caio/clang-p2996
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
dda73336ad22bd0b5ecda17040c50fb10fcbe5fb
clang-p2996/llvm/test/CodeGen/AMDGPU/GlobalISel/flat-scratch.ll
Petar Avramovic 3e35ba53e2 AMDGPU/GFX12: Insert waitcnts before stores with scope_sys (#82996) 
Insert waitcnts for loads and atomics before stores with system scope.
Scope is field in instruction encoding and corresponds to desired
coherence level in cache hierarchy.
Intrinsic stores can set scope in cache policy operand.
If volatile keyword is used on generic stores memory legalizer will set
scope to system. Generic stores, by default, get lowest scope level.
Waitcnts are not required if it is guaranteed that memory is cached.
For example vulkan shaders can guarantee this.
TODO: implement flag for frontends to give us a hint not to insert
waits.
Expecting vulkan flag to be implemented as vulkan:private MMRA.
2024-02-28 16:18:04 +01:00

1501 lines
61 KiB
LLVM
Raw Blame History

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=amdgcn -mcpu=gfx900 -global-isel -mattr=-promote-alloca -mattr=+enable-flat-scratch -verify-machineinstrs < %s | FileCheck -check-prefix=GFX9 %s
; RUN: llc -mtriple=amdgcn -mcpu=gfx1030 -global-isel -mattr=-promote-alloca -mattr=+enable-flat-scratch -verify-machineinstrs < %s | FileCheck -check-prefix=GFX10 %s
; RUN: llc -mtriple=amdgcn -mcpu=gfx940 -global-isel -mattr=-promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=GFX940 %s
; RUN: llc -mtriple=amdgcn -mcpu=gfx1100 -global-isel -mattr=-promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=GFX11 %s
; RUN: llc -mtriple=amdgcn -mcpu=gfx1200 -global-isel -mattr=-promote-alloca -verify-machineinstrs < %s | FileCheck -check-prefix=GFX12 %s
define amdgpu_kernel void @store_load_sindex_kernel(i32 %idx) {
; GFX9-LABEL: store_load_sindex_kernel:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_load_dword s0, s[0:1], 0x24
; GFX9-NEXT: s_add_u32 flat_scratch_lo, s2, s5
; GFX9-NEXT: s_addc_u32 flat_scratch_hi, s3, 0
; GFX9-NEXT: v_mov_b32_e32 v0, 15
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: s_lshl_b32 s1, s0, 2
; GFX9-NEXT: s_and_b32 s0, s0, 15
; GFX9-NEXT: s_add_i32 s1, s1, 0
; GFX9-NEXT: s_lshl_b32 s0, s0, 2
; GFX9-NEXT: scratch_store_dword off, v0, s1
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_add_i32 s0, s0, 0
; GFX9-NEXT: scratch_load_dword v0, off, s0 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_endpgm
;
; GFX10-LABEL: store_load_sindex_kernel:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_add_u32 s2, s2, s5
; GFX10-NEXT: s_addc_u32 s3, s3, 0
; GFX10-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s2
; GFX10-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s3
; GFX10-NEXT: s_load_dword s0, s[0:1], 0x24
; GFX10-NEXT: v_mov_b32_e32 v0, 15
; GFX10-NEXT: s_waitcnt lgkmcnt(0)
; GFX10-NEXT: s_and_b32 s1, s0, 15
; GFX10-NEXT: s_lshl_b32 s0, s0, 2
; GFX10-NEXT: s_lshl_b32 s1, s1, 2
; GFX10-NEXT: s_add_i32 s0, s0, 0
; GFX10-NEXT: s_add_i32 s1, s1, 0
; GFX10-NEXT: scratch_store_dword off, v0, s0
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dword v0, off, s1 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_endpgm
;
; GFX940-LABEL: store_load_sindex_kernel:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: s_load_dword s0, s[0:1], 0x24
; GFX940-NEXT: v_mov_b32_e32 v0, 15
; GFX940-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NEXT: s_lshl_b32 s1, s0, 2
; GFX940-NEXT: s_and_b32 s0, s0, 15
; GFX940-NEXT: s_add_i32 s1, s1, 0
; GFX940-NEXT: s_lshl_b32 s0, s0, 2
; GFX940-NEXT: scratch_store_dword off, v0, s1 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_mov_b32_e32 v0, s0
; GFX940-NEXT: scratch_load_dword v0, v0, off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_endpgm
;
; GFX11-LABEL: store_load_sindex_kernel:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_load_b32 s0, s[0:1], 0x24
; GFX11-NEXT: s_waitcnt lgkmcnt(0)
; GFX11-NEXT: s_and_b32 s1, s0, 15
; GFX11-NEXT: s_lshl_b32 s0, s0, 2
; GFX11-NEXT: s_lshl_b32 s1, s1, 2
; GFX11-NEXT: s_delay_alu instid0(SALU_CYCLE_1)
; GFX11-NEXT: v_dual_mov_b32 v0, 15 :: v_dual_mov_b32 v1, s1
; GFX11-NEXT: s_add_i32 s0, s0, 0
; GFX11-NEXT: scratch_store_b32 off, v0, s0 dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_load_b32 v0, v1, off glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: store_load_sindex_kernel:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: s_load_b32 s0, s[0:1], 0x24
; GFX12-NEXT: v_mov_b32_e32 v1, 15
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: s_lshl_b32 s1, s0, 2
; GFX12-NEXT: s_and_b32 s0, s0, 15
; GFX12-NEXT: v_mov_b32_e32 v0, s1
; GFX12-NEXT: s_lshl_b32 s0, s0, 2
; GFX12-NEXT: s_delay_alu instid0(SALU_CYCLE_1)
; GFX12-NEXT: v_mov_b32_e32 v2, s0
; GFX12-NEXT: scratch_store_b32 v0, v1, off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b32 v0, v2, off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_endpgm
bb:
%i = alloca [32 x float], align 4, addrspace(5)
%i7 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %idx
store volatile i32 15, ptr addrspace(5) %i7, align 4
%i9 = and i32 %idx, 15
%i10 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %i9
%i12 = load volatile i32, ptr addrspace(5) %i10, align 4
ret void
}
define amdgpu_kernel void @store_load_vindex_kernel() {
; GFX9-LABEL: store_load_vindex_kernel:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_add_u32 flat_scratch_lo, s0, s3
; GFX9-NEXT: v_lshlrev_b32_e32 v1, 2, v0
; GFX9-NEXT: v_sub_u32_e32 v0, 0, v0
; GFX9-NEXT: s_addc_u32 flat_scratch_hi, s1, 0
; GFX9-NEXT: v_add_u32_e32 v1, 0, v1
; GFX9-NEXT: v_mov_b32_e32 v2, 15
; GFX9-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT: scratch_store_dword v1, v2, off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_add_u32_e32 v0, 0, v0
; GFX9-NEXT: scratch_load_dword v0, v0, off offset:124 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_endpgm
;
; GFX10-LABEL: store_load_vindex_kernel:
; 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_sub_nc_u32_e32 v1, 0, v0
; GFX10-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX10-NEXT: v_mov_b32_e32 v2, 15
; GFX10-NEXT: v_lshlrev_b32_e32 v1, 2, v1
; GFX10-NEXT: v_add_nc_u32_e32 v0, 0, v0
; GFX10-NEXT: v_add_nc_u32_e32 v1, 0, v1
; GFX10-NEXT: scratch_store_dword v0, v2, off
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dword v0, v1, off offset:124 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_endpgm
;
; GFX940-LABEL: store_load_vindex_kernel:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: v_lshlrev_b32_e32 v1, 2, v0
; GFX940-NEXT: v_sub_u32_e32 v0, 0, v0
; GFX940-NEXT: v_mov_b32_e32 v2, 15
; GFX940-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-NEXT: scratch_store_dword v1, v2, off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_add_u32_e32 v0, 0, v0
; GFX940-NEXT: scratch_load_dword v0, v0, off offset:124 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_endpgm
;
; GFX11-LABEL: store_load_vindex_kernel:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: v_sub_nc_u32_e32 v1, 0, v0
; GFX11-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX11-NEXT: s_delay_alu instid0(VALU_DEP_2)
; GFX11-NEXT: v_dual_mov_b32 v2, 15 :: v_dual_lshlrev_b32 v1, 2, v1
; GFX11-NEXT: scratch_store_b32 v0, v2, off dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: v_add_nc_u32_e32 v1, 0, v1
; GFX11-NEXT: scratch_load_b32 v0, v1, off offset:124 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: store_load_vindex_kernel:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: v_sub_nc_u32_e32 v1, 0, v0
; GFX12-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX12-NEXT: s_delay_alu instid0(VALU_DEP_2)
; GFX12-NEXT: v_dual_mov_b32 v2, 15 :: v_dual_lshlrev_b32 v1, 2, v1
; GFX12-NEXT: scratch_store_b32 v0, v2, off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b32 v0, v1, off offset:124 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_endpgm
bb:
%i = alloca [32 x float], align 4, addrspace(5)
%i2 = tail call i32 @llvm.amdgcn.workitem.id.x()
%i3 = zext i32 %i2 to i64
%i7 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %i2
store volatile i32 15, ptr addrspace(5) %i7, align 4
%i9 = sub nsw i32 31, %i2
%i10 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %i9
%i12 = load volatile i32, ptr addrspace(5) %i10, align 4
ret void
}
define void @store_load_vindex_foo(i32 %idx) {
; GFX9-LABEL: store_load_vindex_foo:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_lshlrev_b32_e32 v1, 2, v0
; GFX9-NEXT: v_and_b32_e32 v0, 15, v0
; GFX9-NEXT: v_add_u32_e32 v1, s32, v1
; GFX9-NEXT: v_mov_b32_e32 v2, 15
; GFX9-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT: scratch_store_dword v1, v2, off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_add_u32_e32 v0, s32, v0
; GFX9-NEXT: scratch_load_dword v0, v0, off glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: store_load_vindex_foo:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_and_b32_e32 v1, 15, v0
; GFX10-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX10-NEXT: v_mov_b32_e32 v2, 15
; GFX10-NEXT: v_lshlrev_b32_e32 v1, 2, v1
; GFX10-NEXT: v_add_nc_u32_e32 v0, s32, v0
; GFX10-NEXT: v_add_nc_u32_e32 v1, s32, v1
; GFX10-NEXT: scratch_store_dword v0, v2, off
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dword v0, v1, off glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_setpc_b64 s[30:31]
;
; GFX940-LABEL: store_load_vindex_foo:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX940-NEXT: v_lshlrev_b32_e32 v1, 2, v0
; GFX940-NEXT: v_add_u32_e32 v1, s32, v1
; GFX940-NEXT: v_mov_b32_e32 v2, 15
; GFX940-NEXT: v_and_b32_e32 v0, 15, v0
; GFX940-NEXT: scratch_store_dword v1, v2, off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-NEXT: scratch_load_dword v0, v0, s32 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_setpc_b64 s[30:31]
;
; GFX11-LABEL: store_load_vindex_foo:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-NEXT: v_dual_mov_b32 v2, 15 :: v_dual_lshlrev_b32 v1, 2, v0
; GFX11-NEXT: v_and_b32_e32 v0, 15, v0
; GFX11-NEXT: s_delay_alu instid0(VALU_DEP_2) | instskip(NEXT) | instid1(VALU_DEP_2)
; GFX11-NEXT: v_add_nc_u32_e32 v1, s32, v1
; GFX11-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX11-NEXT: scratch_store_b32 v1, v2, off dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_load_b32 v0, v0, s32 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_setpc_b64 s[30:31]
;
; GFX12-LABEL: store_load_vindex_foo:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_dual_mov_b32 v2, 15 :: v_dual_and_b32 v1, 15, v0
; GFX12-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX12-NEXT: s_delay_alu instid0(VALU_DEP_2)
; GFX12-NEXT: v_lshlrev_b32_e32 v1, 2, v1
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_store_b32 v0, v2, s32 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b32 v0, v1, s32 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_setpc_b64 s[30:31]
bb:
%i = alloca [32 x float], align 4, addrspace(5)
%i7 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %idx
store volatile i32 15, ptr addrspace(5) %i7, align 4
%i9 = and i32 %idx, 15
%i10 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %i9
%i12 = load volatile i32, ptr addrspace(5) %i10, align 4
ret void
}
define void @private_ptr_foo(ptr addrspace(5) nocapture %arg) {
; GFX9-LABEL: private_ptr_foo:
; GFX9: ; %bb.0:
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v1, 0x41200000
; GFX9-NEXT: scratch_store_dword v0, v1, off offset:4
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: private_ptr_foo:
; GFX10: ; %bb.0:
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_mov_b32_e32 v1, 0x41200000
; GFX10-NEXT: scratch_store_dword v0, v1, off offset:4
; GFX10-NEXT: s_setpc_b64 s[30:31]
;
; GFX940-LABEL: private_ptr_foo:
; GFX940: ; %bb.0:
; GFX940-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX940-NEXT: v_mov_b32_e32 v1, 0x41200000
; GFX940-NEXT: scratch_store_dword v0, v1, off offset:4 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_setpc_b64 s[30:31]
;
; GFX11-LABEL: private_ptr_foo:
; GFX11: ; %bb.0:
; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-NEXT: v_mov_b32_e32 v1, 0x41200000
; GFX11-NEXT: scratch_store_b32 v0, v1, off offset:4
; GFX11-NEXT: s_setpc_b64 s[30:31]
;
; GFX12-LABEL: private_ptr_foo:
; GFX12: ; %bb.0:
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_mov_b32_e32 v1, 0x41200000
; GFX12-NEXT: scratch_store_b32 v0, v1, off offset:4
; GFX12-NEXT: s_setpc_b64 s[30:31]
%gep = getelementptr inbounds float, ptr addrspace(5) %arg, i32 1
store float 1.000000e+01, ptr addrspace(5) %gep, align 4
ret void
}
define amdgpu_kernel void @store_load_sindex_small_offset_kernel(i32 %idx) {
; GFX9-LABEL: store_load_sindex_small_offset_kernel:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_load_dword s0, s[0:1], 0x24
; GFX9-NEXT: s_add_u32 flat_scratch_lo, s2, s5
; GFX9-NEXT: s_addc_u32 flat_scratch_hi, s3, 0
; GFX9-NEXT: s_mov_b32 s1, 0
; GFX9-NEXT: scratch_load_dword v0, off, s1 glc
; GFX9-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT: s_lshl_b32 s1, s0, 2
; GFX9-NEXT: s_and_b32 s0, s0, 15
; GFX9-NEXT: v_mov_b32_e32 v0, 15
; GFX9-NEXT: s_addk_i32 s1, 0x100
; GFX9-NEXT: s_lshl_b32 s0, s0, 2
; GFX9-NEXT: scratch_store_dword off, v0, s1
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_addk_i32 s0, 0x100
; GFX9-NEXT: scratch_load_dword v0, off, s0 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_endpgm
;
; GFX10-LABEL: store_load_sindex_small_offset_kernel:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_add_u32 s2, s2, s5
; GFX10-NEXT: s_addc_u32 s3, s3, 0
; GFX10-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s2
; GFX10-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s3
; GFX10-NEXT: s_load_dword s0, s[0:1], 0x24
; GFX10-NEXT: scratch_load_dword v0, off, off glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: v_mov_b32_e32 v0, 15
; GFX10-NEXT: s_waitcnt lgkmcnt(0)
; GFX10-NEXT: s_and_b32 s1, s0, 15
; GFX10-NEXT: s_lshl_b32 s0, s0, 2
; GFX10-NEXT: s_lshl_b32 s1, s1, 2
; GFX10-NEXT: s_addk_i32 s0, 0x100
; GFX10-NEXT: s_addk_i32 s1, 0x100
; GFX10-NEXT: scratch_store_dword off, v0, s0
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dword v0, off, s1 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_endpgm
;
; GFX940-LABEL: store_load_sindex_small_offset_kernel:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: s_load_dword s0, s[0:1], 0x24
; GFX940-NEXT: scratch_load_dword v0, off, off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_mov_b32_e32 v0, 15
; GFX940-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NEXT: s_lshl_b32 s1, s0, 2
; GFX940-NEXT: s_and_b32 s0, s0, 15
; GFX940-NEXT: s_addk_i32 s1, 0x100
; GFX940-NEXT: s_lshl_b32 s0, s0, 2
; GFX940-NEXT: scratch_store_dword off, v0, s1 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_mov_b32_e32 v0, s0
; GFX940-NEXT: scratch_load_dword v0, v0, off offset:256 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_endpgm
;
; GFX11-LABEL: store_load_sindex_small_offset_kernel:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_load_b32 s0, s[0:1], 0x24
; GFX11-NEXT: scratch_load_b32 v2, off, off glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX11-NEXT: s_and_b32 s1, s0, 15
; GFX11-NEXT: s_lshl_b32 s0, s0, 2
; GFX11-NEXT: s_lshl_b32 s1, s1, 2
; GFX11-NEXT: s_delay_alu instid0(SALU_CYCLE_1)
; GFX11-NEXT: v_dual_mov_b32 v0, 15 :: v_dual_mov_b32 v1, s1
; GFX11-NEXT: s_addk_i32 s0, 0x100
; GFX11-NEXT: scratch_store_b32 off, v0, s0 dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_load_b32 v0, v1, off offset:256 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: store_load_sindex_small_offset_kernel:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: s_load_b32 s0, s[0:1], 0x24
; GFX12-NEXT: scratch_load_b32 v3, off, off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: v_mov_b32_e32 v1, 15
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: s_lshl_b32 s1, s0, 2
; GFX12-NEXT: s_and_b32 s0, s0, 15
; GFX12-NEXT: v_mov_b32_e32 v0, s1
; GFX12-NEXT: s_lshl_b32 s0, s0, 2
; GFX12-NEXT: s_delay_alu instid0(SALU_CYCLE_1)
; GFX12-NEXT: v_mov_b32_e32 v2, s0
; GFX12-NEXT: scratch_store_b32 v0, v1, off offset:256 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b32 v0, v2, off offset:256 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_endpgm
bb:
%padding = alloca [64 x i32], align 4, addrspace(5)
%i = alloca [32 x float], align 4, addrspace(5)
%pad_gep = getelementptr inbounds [64 x i32], ptr addrspace(5) %padding, i32 0, i32 undef
%pad_load = load volatile i32, ptr addrspace(5) %pad_gep, align 4
%i7 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %idx
store volatile i32 15, ptr addrspace(5) %i7, align 4
%i9 = and i32 %idx, 15
%i10 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %i9
%i12 = load volatile i32, ptr addrspace(5) %i10, align 4
ret void
}
define amdgpu_kernel void @store_load_vindex_small_offset_kernel() {
; GFX9-LABEL: store_load_vindex_small_offset_kernel:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_add_u32 flat_scratch_lo, s0, s3
; GFX9-NEXT: s_addc_u32 flat_scratch_hi, s1, 0
; GFX9-NEXT: s_mov_b32 s0, 0
; GFX9-NEXT: scratch_load_dword v1, off, s0 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_lshlrev_b32_e32 v1, 2, v0
; GFX9-NEXT: v_sub_u32_e32 v0, 0, v0
; GFX9-NEXT: v_add_u32_e32 v1, 0x100, v1
; GFX9-NEXT: v_mov_b32_e32 v2, 15
; GFX9-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT: scratch_store_dword v1, v2, off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_add_u32_e32 v0, 0x100, v0
; GFX9-NEXT: scratch_load_dword v0, v0, off offset:124 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_endpgm
;
; GFX10-LABEL: store_load_vindex_small_offset_kernel:
; 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_sub_nc_u32_e32 v1, 0, v0
; GFX10-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX10-NEXT: v_mov_b32_e32 v2, 15
; GFX10-NEXT: scratch_load_dword v3, off, off glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: v_lshlrev_b32_e32 v1, 2, v1
; GFX10-NEXT: v_add_nc_u32_e32 v0, 0x100, v0
; GFX10-NEXT: v_add_nc_u32_e32 v1, 0x100, v1
; GFX10-NEXT: scratch_store_dword v0, v2, off
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dword v0, v1, off offset:124 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_endpgm
;
; GFX940-LABEL: store_load_vindex_small_offset_kernel:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: scratch_load_dword v1, off, off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_lshlrev_b32_e32 v1, 2, v0
; GFX940-NEXT: v_sub_u32_e32 v0, 0, v0
; GFX940-NEXT: v_mov_b32_e32 v2, 15
; GFX940-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-NEXT: scratch_store_dword v1, v2, off offset:256 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_add_u32_e32 v0, 0x100, v0
; GFX940-NEXT: scratch_load_dword v0, v0, off offset:124 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_endpgm
;
; GFX11-LABEL: store_load_vindex_small_offset_kernel:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: v_sub_nc_u32_e32 v1, 0, v0
; GFX11-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX11-NEXT: scratch_load_b32 v3, off, off glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: v_dual_mov_b32 v2, 15 :: v_dual_lshlrev_b32 v1, 2, v1
; GFX11-NEXT: scratch_store_b32 v0, v2, off offset:256 dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: v_add_nc_u32_e32 v1, 0x100, v1
; GFX11-NEXT: scratch_load_b32 v0, v1, off offset:124 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: store_load_vindex_small_offset_kernel:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: v_sub_nc_u32_e32 v1, 0, v0
; GFX12-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX12-NEXT: v_mov_b32_e32 v2, 15
; GFX12-NEXT: scratch_load_b32 v3, off, off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: v_lshlrev_b32_e32 v1, 2, v1
; GFX12-NEXT: scratch_store_b32 v0, v2, off offset:256 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b32 v0, v1, off offset:380 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_endpgm
bb:
%padding = alloca [64 x i32], align 4, addrspace(5)
%i = alloca [32 x float], align 4, addrspace(5)
%pad_gep = getelementptr inbounds [64 x i32], ptr addrspace(5) %padding, i32 0, i32 undef
%pad_load = load volatile i32, ptr addrspace(5) %pad_gep, align 4
%i2 = tail call i32 @llvm.amdgcn.workitem.id.x()
%i3 = zext i32 %i2 to i64
%i7 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %i2
store volatile i32 15, ptr addrspace(5) %i7, align 4
%i9 = sub nsw i32 31, %i2
%i10 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %i9
%i12 = load volatile i32, ptr addrspace(5) %i10, align 4
ret void
}
define void @store_load_vindex_small_offset_foo(i32 %idx) {
; GFX9-LABEL: store_load_vindex_small_offset_foo:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: scratch_load_dword v1, off, s32 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_lshlrev_b32_e32 v1, 2, v0
; GFX9-NEXT: s_add_i32 s0, s32, 0x100
; GFX9-NEXT: v_and_b32_e32 v0, 15, v0
; GFX9-NEXT: v_add_u32_e32 v1, s0, v1
; GFX9-NEXT: v_mov_b32_e32 v2, 15
; GFX9-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT: s_add_i32 s0, s32, 0x100
; GFX9-NEXT: scratch_store_dword v1, v2, off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_add_u32_e32 v0, s0, v0
; GFX9-NEXT: scratch_load_dword v0, v0, off glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: store_load_vindex_small_offset_foo:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_and_b32_e32 v1, 15, v0
; GFX10-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX10-NEXT: s_add_i32 s0, s32, 0x100
; GFX10-NEXT: v_mov_b32_e32 v2, 15
; GFX10-NEXT: scratch_load_dword v3, off, s32 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: v_lshlrev_b32_e32 v1, 2, v1
; GFX10-NEXT: v_add_nc_u32_e32 v0, s0, v0
; GFX10-NEXT: s_add_i32 s0, s32, 0x100
; GFX10-NEXT: v_add_nc_u32_e32 v1, s0, v1
; GFX10-NEXT: scratch_store_dword v0, v2, off
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dword v0, v1, off glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_setpc_b64 s[30:31]
;
; GFX940-LABEL: store_load_vindex_small_offset_foo:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX940-NEXT: scratch_load_dword v1, off, s32 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_lshlrev_b32_e32 v1, 2, v0
; GFX940-NEXT: s_add_i32 s0, s32, 0x100
; GFX940-NEXT: v_add_u32_e32 v1, s0, v1
; GFX940-NEXT: v_mov_b32_e32 v2, 15
; GFX940-NEXT: v_and_b32_e32 v0, 15, v0
; GFX940-NEXT: scratch_store_dword v1, v2, off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-NEXT: scratch_load_dword v0, v0, s32 offset:256 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_setpc_b64 s[30:31]
;
; GFX11-LABEL: store_load_vindex_small_offset_foo:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-NEXT: v_dual_mov_b32 v2, 15 :: v_dual_lshlrev_b32 v1, 2, v0
; GFX11-NEXT: v_and_b32_e32 v0, 15, v0
; GFX11-NEXT: s_add_i32 s0, s32, 0x100
; GFX11-NEXT: scratch_load_b32 v3, off, s32 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: v_add_nc_u32_e32 v1, s0, v1
; GFX11-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX11-NEXT: scratch_store_b32 v1, v2, off dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_load_b32 v0, v0, s32 offset:256 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_setpc_b64 s[30:31]
;
; GFX12-LABEL: store_load_vindex_small_offset_foo:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_dual_mov_b32 v2, 15 :: v_dual_and_b32 v1, 15, v0
; GFX12-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX12-NEXT: scratch_load_b32 v3, off, s32 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: v_lshlrev_b32_e32 v1, 2, v1
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_store_b32 v0, v2, s32 offset:256 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b32 v0, v1, s32 offset:256 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_setpc_b64 s[30:31]
bb:
%padding = alloca [64 x i32], align 4, addrspace(5)
%i = alloca [32 x float], align 4, addrspace(5)
%pad_gep = getelementptr inbounds [64 x i32], ptr addrspace(5) %padding, i32 0, i32 undef
%pad_load = load volatile i32, ptr addrspace(5) %pad_gep, align 4
%i7 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %idx
store volatile i32 15, ptr addrspace(5) %i7, align 4
%i9 = and i32 %idx, 15
%i10 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %i9
%i12 = load volatile i32, ptr addrspace(5) %i10, align 4
ret void
}
define amdgpu_kernel void @store_load_sindex_large_offset_kernel(i32 %idx) {
; GFX9-LABEL: store_load_sindex_large_offset_kernel:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_load_dword s0, s[0:1], 0x24
; GFX9-NEXT: s_add_u32 flat_scratch_lo, s2, s5
; GFX9-NEXT: s_addc_u32 flat_scratch_hi, s3, 0
; GFX9-NEXT: s_mov_b32 s1, 0
; GFX9-NEXT: scratch_load_dword v0, off, s1 offset:4 glc
; GFX9-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX9-NEXT: s_lshl_b32 s1, s0, 2
; GFX9-NEXT: s_and_b32 s0, s0, 15
; GFX9-NEXT: v_mov_b32_e32 v0, 15
; GFX9-NEXT: s_addk_i32 s1, 0x4004
; GFX9-NEXT: s_lshl_b32 s0, s0, 2
; GFX9-NEXT: scratch_store_dword off, v0, s1
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_addk_i32 s0, 0x4004
; GFX9-NEXT: scratch_load_dword v0, off, s0 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_endpgm
;
; GFX10-LABEL: store_load_sindex_large_offset_kernel:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_add_u32 s2, s2, s5
; GFX10-NEXT: s_addc_u32 s3, s3, 0
; GFX10-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s2
; GFX10-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s3
; GFX10-NEXT: s_load_dword s0, s[0:1], 0x24
; GFX10-NEXT: scratch_load_dword v0, off, off offset:4 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: v_mov_b32_e32 v0, 15
; GFX10-NEXT: s_waitcnt lgkmcnt(0)
; GFX10-NEXT: s_and_b32 s1, s0, 15
; GFX10-NEXT: s_lshl_b32 s0, s0, 2
; GFX10-NEXT: s_lshl_b32 s1, s1, 2
; GFX10-NEXT: s_addk_i32 s0, 0x4004
; GFX10-NEXT: s_addk_i32 s1, 0x4004
; GFX10-NEXT: scratch_store_dword off, v0, s0
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dword v0, off, s1 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_endpgm
;
; GFX940-LABEL: store_load_sindex_large_offset_kernel:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: s_load_dword s0, s[0:1], 0x24
; GFX940-NEXT: scratch_load_dword v0, off, off offset:4 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_mov_b32_e32 v0, 15
; GFX940-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NEXT: s_lshl_b32 s1, s0, 2
; GFX940-NEXT: s_and_b32 s0, s0, 15
; GFX940-NEXT: s_addk_i32 s1, 0x4004
; GFX940-NEXT: s_lshl_b32 s0, s0, 2
; GFX940-NEXT: scratch_store_dword off, v0, s1 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_mov_b32_e32 v0, s0
; GFX940-NEXT: s_movk_i32 s0, 0x4004
; GFX940-NEXT: scratch_load_dword v0, v0, s0 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_endpgm
;
; GFX11-LABEL: store_load_sindex_large_offset_kernel:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_load_b32 s0, s[0:1], 0x24
; GFX11-NEXT: scratch_load_b32 v2, off, off offset:4 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0) lgkmcnt(0)
; GFX11-NEXT: s_and_b32 s1, s0, 15
; GFX11-NEXT: s_lshl_b32 s0, s0, 2
; GFX11-NEXT: s_lshl_b32 s1, s1, 2
; GFX11-NEXT: s_delay_alu instid0(SALU_CYCLE_1)
; GFX11-NEXT: v_dual_mov_b32 v0, 15 :: v_dual_mov_b32 v1, s1
; GFX11-NEXT: s_addk_i32 s0, 0x4004
; GFX11-NEXT: scratch_store_b32 off, v0, s0 dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: s_movk_i32 s0, 0x4004
; GFX11-NEXT: scratch_load_b32 v0, v1, s0 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: store_load_sindex_large_offset_kernel:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: s_load_b32 s0, s[0:1], 0x24
; GFX12-NEXT: scratch_load_b32 v3, off, off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: v_mov_b32_e32 v1, 15
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: s_lshl_b32 s1, s0, 2
; GFX12-NEXT: s_and_b32 s0, s0, 15
; GFX12-NEXT: v_mov_b32_e32 v0, s1
; GFX12-NEXT: s_lshl_b32 s0, s0, 2
; GFX12-NEXT: s_delay_alu instid0(SALU_CYCLE_1)
; GFX12-NEXT: v_mov_b32_e32 v2, s0
; GFX12-NEXT: scratch_store_b32 v0, v1, off offset:16384 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b32 v0, v2, off offset:16384 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_endpgm
bb:
%padding = alloca [4096 x i32], align 4, addrspace(5)
%i = alloca [32 x float], align 4, addrspace(5)
%pad_gep = getelementptr inbounds [4096 x i32], ptr addrspace(5) %padding, i32 0, i32 undef
%pad_load = load volatile i32, ptr addrspace(5) %pad_gep, align 4
%i7 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %idx
store volatile i32 15, ptr addrspace(5) %i7, align 4
%i9 = and i32 %idx, 15
%i10 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %i9
%i12 = load volatile i32, ptr addrspace(5) %i10, align 4
ret void
}
define amdgpu_kernel void @store_load_vindex_large_offset_kernel() {
; GFX9-LABEL: store_load_vindex_large_offset_kernel:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_add_u32 flat_scratch_lo, s0, s3
; GFX9-NEXT: s_addc_u32 flat_scratch_hi, s1, 0
; GFX9-NEXT: s_mov_b32 s0, 0
; GFX9-NEXT: scratch_load_dword v1, off, s0 offset:4 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_lshlrev_b32_e32 v1, 2, v0
; GFX9-NEXT: v_sub_u32_e32 v0, 0, v0
; GFX9-NEXT: v_add_u32_e32 v1, 0x4004, v1
; GFX9-NEXT: v_mov_b32_e32 v2, 15
; GFX9-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT: scratch_store_dword v1, v2, off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_add_u32_e32 v0, 0x4004, v0
; GFX9-NEXT: scratch_load_dword v0, v0, off offset:124 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_endpgm
;
; GFX10-LABEL: store_load_vindex_large_offset_kernel:
; 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_sub_nc_u32_e32 v1, 0, v0
; GFX10-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX10-NEXT: v_mov_b32_e32 v2, 15
; GFX10-NEXT: scratch_load_dword v3, off, off offset:4 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: v_lshlrev_b32_e32 v1, 2, v1
; GFX10-NEXT: v_add_nc_u32_e32 v0, 0x4004, v0
; GFX10-NEXT: v_add_nc_u32_e32 v1, 0x4004, v1
; GFX10-NEXT: scratch_store_dword v0, v2, off
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dword v0, v1, off offset:124 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_endpgm
;
; GFX940-LABEL: store_load_vindex_large_offset_kernel:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: scratch_load_dword v1, off, off offset:4 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_lshlrev_b32_e32 v1, 2, v0
; GFX940-NEXT: v_sub_u32_e32 v0, 0, v0
; GFX940-NEXT: v_mov_b32_e32 v2, 15
; GFX940-NEXT: s_movk_i32 s0, 0x4004
; GFX940-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-NEXT: scratch_store_dword v1, v2, s0 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_add_u32_e32 v0, 0x4004, v0
; GFX940-NEXT: scratch_load_dword v0, v0, off offset:124 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_endpgm
;
; GFX11-LABEL: store_load_vindex_large_offset_kernel:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: v_sub_nc_u32_e32 v1, 0, v0
; GFX11-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX11-NEXT: s_movk_i32 s0, 0x4004
; GFX11-NEXT: scratch_load_b32 v3, off, off offset:4 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: v_dual_mov_b32 v2, 15 :: v_dual_lshlrev_b32 v1, 2, v1
; GFX11-NEXT: scratch_store_b32 v0, v2, s0 dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: v_add_nc_u32_e32 v1, 0x4004, v1
; GFX11-NEXT: scratch_load_b32 v0, v1, off offset:124 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: store_load_vindex_large_offset_kernel:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: v_sub_nc_u32_e32 v1, 0, v0
; GFX12-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX12-NEXT: v_mov_b32_e32 v2, 15
; GFX12-NEXT: scratch_load_b32 v3, off, off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: v_lshlrev_b32_e32 v1, 2, v1
; GFX12-NEXT: scratch_store_b32 v0, v2, off offset:16384 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b32 v0, v1, off offset:16508 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_endpgm
bb:
%padding = alloca [4096 x i32], align 4, addrspace(5)
%i = alloca [32 x float], align 4, addrspace(5)
%pad_gep = getelementptr inbounds [4096 x i32], ptr addrspace(5) %padding, i32 0, i32 undef
%pad_load = load volatile i32, ptr addrspace(5) %pad_gep, align 4
%i2 = tail call i32 @llvm.amdgcn.workitem.id.x()
%i3 = zext i32 %i2 to i64
%i7 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %i2
store volatile i32 15, ptr addrspace(5) %i7, align 4
%i9 = sub nsw i32 31, %i2
%i10 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %i9
%i12 = load volatile i32, ptr addrspace(5) %i10, align 4
ret void
}
define void @store_load_vindex_large_offset_foo(i32 %idx) {
; GFX9-LABEL: store_load_vindex_large_offset_foo:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: scratch_load_dword v1, off, s32 offset:4 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_lshlrev_b32_e32 v1, 2, v0
; GFX9-NEXT: s_add_i32 s0, s32, 0x4004
; GFX9-NEXT: v_and_b32_e32 v0, 15, v0
; GFX9-NEXT: v_add_u32_e32 v1, s0, v1
; GFX9-NEXT: v_mov_b32_e32 v2, 15
; GFX9-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX9-NEXT: s_add_i32 s0, s32, 0x4004
; GFX9-NEXT: scratch_store_dword v1, v2, off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_add_u32_e32 v0, s0, v0
; GFX9-NEXT: scratch_load_dword v0, v0, off glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: store_load_vindex_large_offset_foo:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_and_b32_e32 v1, 15, v0
; GFX10-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX10-NEXT: s_add_i32 s0, s32, 0x4004
; GFX10-NEXT: v_mov_b32_e32 v2, 15
; GFX10-NEXT: scratch_load_dword v3, off, s32 offset:4 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: v_lshlrev_b32_e32 v1, 2, v1
; GFX10-NEXT: v_add_nc_u32_e32 v0, s0, v0
; GFX10-NEXT: s_add_i32 s0, s32, 0x4004
; GFX10-NEXT: v_add_nc_u32_e32 v1, s0, v1
; GFX10-NEXT: scratch_store_dword v0, v2, off
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dword v0, v1, off glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_setpc_b64 s[30:31]
;
; GFX940-LABEL: store_load_vindex_large_offset_foo:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX940-NEXT: scratch_load_dword v1, off, s32 offset:4 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_lshlrev_b32_e32 v1, 2, v0
; GFX940-NEXT: s_add_i32 s0, s32, 0x4004
; GFX940-NEXT: v_add_u32_e32 v1, s0, v1
; GFX940-NEXT: v_mov_b32_e32 v2, 15
; GFX940-NEXT: v_and_b32_e32 v0, 15, v0
; GFX940-NEXT: scratch_store_dword v1, v2, off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX940-NEXT: s_add_i32 s0, s32, 0x4004
; GFX940-NEXT: scratch_load_dword v0, v0, s0 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_setpc_b64 s[30:31]
;
; GFX11-LABEL: store_load_vindex_large_offset_foo:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-NEXT: v_dual_mov_b32 v2, 15 :: v_dual_lshlrev_b32 v1, 2, v0
; GFX11-NEXT: v_and_b32_e32 v0, 15, v0
; GFX11-NEXT: s_add_i32 s0, s32, 0x4004
; GFX11-NEXT: scratch_load_b32 v3, off, s32 offset:4 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: v_add_nc_u32_e32 v1, s0, v1
; GFX11-NEXT: s_add_i32 s0, s32, 0x4004
; GFX11-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX11-NEXT: scratch_store_b32 v1, v2, off dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_load_b32 v0, v0, s0 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_setpc_b64 s[30:31]
;
; GFX12-LABEL: store_load_vindex_large_offset_foo:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_dual_mov_b32 v2, 15 :: v_dual_and_b32 v1, 15, v0
; GFX12-NEXT: v_lshlrev_b32_e32 v0, 2, v0
; GFX12-NEXT: scratch_load_b32 v3, off, s32 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: v_lshlrev_b32_e32 v1, 2, v1
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_store_b32 v0, v2, s32 offset:16384 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b32 v0, v1, s32 offset:16384 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_setpc_b64 s[30:31]
bb:
%padding = alloca [4096 x i32], align 4, addrspace(5)
%i = alloca [32 x float], align 4, addrspace(5)
%pad_gep = getelementptr inbounds [4096 x i32], ptr addrspace(5) %padding, i32 0, i32 undef
%pad_load = load volatile i32, ptr addrspace(5) %pad_gep, align 4
%i7 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %idx
store volatile i32 15, ptr addrspace(5) %i7, align 4
%i9 = and i32 %idx, 15
%i10 = getelementptr inbounds [32 x float], ptr addrspace(5) %i, i32 0, i32 %i9
%i12 = load volatile i32, ptr addrspace(5) %i10, align 4
ret void
}
define amdgpu_kernel void @store_load_large_imm_offset_kernel() {
; GFX9-LABEL: store_load_large_imm_offset_kernel:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_add_u32 flat_scratch_lo, s0, s3
; GFX9-NEXT: s_addc_u32 flat_scratch_hi, s1, 0
; GFX9-NEXT: v_mov_b32_e32 v0, 13
; GFX9-NEXT: s_mov_b32 s0, 0
; GFX9-NEXT: scratch_store_dword off, v0, s0 offset:4
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_movk_i32 s0, 0x3e80
; GFX9-NEXT: v_mov_b32_e32 v0, 15
; GFX9-NEXT: s_add_i32 s0, s0, 4
; GFX9-NEXT: scratch_store_dword off, v0, s0
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: scratch_load_dword v0, off, s0 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_endpgm
;
; GFX10-LABEL: store_load_large_imm_offset_kernel:
; 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 v0, 13
; GFX10-NEXT: v_mov_b32_e32 v1, 15
; GFX10-NEXT: s_movk_i32 s0, 0x3e80
; GFX10-NEXT: s_add_i32 s0, s0, 4
; GFX10-NEXT: scratch_store_dword off, v0, off offset:4
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_store_dword off, v1, s0
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dword v0, off, s0 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_endpgm
;
; GFX940-LABEL: store_load_large_imm_offset_kernel:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: v_mov_b32_e32 v0, 13
; GFX940-NEXT: scratch_store_dword off, v0, off offset:4 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_mov_b32_e32 v0, 0x3e80
; GFX940-NEXT: v_mov_b32_e32 v1, 15
; GFX940-NEXT: scratch_store_dword v0, v1, off offset:4 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: scratch_load_dword v0, v0, off offset:4 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_endpgm
;
; GFX11-LABEL: store_load_large_imm_offset_kernel:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: v_dual_mov_b32 v0, 13 :: v_dual_mov_b32 v1, 0x3e80
; GFX11-NEXT: v_mov_b32_e32 v2, 15
; GFX11-NEXT: scratch_store_b32 off, v0, off offset:4 dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_store_b32 v1, v2, off offset:4 dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_load_b32 v0, v1, off offset:4 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: store_load_large_imm_offset_kernel:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: v_dual_mov_b32 v0, 13 :: v_dual_mov_b32 v1, 15
; GFX12-NEXT: scratch_store_b32 off, v0, off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_store_b32 off, v1, off offset:16000 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b32 v0, off, off offset:16000 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_endpgm
bb:
%i = alloca [4096 x i32], align 4, addrspace(5)
%i1 = getelementptr inbounds [4096 x i32], ptr addrspace(5) %i, i32 0, i32 undef
store volatile i32 13, ptr addrspace(5) %i1, align 4
%i7 = getelementptr inbounds [4096 x i32], ptr addrspace(5) %i, i32 0, i32 4000
store volatile i32 15, ptr addrspace(5) %i7, align 4
%i10 = getelementptr inbounds [4096 x i32], ptr addrspace(5) %i, i32 0, i32 4000
%i12 = load volatile i32, ptr addrspace(5) %i10, align 4
ret void
}
define void @store_load_large_imm_offset_foo() {
; GFX9-LABEL: store_load_large_imm_offset_foo:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v0, 13
; GFX9-NEXT: s_movk_i32 s0, 0x3e80
; GFX9-NEXT: s_add_i32 s1, s32, 4
; GFX9-NEXT: scratch_store_dword off, v0, s32 offset:4
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v0, 15
; GFX9-NEXT: s_add_i32 s0, s0, s1
; GFX9-NEXT: scratch_store_dword off, v0, s0
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: scratch_load_dword v0, off, s0 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: store_load_large_imm_offset_foo:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_mov_b32_e32 v0, 13
; GFX10-NEXT: v_mov_b32_e32 v1, 15
; GFX10-NEXT: s_movk_i32 s0, 0x3e80
; GFX10-NEXT: s_add_i32 s1, s32, 4
; GFX10-NEXT: s_add_i32 s0, s0, s1
; GFX10-NEXT: scratch_store_dword off, v0, s32 offset:4
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_store_dword off, v1, s0
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dword v0, off, s0 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_setpc_b64 s[30:31]
;
; GFX940-LABEL: store_load_large_imm_offset_foo:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX940-NEXT: v_mov_b32_e32 v0, 13
; GFX940-NEXT: scratch_store_dword off, v0, s32 offset:4 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: v_mov_b32_e32 v0, 0x3e80
; GFX940-NEXT: v_mov_b32_e32 v1, 15
; GFX940-NEXT: scratch_store_dword v0, v1, s32 offset:4 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: scratch_load_dword v0, v0, s32 offset:4 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_setpc_b64 s[30:31]
;
; GFX11-LABEL: store_load_large_imm_offset_foo:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-NEXT: v_dual_mov_b32 v0, 13 :: v_dual_mov_b32 v1, 0x3e80
; GFX11-NEXT: v_mov_b32_e32 v2, 15
; GFX11-NEXT: scratch_store_b32 off, v0, s32 offset:4 dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_store_b32 v1, v2, s32 offset:4 dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_load_b32 v0, v1, s32 offset:4 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_setpc_b64 s[30:31]
;
; GFX12-LABEL: store_load_large_imm_offset_foo:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_dual_mov_b32 v0, 13 :: v_dual_mov_b32 v1, 15
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_store_b32 off, v0, s32 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_store_b32 off, v1, s32 offset:16000 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b32 v0, off, s32 offset:16000 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_setpc_b64 s[30:31]
bb:
%i = alloca [4096 x i32], align 4, addrspace(5)
%i1 = getelementptr inbounds [4096 x i32], ptr addrspace(5) %i, i32 0, i32 undef
store volatile i32 13, ptr addrspace(5) %i1, align 4
%i7 = getelementptr inbounds [4096 x i32], ptr addrspace(5) %i, i32 0, i32 4000
store volatile i32 15, ptr addrspace(5) %i7, align 4
%i10 = getelementptr inbounds [4096 x i32], ptr addrspace(5) %i, i32 0, i32 4000
%i12 = load volatile i32, ptr addrspace(5) %i10, align 4
ret void
}
define amdgpu_kernel void @store_load_vidx_sidx_offset(i32 %sidx) {
; GFX9-LABEL: store_load_vidx_sidx_offset:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_load_dword s0, s[0:1], 0x24
; GFX9-NEXT: s_add_u32 flat_scratch_lo, s2, s5
; GFX9-NEXT: s_addc_u32 flat_scratch_hi, s3, 0
; GFX9-NEXT: v_mov_b32_e32 v1, 15
; GFX9-NEXT: s_waitcnt lgkmcnt(0)
; GFX9-NEXT: v_add_lshl_u32 v0, s0, v0, 2
; GFX9-NEXT: v_add_u32_e32 v0, 0, v0
; GFX9-NEXT: scratch_store_dword v0, v1, off offset:1024
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: scratch_load_dword v0, v0, off offset:1024 glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_endpgm
;
; GFX10-LABEL: store_load_vidx_sidx_offset:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_add_u32 s2, s2, s5
; GFX10-NEXT: s_addc_u32 s3, s3, 0
; GFX10-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_LO), s2
; GFX10-NEXT: s_setreg_b32 hwreg(HW_REG_FLAT_SCR_HI), s3
; GFX10-NEXT: s_load_dword s0, s[0:1], 0x24
; GFX10-NEXT: v_mov_b32_e32 v1, 15
; GFX10-NEXT: s_waitcnt lgkmcnt(0)
; GFX10-NEXT: v_add_lshl_u32 v0, s0, v0, 2
; GFX10-NEXT: v_add_nc_u32_e32 v0, 0, v0
; GFX10-NEXT: scratch_store_dword v0, v1, off offset:1024
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dword v0, v0, off offset:1024 glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_endpgm
;
; GFX940-LABEL: store_load_vidx_sidx_offset:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: s_load_dword s0, s[0:1], 0x24
; GFX940-NEXT: v_mov_b32_e32 v1, 15
; GFX940-NEXT: s_waitcnt lgkmcnt(0)
; GFX940-NEXT: v_add_lshl_u32 v0, s0, v0, 2
; GFX940-NEXT: v_add_u32_e32 v0, 0, v0
; GFX940-NEXT: scratch_store_dword v0, v1, off offset:1024 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: scratch_load_dword v0, v0, off offset:1024 sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_endpgm
;
; GFX11-LABEL: store_load_vidx_sidx_offset:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_load_b32 s0, s[0:1], 0x24
; GFX11-NEXT: v_mov_b32_e32 v1, 15
; GFX11-NEXT: s_waitcnt lgkmcnt(0)
; GFX11-NEXT: v_add_lshl_u32 v0, s0, v0, 2
; GFX11-NEXT: s_delay_alu instid0(VALU_DEP_1)
; GFX11-NEXT: v_add_nc_u32_e32 v0, 0, v0
; GFX11-NEXT: scratch_store_b32 v0, v1, off offset:1024 dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_load_b32 v0, v0, off offset:1024 glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_endpgm
;
; GFX12-LABEL: store_load_vidx_sidx_offset:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: s_load_b32 s0, s[0:1], 0x24
; GFX12-NEXT: v_mov_b32_e32 v1, 15
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_add_lshl_u32 v0, s0, v0, 2
; GFX12-NEXT: scratch_store_b32 v0, v1, off offset:1024 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b32 v0, v0, off offset:1024 scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_endpgm
bb:
%alloca = alloca [32 x i32], align 4, addrspace(5)
%vidx = tail call i32 @llvm.amdgcn.workitem.id.x()
%add1 = add nsw i32 %sidx, %vidx
%add2 = add nsw i32 %add1, 256
%gep = getelementptr inbounds [32 x i32], ptr addrspace(5) %alloca, i32 0, i32 %add2
store volatile i32 15, ptr addrspace(5) %gep, align 4
%load = load volatile i32, ptr addrspace(5) %gep, align 4
ret void
}
define void @store_load_i64_aligned(ptr addrspace(5) nocapture %arg) {
; GFX9-LABEL: store_load_i64_aligned:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v1, 15
; GFX9-NEXT: v_mov_b32_e32 v2, 0
; GFX9-NEXT: scratch_store_dwordx2 v0, v[1:2], off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: scratch_load_dwordx2 v[0:1], v0, off glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: store_load_i64_aligned:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_mov_b32_e32 v1, 15
; GFX10-NEXT: v_mov_b32_e32 v2, 0
; GFX10-NEXT: scratch_store_dwordx2 v0, v[1:2], off
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dwordx2 v[0:1], v0, off glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_setpc_b64 s[30:31]
;
; GFX940-LABEL: store_load_i64_aligned:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX940-NEXT: v_mov_b64_e32 v[2:3], 15
; GFX940-NEXT: scratch_store_dwordx2 v0, v[2:3], off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: scratch_load_dwordx2 v[0:1], v0, off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_setpc_b64 s[30:31]
;
; GFX11-LABEL: store_load_i64_aligned:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-NEXT: v_mov_b32_e32 v1, 15
; GFX11-NEXT: v_mov_b32_e32 v2, 0
; GFX11-NEXT: scratch_store_b64 v0, v[1:2], off dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_load_b64 v[0:1], v0, off glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_setpc_b64 s[30:31]
;
; GFX12-LABEL: store_load_i64_aligned:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_mov_b32_e32 v1, 15
; GFX12-NEXT: v_mov_b32_e32 v2, 0
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_store_b64 v0, v[1:2], off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b64 v[0:1], v0, off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_setpc_b64 s[30:31]
bb:
store volatile i64 15, ptr addrspace(5) %arg, align 8
%load = load volatile i64, ptr addrspace(5) %arg, align 8
ret void
}
define void @store_load_i64_unaligned(ptr addrspace(5) nocapture %arg) {
; GFX9-LABEL: store_load_i64_unaligned:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: v_mov_b32_e32 v1, 15
; GFX9-NEXT: v_mov_b32_e32 v2, 0
; GFX9-NEXT: scratch_store_dwordx2 v0, v[1:2], off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: scratch_load_dwordx2 v[0:1], v0, off glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: store_load_i64_unaligned:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: v_mov_b32_e32 v1, 15
; GFX10-NEXT: v_mov_b32_e32 v2, 0
; GFX10-NEXT: scratch_store_dwordx2 v0, v[1:2], off
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dwordx2 v[0:1], v0, off glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_setpc_b64 s[30:31]
;
; GFX940-LABEL: store_load_i64_unaligned:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX940-NEXT: v_mov_b64_e32 v[2:3], 15
; GFX940-NEXT: scratch_store_dwordx2 v0, v[2:3], off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: scratch_load_dwordx2 v[0:1], v0, off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_setpc_b64 s[30:31]
;
; GFX11-LABEL: store_load_i64_unaligned:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-NEXT: v_mov_b32_e32 v1, 15
; GFX11-NEXT: v_mov_b32_e32 v2, 0
; GFX11-NEXT: scratch_store_b64 v0, v[1:2], off dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_load_b64 v[0:1], v0, off glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_setpc_b64 s[30:31]
;
; GFX12-LABEL: store_load_i64_unaligned:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: v_mov_b32_e32 v1, 15
; GFX12-NEXT: v_mov_b32_e32 v2, 0
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_store_b64 v0, v[1:2], off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b64 v[0:1], v0, off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_setpc_b64 s[30:31]
bb:
store volatile i64 15, ptr addrspace(5) %arg, align 1
%load = load volatile i64, ptr addrspace(5) %arg, align 1
ret void
}
define void @store_load_v3i32_unaligned(ptr addrspace(5) nocapture %arg) {
; GFX9-LABEL: store_load_v3i32_unaligned:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: s_mov_b32 s2, 3
; GFX9-NEXT: s_mov_b32 s1, 2
; GFX9-NEXT: s_mov_b32 s0, 1
; GFX9-NEXT: v_mov_b32_e32 v3, s2
; GFX9-NEXT: v_mov_b32_e32 v2, s1
; GFX9-NEXT: v_mov_b32_e32 v1, s0
; GFX9-NEXT: scratch_store_dwordx3 v0, v[1:3], off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: scratch_load_dwordx3 v[0:2], v0, off glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: store_load_v3i32_unaligned:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: s_mov_b32 s2, 3
; GFX10-NEXT: s_mov_b32 s1, 2
; GFX10-NEXT: s_mov_b32 s0, 1
; GFX10-NEXT: v_mov_b32_e32 v3, s2
; GFX10-NEXT: v_mov_b32_e32 v2, s1
; GFX10-NEXT: v_mov_b32_e32 v1, s0
; GFX10-NEXT: scratch_store_dwordx3 v0, v[1:3], off
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dwordx3 v[0:2], v0, off glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_setpc_b64 s[30:31]
;
; GFX940-LABEL: store_load_v3i32_unaligned:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX940-NEXT: s_mov_b32 s2, 3
; GFX940-NEXT: s_mov_b32 s1, 2
; GFX940-NEXT: s_mov_b32 s0, 1
; GFX940-NEXT: v_mov_b32_e32 v4, s2
; GFX940-NEXT: v_mov_b32_e32 v3, s1
; GFX940-NEXT: v_mov_b32_e32 v2, s0
; GFX940-NEXT: scratch_store_dwordx3 v0, v[2:4], off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: scratch_load_dwordx3 v[0:2], v0, off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_setpc_b64 s[30:31]
;
; GFX11-LABEL: store_load_v3i32_unaligned:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-NEXT: s_mov_b32 s2, 3
; GFX11-NEXT: s_mov_b32 s1, 2
; GFX11-NEXT: s_mov_b32 s0, 1
; GFX11-NEXT: v_dual_mov_b32 v3, s2 :: v_dual_mov_b32 v2, s1
; GFX11-NEXT: v_mov_b32_e32 v1, s0
; GFX11-NEXT: scratch_store_b96 v0, v[1:3], off dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_load_b96 v[0:2], v0, off glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_setpc_b64 s[30:31]
;
; GFX12-LABEL: store_load_v3i32_unaligned:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: s_mov_b32 s2, 3
; GFX12-NEXT: s_mov_b32 s1, 2
; GFX12-NEXT: s_mov_b32 s0, 1
; GFX12-NEXT: v_dual_mov_b32 v3, s2 :: v_dual_mov_b32 v2, s1
; GFX12-NEXT: v_mov_b32_e32 v1, s0
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_store_b96 v0, v[1:3], off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b96 v[0:2], v0, off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_setpc_b64 s[30:31]
bb:
store volatile <3 x i32> <i32 1, i32 2, i32 3>, ptr addrspace(5) %arg, align 1
%load = load volatile <3 x i32>, ptr addrspace(5) %arg, align 1
ret void
}
define void @store_load_v4i32_unaligned(ptr addrspace(5) nocapture %arg) {
; GFX9-LABEL: store_load_v4i32_unaligned:
; GFX9: ; %bb.0: ; %bb
; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX9-NEXT: s_mov_b32 s3, 4
; GFX9-NEXT: s_mov_b32 s2, 3
; GFX9-NEXT: s_mov_b32 s1, 2
; GFX9-NEXT: s_mov_b32 s0, 1
; GFX9-NEXT: v_mov_b32_e32 v4, s3
; GFX9-NEXT: v_mov_b32_e32 v3, s2
; GFX9-NEXT: v_mov_b32_e32 v2, s1
; GFX9-NEXT: v_mov_b32_e32 v1, s0
; GFX9-NEXT: scratch_store_dwordx4 v0, v[1:4], off
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: scratch_load_dwordx4 v[0:3], v0, off glc
; GFX9-NEXT: s_waitcnt vmcnt(0)
; GFX9-NEXT: s_setpc_b64 s[30:31]
;
; GFX10-LABEL: store_load_v4i32_unaligned:
; GFX10: ; %bb.0: ; %bb
; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX10-NEXT: s_mov_b32 s3, 4
; GFX10-NEXT: s_mov_b32 s2, 3
; GFX10-NEXT: s_mov_b32 s1, 2
; GFX10-NEXT: s_mov_b32 s0, 1
; GFX10-NEXT: v_mov_b32_e32 v4, s3
; GFX10-NEXT: v_mov_b32_e32 v3, s2
; GFX10-NEXT: v_mov_b32_e32 v2, s1
; GFX10-NEXT: v_mov_b32_e32 v1, s0
; GFX10-NEXT: scratch_store_dwordx4 v0, v[1:4], off
; GFX10-NEXT: s_waitcnt_vscnt null, 0x0
; GFX10-NEXT: scratch_load_dwordx4 v[0:3], v0, off glc dlc
; GFX10-NEXT: s_waitcnt vmcnt(0)
; GFX10-NEXT: s_setpc_b64 s[30:31]
;
; GFX940-LABEL: store_load_v4i32_unaligned:
; GFX940: ; %bb.0: ; %bb
; GFX940-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX940-NEXT: s_mov_b32 s3, 4
; GFX940-NEXT: s_mov_b32 s2, 3
; GFX940-NEXT: s_mov_b32 s1, 2
; GFX940-NEXT: s_mov_b32 s0, 1
; GFX940-NEXT: v_mov_b64_e32 v[4:5], s[2:3]
; GFX940-NEXT: v_mov_b64_e32 v[2:3], s[0:1]
; GFX940-NEXT: scratch_store_dwordx4 v0, v[2:5], off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: scratch_load_dwordx4 v[0:3], v0, off sc0 sc1
; GFX940-NEXT: s_waitcnt vmcnt(0)
; GFX940-NEXT: s_setpc_b64 s[30:31]
;
; GFX11-LABEL: store_load_v4i32_unaligned:
; GFX11: ; %bb.0: ; %bb
; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
; GFX11-NEXT: s_mov_b32 s3, 4
; GFX11-NEXT: s_mov_b32 s2, 3
; GFX11-NEXT: s_mov_b32 s1, 2
; GFX11-NEXT: s_mov_b32 s0, 1
; GFX11-NEXT: v_dual_mov_b32 v4, s3 :: v_dual_mov_b32 v3, s2
; GFX11-NEXT: v_dual_mov_b32 v2, s1 :: v_dual_mov_b32 v1, s0
; GFX11-NEXT: scratch_store_b128 v0, v[1:4], off dlc
; GFX11-NEXT: s_waitcnt_vscnt null, 0x0
; GFX11-NEXT: scratch_load_b128 v[0:3], v0, off glc dlc
; GFX11-NEXT: s_waitcnt vmcnt(0)
; GFX11-NEXT: s_setpc_b64 s[30:31]
;
; GFX12-LABEL: store_load_v4i32_unaligned:
; GFX12: ; %bb.0: ; %bb
; GFX12-NEXT: s_wait_loadcnt_dscnt 0x0
; GFX12-NEXT: s_wait_expcnt 0x0
; GFX12-NEXT: s_wait_samplecnt 0x0
; GFX12-NEXT: s_wait_bvhcnt 0x0
; GFX12-NEXT: s_wait_kmcnt 0x0
; GFX12-NEXT: s_mov_b32 s3, 4
; GFX12-NEXT: s_mov_b32 s2, 3
; GFX12-NEXT: s_mov_b32 s1, 2
; GFX12-NEXT: s_mov_b32 s0, 1
; GFX12-NEXT: v_dual_mov_b32 v4, s3 :: v_dual_mov_b32 v3, s2
; GFX12-NEXT: v_dual_mov_b32 v2, s1 :: v_dual_mov_b32 v1, s0
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_store_b128 v0, v[1:4], off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_storecnt 0x0
; GFX12-NEXT: scratch_load_b128 v[0:3], v0, off scope:SCOPE_SYS
; GFX12-NEXT: s_wait_loadcnt 0x0
; GFX12-NEXT: s_setpc_b64 s[30:31]
bb:
store volatile <4 x i32> <i32 1, i32 2, i32 3, i32 4>, ptr addrspace(5) %arg, align 1
%load = load volatile <4 x i32>, ptr addrspace(5) %arg, align 1
ret void
}
declare i32 @llvm.amdgcn.workitem.id.x()
Reference in New Issue View Git Blame Copy Permalink