:$sampler, imm),
+    (name vt:$addr, v8i32:$rsrc, v4i32:$sampler, imm),
     (opcode 0xf, 0, 0, 0, 0, 0, 0, 0, $addr, $rsrc, $sampler)
 >;
 
 class SampleRectPattern<SDNode name, MIMG opcode, ValueType vt> : Pat <
-    (name vt:$addr, v32i8:$rsrc, v4i32:$sampler, TEX_RECT),
+    (name vt:$addr, v8i32:$rsrc, v4i32:$sampler, TEX_RECT),
     (opcode 0xf, 1, 0, 0, 0, 0, 0, 0, $addr, $rsrc, $sampler)
 >;
 
 class SampleArrayPattern<SDNode name, MIMG opcode, ValueType vt> : Pat <
-    (name vt:$addr, v32i8:$rsrc, v4i32:$sampler, TEX_ARRAY),
+    (name vt:$addr, v8i32:$rsrc, v4i32:$sampler, TEX_ARRAY),
     (opcode 0xf, 0, 0, 1, 0, 0, 0, 0, $addr, $rsrc, $sampler)
 >;
 
 class SampleShadowPattern<SDNode name, MIMG opcode,
                           ValueType vt> : Pat <
-    (name vt:$addr, v32i8:$rsrc, v4i32:$sampler, TEX_SHADOW),
+    (name vt:$addr, v8i32:$rsrc, v4i32:$sampler, TEX_SHADOW),
     (opcode 0xf, 0, 0, 0, 0, 0, 0, 0, $addr, $rsrc, $sampler)
 >;
 
 class SampleShadowArrayPattern<SDNode name, MIMG opcode,
                                ValueType vt> : Pat <
-    (name vt:$addr, v32i8:$rsrc, v4i32:$sampler, TEX_SHADOW_ARRAY),
+    (name vt:$addr, v8i32:$rsrc, v4i32:$sampler, TEX_SHADOW_ARRAY),
     (opcode 0xf, 0, 0, 1, 0, 0, 0, 0, $addr, $rsrc, $sampler)
 >;
 
@@ -2423,22 +2422,22 @@ defm : SamplePatterns<IMAGE_SAMPLE_V4_V16, IMAGE_SAMPLE_C_V4_V16,
 
 /* int_SI_imageload for texture fetches consuming varying address parameters */
 class ImageLoadPattern<Intrinsic name, MIMG opcode, ValueType addr_type> : Pat <
-    (name addr_type:$addr, v32i8:$rsrc, imm),
+    (name addr_type:$addr, v8i32:$rsrc, imm),
     (opcode 0xf, 0, 0, 0, 0, 0, 0, 0, $addr, $rsrc)
 >;
 
 class ImageLoadArrayPattern<Intrinsic name, MIMG opcode, ValueType addr_type> : Pat <
-    (name addr_type:$addr, v32i8:$rsrc, TEX_ARRAY),
+    (name addr_type:$addr, v8i32:$rsrc, TEX_ARRAY),
     (opcode 0xf, 0, 0, 1, 0, 0, 0, 0, $addr, $rsrc)
 >;
 
 class ImageLoadMSAAPattern<Intrinsic name, MIMG opcode, ValueType addr_type> : Pat <
-    (name addr_type:$addr, v32i8:$rsrc, TEX_MSAA),
+    (name addr_type:$addr, v8i32:$rsrc, TEX_MSAA),
     (opcode 0xf, 0, 0, 0, 0, 0, 0, 0, $addr, $rsrc)
 >;
 
 class ImageLoadArrayMSAAPattern<Intrinsic name, MIMG opcode, ValueType addr_type> : Pat <
-    (name addr_type:$addr, v32i8:$rsrc, TEX_ARRAY_MSAA),
+    (name addr_type:$addr, v8i32:$rsrc, TEX_ARRAY_MSAA),
     (opcode 0xf, 0, 0, 1, 0, 0, 0, 0, $addr, $rsrc)
 >;
 
@@ -2552,12 +2551,8 @@ def : BitConvert <v4i32, v2f64, VReg_128>;
 
 def : BitConvert <v8f32, v8i32, SReg_256>;
 def : BitConvert <v8i32, v8f32, SReg_256>;
-def : BitConvert <v8i32, v32i8, SReg_256>;
-def : BitConvert <v32i8, v8i32, SReg_256>;
-def : BitConvert <v8i32, v32i8, VReg_256>;
 def : BitConvert <v8i32, v8f32, VReg_256>;
 def : BitConvert <v8f32, v8i32, VReg_256>;
-def : BitConvert <v32i8, v8i32, VReg_256>;
 
 def : BitConvert <v16i32, v16f32, VReg_512>;
 def : BitConvert <v16f32, v16i32, VReg_512>;
diff --git a/lib/Target/AMDGPU/SIRegisterInfo.td b/lib/Target/AMDGPU/SIRegisterInfo.td
index bfaf93709d8..5ab8830d1d9 100644
--- a/lib/Target/AMDGPU/SIRegisterInfo.td
+++ b/lib/Target/AMDGPU/SIRegisterInfo.td
@@ -208,7 +208,7 @@ def SReg_128 : RegisterClass<"AMDGPU", [v4i32, v16i8, v2i64], 32, (add SGPR_128)
   let CopyCost = 2;
 }
 
-def SReg_256 : RegisterClass<"AMDGPU", [v32i8, v8i32, v8f32], 32, (add SGPR_256)> {
+def SReg_256 : RegisterClass<"AMDGPU", [v8i32, v8f32], 32, (add SGPR_256)> {
   // Requires 4 s_mov_b64 to copy
   let CopyCost = 4;
 }
@@ -236,7 +236,7 @@ def VReg_128 : RegisterClass<"AMDGPU", [v4i32, v4f32, v2i64, v2f64], 32, (add VG
   let CopyCost = 4;
 }
 
-def VReg_256 : RegisterClass<"AMDGPU", [v32i8, v8i32, v8f32], 32, (add VGPR_256)> {
+def VReg_256 : RegisterClass<"AMDGPU", [v8i32, v8f32], 32, (add VGPR_256)> {
   let CopyCost = 8;
 }
 
diff --git a/test/CodeGen/AMDGPU/extract-vector-elt-i8.ll b/test/CodeGen/AMDGPU/extract-vector-elt-i8.ll
new file mode 100644
index 00000000000..ff1006b80ef
--- /dev/null
+++ b/test/CodeGen/AMDGPU/extract-vector-elt-i8.ll
@@ -0,0 +1,111 @@
+; RUN: llc -march=amdgcn -mcpu=SI -verify-machineinstrs< %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs< %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
+
+; FUNC-LABEL: {{^}}extract_vector_elt_v1i8:
+; SI: buffer_load_ubyte
+; SI: buffer_store_byte
+define void @extract_vector_elt_v1i8(i8 addrspace(1)* %out, <1 x i8> %foo) #0 {
+  %p0 = extractelement <1 x i8> %foo, i32 0
+  store i8 %p0, i8 addrspace(1)* %out
+  ret void
+}
+
+; FUNC-LABEL: {{^}}extract_vector_elt_v2i8:
+; SI: buffer_load_ubyte
+; SI: buffer_load_ubyte
+; SI: buffer_store_byte
+; SI: buffer_store_byte
+define void @extract_vector_elt_v2i8(i8 addrspace(1)* %out, <2 x i8> %foo) #0 {
+  %p0 = extractelement <2 x i8> %foo, i32 0
+  %p1 = extractelement <2 x i8> %foo, i32 1
+  %out1 = getelementptr i8, i8 addrspace(1)* %out, i32 1
+  store i8 %p1, i8 addrspace(1)* %out
+  store i8 %p0, i8 addrspace(1)* %out1
+  ret void
+}
+
+; FUNC-LABEL: {{^}}extract_vector_elt_v3i8:
+; SI: buffer_load_ubyte
+; SI: buffer_load_ubyte
+; SI: buffer_store_byte
+; SI: buffer_store_byte
+define void @extract_vector_elt_v3i8(i8 addrspace(1)* %out, <3 x i8> %foo) #0 {
+  %p0 = extractelement <3 x i8> %foo, i32 0
+  %p1 = extractelement <3 x i8> %foo, i32 2
+  %out1 = getelementptr i8, i8 addrspace(1)* %out, i32 1
+  store i8 %p1, i8 addrspace(1)* %out
+  store i8 %p0, i8 addrspace(1)* %out1
+  ret void
+}
+
+; FUNC-LABEL: {{^}}extract_vector_elt_v4i8:
+; SI: buffer_load_ubyte
+; SI: buffer_load_ubyte
+; SI: buffer_store_byte
+; SI: buffer_store_byte
+define void @extract_vector_elt_v4i8(i8 addrspace(1)* %out, <4 x i8> %foo) #0 {
+  %p0 = extractelement <4 x i8> %foo, i32 0
+  %p1 = extractelement <4 x i8> %foo, i32 2
+  %out1 = getelementptr i8, i8 addrspace(1)* %out, i32 1
+  store i8 %p1, i8 addrspace(1)* %out
+  store i8 %p0, i8 addrspace(1)* %out1
+  ret void
+}
+
+; FUNC-LABEL: {{^}}extract_vector_elt_v8i8:
+; SI: buffer_load_ubyte
+; SI: buffer_load_ubyte
+; SI: buffer_store_byte
+; SI: buffer_store_byte
+define void @extract_vector_elt_v8i8(i8 addrspace(1)* %out, <8 x i8> %foo) #0 {
+  %p0 = extractelement <8 x i8> %foo, i32 0
+  %p1 = extractelement <8 x i8> %foo, i32 2
+  %out1 = getelementptr i8, i8 addrspace(1)* %out, i32 1
+  store i8 %p1, i8 addrspace(1)* %out
+  store i8 %p0, i8 addrspace(1)* %out1
+  ret void
+}
+
+; FUNC-LABEL: {{^}}extract_vector_elt_v16i8:
+; SI: buffer_load_ubyte
+; SI: buffer_load_ubyte
+; SI: buffer_store_byte
+; SI: buffer_store_byte
+define void @extract_vector_elt_v16i8(i8 addrspace(1)* %out, <16 x i8> %foo) #0 {
+  %p0 = extractelement <16 x i8> %foo, i32 0
+  %p1 = extractelement <16 x i8> %foo, i32 2
+  %out1 = getelementptr i8, i8 addrspace(1)* %out, i32 1
+  store i8 %p1, i8 addrspace(1)* %out
+  store i8 %p0, i8 addrspace(1)* %out1
+  ret void
+}
+
+; FUNC-LABEL: {{^}}extract_vector_elt_v32i8:
+; SI: buffer_load_ubyte
+; SI: buffer_load_ubyte
+; SI: buffer_store_byte
+; SI: buffer_store_byte
+define void @extract_vector_elt_v32i8(i8 addrspace(1)* %out, <32 x i8> %foo) #0 {
+  %p0 = extractelement <32 x i8> %foo, i32 0
+  %p1 = extractelement <32 x i8> %foo, i32 2
+  %out1 = getelementptr i8, i8 addrspace(1)* %out, i32 1
+  store i8 %p1, i8 addrspace(1)* %out
+  store i8 %p0, i8 addrspace(1)* %out1
+  ret void
+}
+
+; FUNC-LABEL: {{^}}extract_vector_elt_v64i8:
+; SI: buffer_load_ubyte
+; SI: buffer_load_ubyte
+; SI: buffer_store_byte
+; SI: buffer_store_byte
+define void @extract_vector_elt_v64i8(i8 addrspace(1)* %out, <64 x i8> %foo) #0 {
+  %p0 = extractelement <64 x i8> %foo, i32 0
+  %p1 = extractelement <64 x i8> %foo, i32 2
+  %out1 = getelementptr i8, i8 addrspace(1)* %out, i32 1
+  store i8 %p1, i8 addrspace(1)* %out
+  store i8 %p0, i8 addrspace(1)* %out1
+  ret void
+}
+
+attributes #0 = { nounwind }
diff --git a/test/CodeGen/AMDGPU/llvm.SI.gather4.ll b/test/CodeGen/AMDGPU/llvm.SI.gather4.ll
index 275cb580bc9..13f12d0d0ad 100644
--- a/test/CodeGen/AMDGPU/llvm.SI.gather4.ll
+++ b/test/CodeGen/AMDGPU/llvm.SI.gather4.ll
@@ -5,7 +5,7 @@
 ;CHECK: image_gather4 {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_v2() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.v2i32(<2 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.v2i32(<2 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -18,7 +18,7 @@ main_body:
 ;CHECK: image_gather4 {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -31,7 +31,7 @@ main_body:
 ;CHECK: image_gather4_cl {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_cl() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.cl.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.cl.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -44,7 +44,7 @@ main_body:
 ;CHECK: image_gather4_l {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_l() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.l.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.l.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -57,7 +57,7 @@ main_body:
 ;CHECK: image_gather4_b {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_b() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.b.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.b.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -70,7 +70,7 @@ main_body:
 ;CHECK: image_gather4_b_cl {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_b_cl() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.b.cl.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.b.cl.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -83,7 +83,7 @@ main_body:
 ;CHECK: image_gather4_b_cl {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_b_cl_v8() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.b.cl.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.b.cl.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -96,7 +96,7 @@ main_body:
 ;CHECK: image_gather4_lz {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_lz_v2() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.lz.v2i32(<2 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.lz.v2i32(<2 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -109,7 +109,7 @@ main_body:
 ;CHECK: image_gather4_lz {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_lz() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.lz.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.lz.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -124,7 +124,7 @@ main_body:
 ;CHECK: image_gather4_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_o() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.o.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.o.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -137,7 +137,7 @@ main_body:
 ;CHECK: image_gather4_cl_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_cl_o() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.cl.o.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.cl.o.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -150,7 +150,7 @@ main_body:
 ;CHECK: image_gather4_cl_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_cl_o_v8() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.cl.o.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.cl.o.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -163,7 +163,7 @@ main_body:
 ;CHECK: image_gather4_l_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_l_o() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.l.o.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.l.o.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -176,7 +176,7 @@ main_body:
 ;CHECK: image_gather4_l_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_l_o_v8() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.l.o.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.l.o.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -189,7 +189,7 @@ main_body:
 ;CHECK: image_gather4_b_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_b_o() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.b.o.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.b.o.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -202,7 +202,7 @@ main_body:
 ;CHECK: image_gather4_b_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_b_o_v8() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.b.o.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.b.o.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -215,7 +215,7 @@ main_body:
 ;CHECK: image_gather4_b_cl_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_b_cl_o() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.b.cl.o.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.b.cl.o.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -228,7 +228,7 @@ main_body:
 ;CHECK: image_gather4_lz_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_lz_o() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.lz.o.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.lz.o.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -243,7 +243,7 @@ main_body:
 ;CHECK: image_gather4_c {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -256,7 +256,7 @@ main_body:
 ;CHECK: image_gather4_c_cl {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_cl() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.cl.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.cl.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -269,7 +269,7 @@ main_body:
 ;CHECK: image_gather4_c_cl {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_cl_v8() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.cl.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.cl.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -282,7 +282,7 @@ main_body:
 ;CHECK: image_gather4_c_l {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_l() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.l.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.l.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -295,7 +295,7 @@ main_body:
 ;CHECK: image_gather4_c_l {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_l_v8() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.l.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.l.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -308,7 +308,7 @@ main_body:
 ;CHECK: image_gather4_c_b {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_b() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.b.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.b.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -321,7 +321,7 @@ main_body:
 ;CHECK: image_gather4_c_b {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_b_v8() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.b.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.b.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -334,7 +334,7 @@ main_body:
 ;CHECK: image_gather4_c_b_cl {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_b_cl() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.b.cl.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.b.cl.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -347,7 +347,7 @@ main_body:
 ;CHECK: image_gather4_c_lz {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_lz() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.lz.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.lz.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -362,7 +362,7 @@ main_body:
 ;CHECK: image_gather4_c_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_o() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.o.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.o.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -375,7 +375,7 @@ main_body:
 ;CHECK: image_gather4_c_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_o_v8() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.o.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.o.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -388,7 +388,7 @@ main_body:
 ;CHECK: image_gather4_c_cl_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_cl_o() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.cl.o.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.cl.o.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -401,7 +401,7 @@ main_body:
 ;CHECK: image_gather4_c_l_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_l_o() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.l.o.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.l.o.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -414,7 +414,7 @@ main_body:
 ;CHECK: image_gather4_c_b_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_b_o() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.b.o.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.b.o.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -427,7 +427,7 @@ main_body:
 ;CHECK: image_gather4_c_b_cl_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_b_cl_o() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.b.cl.o.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.b.cl.o.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -440,7 +440,7 @@ main_body:
 ;CHECK: image_gather4_c_lz_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_lz_o() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.lz.o.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.lz.o.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -453,7 +453,7 @@ main_body:
 ;CHECK: image_gather4_c_lz_o {{v\[[0-9]+:[0-9]+\]}}, 1, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @gather4_c_lz_o_v8() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.gather4.c.lz.o.v8i32(<8 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.gather4.c.lz.o.v8i32(<8 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 1, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   %r2 = extractelement <4 x float> %r, i32 2
@@ -464,44 +464,44 @@ main_body:
 
 
 
-declare <4 x float> @llvm.SI.gather4.v2i32(<2 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.cl.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.l.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.b.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.b.cl.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.b.cl.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.lz.v2i32(<2 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.lz.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.v2i32(<2 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.cl.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.l.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.b.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.b.cl.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.b.cl.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.lz.v2i32(<2 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.lz.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
 
-declare <4 x float> @llvm.SI.gather4.o.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.cl.o.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.cl.o.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.l.o.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.l.o.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.b.o.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.b.o.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.b.cl.o.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.lz.o.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.o.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.cl.o.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.cl.o.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.l.o.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.l.o.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.b.o.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.b.o.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.b.cl.o.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.lz.o.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
 
-declare <4 x float> @llvm.SI.gather4.c.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.cl.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.cl.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.l.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.l.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.b.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.b.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.b.cl.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.lz.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.cl.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.cl.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.l.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.l.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.b.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.b.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.b.cl.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.lz.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
 
-declare <4 x float> @llvm.SI.gather4.c.o.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.o.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.cl.o.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.l.o.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.b.o.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.b.cl.o.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.lz.o.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.gather4.c.lz.o.v8i32(<8 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.o.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.o.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.cl.o.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.l.o.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.b.o.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.b.cl.o.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.lz.o.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.gather4.c.lz.o.v8i32(<8 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
 
 declare void @llvm.SI.export(i32, i32, i32, i32, i32, float, float, float, float)
 
diff --git a/test/CodeGen/AMDGPU/llvm.SI.getlod.ll b/test/CodeGen/AMDGPU/llvm.SI.getlod.ll
index 06ee98e91b3..4d5b87e378b 100644
--- a/test/CodeGen/AMDGPU/llvm.SI.getlod.ll
+++ b/test/CodeGen/AMDGPU/llvm.SI.getlod.ll
@@ -5,7 +5,7 @@
 ;CHECK: image_get_lod {{v\[[0-9]+:[0-9]+\]}}, 3, 0, 0, -1, 0, 0, 0, 0, {{v[0-9]+}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @getlod() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.getlod.i32(i32 undef, <32 x i8> undef, <16 x i8> undef, i32 15, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.getlod.i32(i32 undef, <8 x i32> undef, <4 x i32> undef, i32 15, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   call void @llvm.SI.export(i32 15, i32 1, i32 1, i32 0, i32 1, float %r0, float %r1, float %r0, float %r1)
@@ -16,7 +16,7 @@ main_body:
 ;CHECK: image_get_lod {{v\[[0-9]+:[0-9]+\]}}, 3, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @getlod_v2() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.getlod.v2i32(<2 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 15, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.getlod.v2i32(<2 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 15, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   call void @llvm.SI.export(i32 15, i32 1, i32 1, i32 0, i32 1, float %r0, float %r1, float %r0, float %r1)
@@ -27,7 +27,7 @@ main_body:
 ;CHECK: image_get_lod {{v\[[0-9]+:[0-9]+\]}}, 3, 0, 0, -1, 0, 0, 0, 0, {{v\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}
 define void @getlod_v4() #0 {
 main_body:
-  %r = call <4 x float> @llvm.SI.getlod.v4i32(<4 x i32> undef, <32 x i8> undef, <16 x i8> undef, i32 15, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
+  %r = call <4 x float> @llvm.SI.getlod.v4i32(<4 x i32> undef, <8 x i32> undef, <4 x i32> undef, i32 15, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0)
   %r0 = extractelement <4 x float> %r, i32 0
   %r1 = extractelement <4 x float> %r, i32 1
   call void @llvm.SI.export(i32 15, i32 1, i32 1, i32 0, i32 1, float %r0, float %r1, float %r0, float %r1)
@@ -35,9 +35,9 @@ main_body:
 }
 
 
-declare <4 x float> @llvm.SI.getlod.i32(i32, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.getlod.v2i32(<2 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
-declare <4 x float> @llvm.SI.getlod.v4i32(<4 x i32>, <32 x i8>, <16 x i8>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.getlod.i32(i32, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.getlod.v2i32(<2 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
+declare <4 x float> @llvm.SI.getlod.v4i32(<4 x i32>, <8 x i32>, <4 x i32>, i32, i32, i32, i32, i32, i32, i32, i32) #1
 
 declare void @llvm.SI.export(i32, i32, i32, i32, i32, float, float, float, float)