diff --git a/llvm/test/CodeGen/AMDGPU/flat-scratch-i8-i16.ll b/llvm/test/CodeGen/AMDGPU/flat-scratch-i8-i16.ll
new file mode 100644
index 00000000000000..499ee9acbbf34a
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/flat-scratch-i8-i16.ll
@@ -0,0 +1,1068 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=amdgcn -march=amdgcn -mcpu=gfx1010 -mattr=+enable-flat-scratch -verify-machineinstrs < %s | FileCheck --check-prefixes=GFX10 %s
+; RUN: llc -mtriple=amdgcn -march=amdgcn -mcpu=gfx1100 -mattr=+enable-flat-scratch -verify-machineinstrs < %s | FileCheck --check-prefixes=GFX11 %s
+
+; vgpr offset
+
+define amdgpu_ps void @test_scratch_load_i8_zext_v(ptr addrspace(5) %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_zext_v:
+; GFX10:       ; %bb.0:
+; GFX10-NEXT:    s_add_u32 s0, s0, s2
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    scratch_load_ubyte v0, v0, off offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v0
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_zext_v:
+; GFX11:       ; %bb.0:
+; GFX11-NEXT:    scratch_load_u8 v0, v0, off offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v0
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+  %gep = getelementptr inbounds i8, ptr addrspace(5) %in, i32 1
+  %load = load i8, ptr addrspace(5) %gep, align 4
+  %ext = zext i8 %load to i32
+  store i32 %ext, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i8_sext_v(ptr addrspace(5) %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_sext_v:
+; GFX10:       ; %bb.0:
+; GFX10-NEXT:    s_add_u32 s0, s0, s2
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    scratch_load_sbyte v0, v0, off offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v0
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_sext_v:
+; GFX11:       ; %bb.0:
+; GFX11-NEXT:    scratch_load_i8 v0, v0, off offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v0
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+  %gep = getelementptr inbounds i8, ptr addrspace(5) %in, i32 1
+  %load = load i8, ptr addrspace(5) %gep, align 4
+  %ext = sext i8 %load to i32
+  store i32 %ext, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i16_zext_v(ptr addrspace(5) %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i16_zext_v:
+; GFX10:       ; %bb.0:
+; GFX10-NEXT:    s_add_u32 s0, s0, s2
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    scratch_load_ushort v0, v0, off offset:2
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v0
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i16_zext_v:
+; GFX11:       ; %bb.0:
+; GFX11-NEXT:    scratch_load_u16 v0, v0, off offset:2
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v0
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+  %gep = getelementptr inbounds i16, ptr addrspace(5) %in, i32 1
+  %load = load i16, ptr addrspace(5) %gep, align 4
+  %ext = zext i16 %load to i32
+  store i32 %ext, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i16_sext_v(ptr addrspace(5) %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i16_sext_v:
+; GFX10:       ; %bb.0:
+; GFX10-NEXT:    s_add_u32 s0, s0, s2
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    scratch_load_sshort v0, v0, off offset:2
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v0
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i16_sext_v:
+; GFX11:       ; %bb.0:
+; GFX11-NEXT:    scratch_load_i16 v0, v0, off offset:2
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v0
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+  %gep = getelementptr inbounds i16, ptr addrspace(5) %in, i32 1
+  %load = load i16, ptr addrspace(5) %gep, align 4
+  %ext = sext i16 %load to i32
+  store i32 %ext, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i8_zext_to_d16_lo_v(ptr addrspace(5) %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_zext_to_d16_lo_v:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s2
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_mov_b32_e32 v3, 0xffff0000
+; GFX10-NEXT:    scratch_load_ubyte_d16 v3, v0, off offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v3
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_zext_to_d16_lo_v:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_mov_b32_e32 v3, 0xffff0000
+; GFX11-NEXT:    scratch_load_d16_u8 v3, v0, off offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v3
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %gep = getelementptr i8, ptr addrspace(5) %in, i64 1
+  %load_lo = load i8, ptr addrspace(5) %gep
+  %ext = zext i8 %load_lo to i16
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %ext, i32 0
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i8_sext_to_d16_lo_v(ptr addrspace(5) %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_sext_to_d16_lo_v:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s2
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_mov_b32_e32 v3, 0xffff0000
+; GFX10-NEXT:    scratch_load_sbyte_d16 v3, v0, off offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v3
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_sext_to_d16_lo_v:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_mov_b32_e32 v3, 0xffff0000
+; GFX11-NEXT:    scratch_load_d16_i8 v3, v0, off offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v3
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %gep = getelementptr i8, ptr addrspace(5) %in, i64 1
+  %load_lo = load i8, ptr addrspace(5) %gep
+  %ext = sext i8 %load_lo to i16
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %ext, i32 0
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i16_to_d16_lo_v(ptr addrspace(5) %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i16_to_d16_lo_v:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s2
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_mov_b32_e32 v3, 0xffff0000
+; GFX10-NEXT:    scratch_load_short_d16 v3, v0, off offset:2
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v3
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i16_to_d16_lo_v:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_mov_b32_e32 v3, 0xffff0000
+; GFX11-NEXT:    scratch_load_d16_b16 v3, v0, off offset:2
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v3
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %gep = getelementptr i16, ptr addrspace(5) %in, i64 1
+  %load_lo = load i16, ptr addrspace(5) %gep
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %load_lo, i32 0
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+
+define amdgpu_ps void @test_scratch_load_i8_zext_to_d16_hi_v(ptr addrspace(5) %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_zext_to_d16_hi_v:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s2
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_mov_b32_e32 v3, -1
+; GFX10-NEXT:    scratch_load_ubyte_d16_hi v3, v0, off offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v3
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_zext_to_d16_hi_v:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_mov_b32_e32 v3, -1
+; GFX11-NEXT:    scratch_load_d16_hi_u8 v3, v0, off offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v3
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %gep = getelementptr i8, ptr addrspace(5) %in, i64 1
+  %load_lo = load i8, ptr addrspace(5) %gep
+  %ext = zext i8 %load_lo to i16
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %ext, i32 1
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i8_sext_to_d16_hi_v(ptr addrspace(5) %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_sext_to_d16_hi_v:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s2
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_mov_b32_e32 v3, -1
+; GFX10-NEXT:    scratch_load_sbyte_d16_hi v3, v0, off offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v3
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_sext_to_d16_hi_v:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_mov_b32_e32 v3, -1
+; GFX11-NEXT:    scratch_load_d16_hi_i8 v3, v0, off offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v3
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %gep = getelementptr i8, ptr addrspace(5) %in, i64 1
+  %load_lo = load i8, ptr addrspace(5) %gep
+  %ext = sext i8 %load_lo to i16
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %ext, i32 1
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i16_to_d16_hi_v(ptr addrspace(5) %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i16_to_d16_hi_v:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s2
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_mov_b32_e32 v3, -1
+; GFX10-NEXT:    scratch_load_short_d16_hi v3, v0, off offset:2
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v3
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i16_to_d16_hi_v:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_mov_b32_e32 v3, -1
+; GFX11-NEXT:    scratch_load_d16_hi_b16 v3, v0, off offset:2
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v3
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %gep = getelementptr i16, ptr addrspace(5) %in, i64 1
+  %load_lo = load i16, ptr addrspace(5) %gep
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %load_lo, i32 1
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_store_b8_from_d16_hi_v(ptr %in, ptr addrspace(5) %out) {
+; GFX10-LABEL: test_scratch_store_b8_from_d16_hi_v:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s2
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    flat_load_dword v0, v[0:1]
+; GFX10-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX10-NEXT:    scratch_store_byte_d16_hi v2, v0, off offset:4
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_store_b8_from_d16_hi_v:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    flat_load_b32 v0, v[0:1]
+; GFX11-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX11-NEXT:    scratch_store_d16_hi_b8 v2, v0, off offset:4
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %load = load <4 x i8>, ptr %in
+  %element = extractelement <4 x i8> %load, i32 2
+  %gep = getelementptr <4 x i8>, ptr addrspace(5) %out, i64 1
+  store i8 %element, ptr addrspace(5) %gep, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_store_b16_from_d16_hi_v(ptr %in, ptr addrspace(5) %out) {
+; GFX10-LABEL: test_scratch_store_b16_from_d16_hi_v:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s2
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    flat_load_dword v0, v[0:1]
+; GFX10-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX10-NEXT:    scratch_store_short_d16_hi v2, v0, off offset:2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_store_b16_from_d16_hi_v:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    flat_load_b32 v0, v[0:1]
+; GFX11-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX11-NEXT:    scratch_store_d16_hi_b16 v2, v0, off offset:2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %load = load <2 x i16>, ptr %in
+  %element = extractelement <2 x i16> %load, i32 1
+  %gep = getelementptr <2 x i8>, ptr addrspace(5) %out, i64 1
+  store i16 %element, ptr addrspace(5) %gep, align 4
+  ret void
+}
+
+
+
+
+; sgpr offset
+
+define amdgpu_ps void @test_scratch_load_i8_zext_s(ptr addrspace(5) inreg %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_zext_s:
+; GFX10:       ; %bb.0:
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    scratch_load_ubyte v2, off, s2 offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[0:1], v2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_zext_s:
+; GFX11:       ; %bb.0:
+; GFX11-NEXT:    scratch_load_u8 v2, off, s0 offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[0:1], v2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+  %gep = getelementptr inbounds i8, ptr addrspace(5) %in, i32 1
+  %load = load i8, ptr addrspace(5) %gep, align 4
+  %ext = zext i8 %load to i32
+  store i32 %ext, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i8_sext_s(ptr addrspace(5) inreg %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_sext_s:
+; GFX10:       ; %bb.0:
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    scratch_load_sbyte v2, off, s2 offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[0:1], v2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_sext_s:
+; GFX11:       ; %bb.0:
+; GFX11-NEXT:    scratch_load_i8 v2, off, s0 offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[0:1], v2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+  %gep = getelementptr inbounds i8, ptr addrspace(5) %in, i32 1
+  %load = load i8, ptr addrspace(5) %gep, align 4
+  %ext = sext i8 %load to i32
+  store i32 %ext, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i16_zext_s(ptr addrspace(5) inreg %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i16_zext_s:
+; GFX10:       ; %bb.0:
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    scratch_load_ushort v2, off, s2 offset:2
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[0:1], v2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i16_zext_s:
+; GFX11:       ; %bb.0:
+; GFX11-NEXT:    scratch_load_u16 v2, off, s0 offset:2
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[0:1], v2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+  %gep = getelementptr inbounds i16, ptr addrspace(5) %in, i32 1
+  %load = load i16, ptr addrspace(5) %gep, align 4
+  %ext = zext i16 %load to i32
+  store i32 %ext, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i16_sext_s(ptr addrspace(5) inreg %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i16_sext_s:
+; GFX10:       ; %bb.0:
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    scratch_load_sshort v2, off, s2 offset:2
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[0:1], v2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i16_sext_s:
+; GFX11:       ; %bb.0:
+; GFX11-NEXT:    scratch_load_i16 v2, off, s0 offset:2
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[0:1], v2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+  %gep = getelementptr inbounds i16, ptr addrspace(5) %in, i32 1
+  %load = load i16, ptr addrspace(5) %gep, align 4
+  %ext = sext i16 %load to i32
+  store i32 %ext, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i8_zext_to_d16_lo_s(ptr addrspace(5) inreg %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_zext_to_d16_lo_s:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_mov_b32_e32 v2, 0xffff0000
+; GFX10-NEXT:    scratch_load_ubyte_d16 v2, off, s2 offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[0:1], v2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_zext_to_d16_lo_s:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_mov_b32_e32 v2, 0xffff0000
+; GFX11-NEXT:    scratch_load_d16_u8 v2, off, s0 offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[0:1], v2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %gep = getelementptr i8, ptr addrspace(5) %in, i64 1
+  %load_lo = load i8, ptr addrspace(5) %gep
+  %ext = zext i8 %load_lo to i16
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %ext, i32 0
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i8_sext_to_d16_lo_s(ptr addrspace(5) inreg %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_sext_to_d16_lo_s:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_mov_b32_e32 v2, 0xffff0000
+; GFX10-NEXT:    scratch_load_sbyte_d16 v2, off, s2 offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[0:1], v2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_sext_to_d16_lo_s:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_mov_b32_e32 v2, 0xffff0000
+; GFX11-NEXT:    scratch_load_d16_i8 v2, off, s0 offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[0:1], v2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %gep = getelementptr i8, ptr addrspace(5) %in, i64 1
+  %load_lo = load i8, ptr addrspace(5) %gep
+  %ext = sext i8 %load_lo to i16
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %ext, i32 0
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i16_to_d16_lo_s(ptr addrspace(5) inreg %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i16_to_d16_lo_s:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_mov_b32_e32 v2, 0xffff0000
+; GFX10-NEXT:    scratch_load_short_d16 v2, off, s2 offset:2
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[0:1], v2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i16_to_d16_lo_s:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_mov_b32_e32 v2, 0xffff0000
+; GFX11-NEXT:    scratch_load_d16_b16 v2, off, s0 offset:2
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[0:1], v2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %gep = getelementptr i16, ptr addrspace(5) %in, i64 1
+  %load_lo = load i16, ptr addrspace(5) %gep
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %load_lo, i32 0
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+
+define amdgpu_ps void @test_scratch_load_i8_zext_to_d16_hi_s(ptr addrspace(5) inreg %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_zext_to_d16_hi_s:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_mov_b32_e32 v2, -1
+; GFX10-NEXT:    scratch_load_ubyte_d16_hi v2, off, s2 offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[0:1], v2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_zext_to_d16_hi_s:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_mov_b32_e32 v2, -1
+; GFX11-NEXT:    scratch_load_d16_hi_u8 v2, off, s0 offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[0:1], v2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %gep = getelementptr i8, ptr addrspace(5) %in, i64 1
+  %load_lo = load i8, ptr addrspace(5) %gep
+  %ext = zext i8 %load_lo to i16
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %ext, i32 1
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i8_sext_to_d16_hi_s(ptr addrspace(5) inreg %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_sext_to_d16_hi_s:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_mov_b32_e32 v2, -1
+; GFX10-NEXT:    scratch_load_sbyte_d16_hi v2, off, s2 offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[0:1], v2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_sext_to_d16_hi_s:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_mov_b32_e32 v2, -1
+; GFX11-NEXT:    scratch_load_d16_hi_i8 v2, off, s0 offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[0:1], v2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %gep = getelementptr i8, ptr addrspace(5) %in, i64 1
+  %load_lo = load i8, ptr addrspace(5) %gep
+  %ext = sext i8 %load_lo to i16
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %ext, i32 1
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i16_to_d16_hi_s(ptr addrspace(5) inreg %in, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i16_to_d16_hi_s:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_mov_b32_e32 v2, -1
+; GFX10-NEXT:    scratch_load_short_d16_hi v2, off, s2 offset:2
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[0:1], v2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i16_to_d16_hi_s:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_mov_b32_e32 v2, -1
+; GFX11-NEXT:    scratch_load_d16_hi_b16 v2, off, s0 offset:2
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[0:1], v2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %gep = getelementptr i16, ptr addrspace(5) %in, i64 1
+  %load_lo = load i16, ptr addrspace(5) %gep
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %load_lo, i32 1
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_store_b8_from_d16_hi_s(ptr %in, ptr addrspace(5) inreg %out) {
+; GFX10-LABEL: test_scratch_store_b8_from_d16_hi_s:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    flat_load_dword v0, v[0:1]
+; GFX10-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX10-NEXT:    scratch_store_byte_d16_hi off, v0, s2 offset:4
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_store_b8_from_d16_hi_s:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    flat_load_b32 v0, v[0:1]
+; GFX11-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX11-NEXT:    scratch_store_d16_hi_b8 off, v0, s0 offset:4
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %load = load <4 x i8>, ptr %in
+  %element = extractelement <4 x i8> %load, i32 2
+  %gep = getelementptr <4 x i8>, ptr addrspace(5) %out, i64 1
+  store i8 %element, ptr addrspace(5) %gep, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_store_b16_from_d16_hi_s(ptr %in, ptr addrspace(5) inreg %out) {
+; GFX10-LABEL: test_scratch_store_b16_from_d16_hi_s:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    flat_load_dword v0, v[0:1]
+; GFX10-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX10-NEXT:    scratch_store_short_d16_hi off, v0, s2 offset:2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_store_b16_from_d16_hi_s:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    flat_load_b32 v0, v[0:1]
+; GFX11-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX11-NEXT:    scratch_store_d16_hi_b16 off, v0, s0 offset:2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %load = load <2 x i16>, ptr %in
+  %element = extractelement <2 x i16> %load, i32 1
+  %gep = getelementptr <2 x i8>, ptr addrspace(5) %out, i64 1
+  store i16 %element, ptr addrspace(5) %gep, align 4
+  ret void
+}
+
+
+
+
+; sgpr + vgpr offset
+
+define amdgpu_ps void @test_scratch_load_i8_zext_svs(ptr addrspace(5) inreg %in, i32 %voffset, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_zext_svs:
+; GFX10:       ; %bb.0:
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_lshl_add_u32 v0, v0, 2, s2
+; GFX10-NEXT:    scratch_load_ubyte v0, v0, off offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v0
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_zext_svs:
+; GFX11:       ; %bb.0:
+; GFX11-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
+; GFX11-NEXT:    scratch_load_u8 v0, v0, s0 offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v0
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+  %voffset4 = mul i32 %voffset, 4
+  %gep0 = getelementptr inbounds i8, ptr addrspace(5) %in, i32 %voffset4
+  %gep = getelementptr inbounds i8, ptr addrspace(5) %gep0, i32 1
+  %load = load i8, ptr addrspace(5) %gep, align 4
+  %ext = zext i8 %load to i32
+  store i32 %ext, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i8_sext_svs(ptr addrspace(5) inreg %in, i32 %voffset, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_sext_svs:
+; GFX10:       ; %bb.0:
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_lshl_add_u32 v0, v0, 2, s2
+; GFX10-NEXT:    scratch_load_sbyte v0, v0, off offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v0
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_sext_svs:
+; GFX11:       ; %bb.0:
+; GFX11-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
+; GFX11-NEXT:    scratch_load_i8 v0, v0, s0 offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v0
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+  %voffset4 = mul i32 %voffset, 4
+  %gep0 = getelementptr inbounds i8, ptr addrspace(5) %in, i32 %voffset4
+  %gep = getelementptr inbounds i8, ptr addrspace(5) %gep0, i32 1
+  %load = load i8, ptr addrspace(5) %gep, align 4
+  %ext = sext i8 %load to i32
+  store i32 %ext, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i16_zext_svs(ptr addrspace(5) inreg %in, i32 %voffset, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i16_zext_svs:
+; GFX10:       ; %bb.0:
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_lshl_add_u32 v0, v0, 2, s2
+; GFX10-NEXT:    scratch_load_ushort v0, v0, off offset:2
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v0
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i16_zext_svs:
+; GFX11:       ; %bb.0:
+; GFX11-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
+; GFX11-NEXT:    scratch_load_u16 v0, v0, s0 offset:2
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v0
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+  %voffset4 = mul i32 %voffset, 4
+  %gep0 = getelementptr inbounds i8, ptr addrspace(5) %in, i32 %voffset4
+  %gep = getelementptr inbounds i16, ptr addrspace(5) %gep0, i32 1
+  %load = load i16, ptr addrspace(5) %gep, align 4
+  %ext = zext i16 %load to i32
+  store i32 %ext, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i16_sext_svs(ptr addrspace(5) inreg %in, i32 %voffset, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i16_sext_svs:
+; GFX10:       ; %bb.0:
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_lshl_add_u32 v0, v0, 2, s2
+; GFX10-NEXT:    scratch_load_sshort v0, v0, off offset:2
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v0
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i16_sext_svs:
+; GFX11:       ; %bb.0:
+; GFX11-NEXT:    v_lshlrev_b32_e32 v0, 2, v0
+; GFX11-NEXT:    scratch_load_i16 v0, v0, s0 offset:2
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v0
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+  %voffset4 = mul i32 %voffset, 4
+  %gep0 = getelementptr inbounds i8, ptr addrspace(5) %in, i32 %voffset4
+  %gep = getelementptr inbounds i16, ptr addrspace(5) %gep0, i32 1
+  %load = load i16, ptr addrspace(5) %gep, align 4
+  %ext = sext i16 %load to i32
+  store i32 %ext, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i8_zext_to_d16_lo_svs(ptr addrspace(5) inreg %in, i32 %voffset, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_zext_to_d16_lo_svs:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_lshl_add_u32 v0, v0, 2, s2
+; GFX10-NEXT:    v_mov_b32_e32 v3, 0xffff0000
+; GFX10-NEXT:    scratch_load_ubyte_d16 v3, v0, off offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v3
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_zext_to_d16_lo_svs:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_dual_mov_b32 v3, 0xffff0000 :: v_dual_lshlrev_b32 v0, 2, v0
+; GFX11-NEXT:    scratch_load_d16_u8 v3, v0, s0 offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v3
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %voffset4 = mul i32 %voffset, 4
+  %gep0 = getelementptr inbounds i8, ptr addrspace(5) %in, i32 %voffset4
+  %gep = getelementptr i8, ptr addrspace(5) %gep0, i64 1
+  %load_lo = load i8, ptr addrspace(5) %gep
+  %ext = zext i8 %load_lo to i16
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %ext, i32 0
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i8_sext_to_d16_lo_svs(ptr addrspace(5) inreg %in, i32 %voffset, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_sext_to_d16_lo_svs:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_lshl_add_u32 v0, v0, 2, s2
+; GFX10-NEXT:    v_mov_b32_e32 v3, 0xffff0000
+; GFX10-NEXT:    scratch_load_sbyte_d16 v3, v0, off offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v3
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_sext_to_d16_lo_svs:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_dual_mov_b32 v3, 0xffff0000 :: v_dual_lshlrev_b32 v0, 2, v0
+; GFX11-NEXT:    scratch_load_d16_i8 v3, v0, s0 offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v3
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %voffset4 = mul i32 %voffset, 4
+  %gep0 = getelementptr inbounds i8, ptr addrspace(5) %in, i32 %voffset4
+  %gep = getelementptr i8, ptr addrspace(5) %gep0, i64 1
+  %load_lo = load i8, ptr addrspace(5) %gep
+  %ext = sext i8 %load_lo to i16
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %ext, i32 0
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i16_to_d16_lo_svs(ptr addrspace(5) inreg %in, i32 %voffset, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i16_to_d16_lo_svs:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_lshl_add_u32 v0, v0, 2, s2
+; GFX10-NEXT:    v_mov_b32_e32 v3, 0xffff0000
+; GFX10-NEXT:    scratch_load_short_d16 v3, v0, off offset:2
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v3
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i16_to_d16_lo_svs:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_dual_mov_b32 v3, 0xffff0000 :: v_dual_lshlrev_b32 v0, 2, v0
+; GFX11-NEXT:    scratch_load_d16_b16 v3, v0, s0 offset:2
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v3
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %voffset4 = mul i32 %voffset, 4
+  %gep0 = getelementptr inbounds i8, ptr addrspace(5) %in, i32 %voffset4
+  %gep = getelementptr i16, ptr addrspace(5) %gep0, i64 1
+  %load_lo = load i16, ptr addrspace(5) %gep
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %load_lo, i32 0
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+
+define amdgpu_ps void @test_scratch_load_i8_zext_to_d16_hi_svs(ptr addrspace(5) inreg %in, i32 %voffset, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_zext_to_d16_hi_svs:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_lshl_add_u32 v0, v0, 2, s2
+; GFX10-NEXT:    v_mov_b32_e32 v3, -1
+; GFX10-NEXT:    scratch_load_ubyte_d16_hi v3, v0, off offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v3
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_zext_to_d16_hi_svs:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_dual_mov_b32 v3, -1 :: v_dual_lshlrev_b32 v0, 2, v0
+; GFX11-NEXT:    scratch_load_d16_hi_u8 v3, v0, s0 offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v3
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %voffset4 = mul i32 %voffset, 4
+  %gep0 = getelementptr inbounds i8, ptr addrspace(5) %in, i32 %voffset4
+  %gep = getelementptr i8, ptr addrspace(5) %gep0, i64 1
+  %load_lo = load i8, ptr addrspace(5) %gep
+  %ext = zext i8 %load_lo to i16
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %ext, i32 1
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i8_sext_to_d16_hi_svs(ptr addrspace(5) inreg %in, i32 %voffset, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i8_sext_to_d16_hi_svs:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_lshl_add_u32 v0, v0, 2, s2
+; GFX10-NEXT:    v_mov_b32_e32 v3, -1
+; GFX10-NEXT:    scratch_load_sbyte_d16_hi v3, v0, off offset:1
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v3
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i8_sext_to_d16_hi_svs:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_dual_mov_b32 v3, -1 :: v_dual_lshlrev_b32 v0, 2, v0
+; GFX11-NEXT:    scratch_load_d16_hi_i8 v3, v0, s0 offset:1
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v3
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %voffset4 = mul i32 %voffset, 4
+  %gep0 = getelementptr inbounds i8, ptr addrspace(5) %in, i32 %voffset4
+  %gep = getelementptr i8, ptr addrspace(5) %gep0, i64 1
+  %load_lo = load i8, ptr addrspace(5) %gep
+  %ext = sext i8 %load_lo to i16
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %ext, i32 1
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_load_i16_to_d16_hi_svs(ptr addrspace(5) inreg %in, i32 %voffset, ptr %out) {
+; GFX10-LABEL: test_scratch_load_i16_to_d16_hi_svs:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    v_lshl_add_u32 v0, v0, 2, s2
+; GFX10-NEXT:    v_mov_b32_e32 v3, -1
+; GFX10-NEXT:    scratch_load_short_d16_hi v3, v0, off offset:2
+; GFX10-NEXT:    s_waitcnt vmcnt(0)
+; GFX10-NEXT:    flat_store_dword v[1:2], v3
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_load_i16_to_d16_hi_svs:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    v_dual_mov_b32 v3, -1 :: v_dual_lshlrev_b32 v0, 2, v0
+; GFX11-NEXT:    scratch_load_d16_hi_b16 v3, v0, s0 offset:2
+; GFX11-NEXT:    s_waitcnt vmcnt(0)
+; GFX11-NEXT:    flat_store_b32 v[1:2], v3
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %voffset4 = mul i32 %voffset, 4
+  %gep0 = getelementptr inbounds i8, ptr addrspace(5) %in, i32 %voffset4
+  %gep = getelementptr i16, ptr addrspace(5) %gep0, i64 1
+  %load_lo = load i16, ptr addrspace(5) %gep
+  %result = insertelement <2 x i16> <i16 -1, i16 -1>, i16 %load_lo, i32 1
+  store <2 x i16> %result, ptr %out, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_store_b8_from_d16_hi_svs(ptr %in, ptr addrspace(5) inreg %out, i32 %voffset) {
+; GFX10-LABEL: test_scratch_store_b8_from_d16_hi_svs:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    flat_load_dword v0, v[0:1]
+; GFX10-NEXT:    v_lshl_add_u32 v1, v2, 2, s2
+; GFX10-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX10-NEXT:    scratch_store_byte_d16_hi v1, v0, off offset:4
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_store_b8_from_d16_hi_svs:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    flat_load_b32 v0, v[0:1]
+; GFX11-NEXT:    v_lshlrev_b32_e32 v1, 2, v2
+; GFX11-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX11-NEXT:    scratch_store_d16_hi_b8 v1, v0, s0 offset:4
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %load = load <4 x i8>, ptr %in
+  %element = extractelement <4 x i8> %load, i32 2
+  %voffset4 = mul i32 %voffset, 4
+  %gep0 = getelementptr inbounds i8, ptr addrspace(5) %out, i32 %voffset4
+  %gep = getelementptr <4 x i8>, ptr addrspace(5) %gep0, i64 1
+  store i8 %element, ptr addrspace(5) %gep, align 4
+  ret void
+}
+
+define amdgpu_ps void @test_scratch_store_b16_from_d16_hi_svs(ptr %in, ptr addrspace(5) inreg %out, i32 %voffset) {
+; GFX10-LABEL: test_scratch_store_b16_from_d16_hi_svs:
+; GFX10:       ; %bb.0: ; %bb
+; GFX10-NEXT:    s_add_u32 s0, s0, s3
+; GFX10-NEXT:    s_addc_u32 s1, s1, 0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s0
+; GFX10-NEXT:    s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s1
+; GFX10-NEXT:    flat_load_dword v0, v[0:1]
+; GFX10-NEXT:    v_lshl_add_u32 v1, v2, 2, s2
+; GFX10-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX10-NEXT:    scratch_store_short_d16_hi v1, v0, off offset:2
+; GFX10-NEXT:    s_endpgm
+;
+; GFX11-LABEL: test_scratch_store_b16_from_d16_hi_svs:
+; GFX11:       ; %bb.0: ; %bb
+; GFX11-NEXT:    flat_load_b32 v0, v[0:1]
+; GFX11-NEXT:    v_lshlrev_b32_e32 v1, 2, v2
+; GFX11-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GFX11-NEXT:    scratch_store_d16_hi_b16 v1, v0, s0 offset:2
+; GFX11-NEXT:    s_sendmsg sendmsg(MSG_DEALLOC_VGPRS)
+; GFX11-NEXT:    s_endpgm
+bb:
+  %load = load <2 x i16>, ptr %in
+  %element = extractelement <2 x i16> %load, i32 1
+  %voffset4 = mul i32 %voffset, 4
+  %gep0 = getelementptr inbounds i8, ptr addrspace(5) %out, i32 %voffset4
+  %gep = getelementptr <2 x i8>, ptr addrspace(5) %gep0, i64 1
+  store i16 %element, ptr addrspace(5) %gep, align 4
+  ret void
+}