mirror of
https://github.com/RPCSX/llvm.git
synced 2024-12-13 06:38:44 +00:00
7adbf112c7
Re-comitting with a change that avoids undefined uses getting put into the VRegUses list. The new algorithm remembers the uses encountered while walking backwards until a matching def is found. Contrary to the previous version this: - Works without LiveIntervals being available - Allows to increase the precision to subregisters/lanemasks (not used for now) The changes in the AMDGPU tests are necessary because the R600 scheduler is not stable with respect to the order of nodes in the ready queues. Differential Revision: http://reviews.llvm.org/D9068 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@254683 91177308-0d34-0410-b5e6-96231b3b80d8
217 lines
9.4 KiB
LLVM
217 lines
9.4 KiB
LLVM
; RUN: llc -march=amdgcn -mcpu=verde -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=GCN -check-prefix=FUNC %s
|
|
; XUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=VI -check-prefix=GCN -check-prefix=FUNC %s
|
|
; RUN: llc -march=r600 -mcpu=redwood < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
|
|
|
|
declare i32 @llvm.r600.read.tidig.x() #0
|
|
|
|
; FUNC-LABEL: {{^}}lshr_i32:
|
|
; SI: v_lshrrev_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
; VI: v_lshrrev_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
; EG: LSHR {{\*? *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
define void @lshr_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
|
|
%b_ptr = getelementptr i32, i32 addrspace(1)* %in, i32 1
|
|
%a = load i32, i32 addrspace(1)* %in
|
|
%b = load i32, i32 addrspace(1)* %b_ptr
|
|
%result = lshr i32 %a, %b
|
|
store i32 %result, i32 addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}lshr_v2i32:
|
|
; SI: v_lshr_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
; SI: v_lshr_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
|
|
; VI: v_lshrrev_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
; VI: v_lshrrev_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
|
|
; EG: LSHR {{\*? *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
; EG: LSHR {{\*? *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
define void @lshr_v2i32(<2 x i32> addrspace(1)* %out, <2 x i32> addrspace(1)* %in) {
|
|
%b_ptr = getelementptr <2 x i32>, <2 x i32> addrspace(1)* %in, i32 1
|
|
%a = load <2 x i32>, <2 x i32> addrspace(1)* %in
|
|
%b = load <2 x i32>, <2 x i32> addrspace(1)* %b_ptr
|
|
%result = lshr <2 x i32> %a, %b
|
|
store <2 x i32> %result, <2 x i32> addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}lshr_v4i32:
|
|
; SI: v_lshr_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
; SI: v_lshr_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
; SI: v_lshr_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
; SI: v_lshr_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
|
|
; VI: v_lshrrev_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
; VI: v_lshrrev_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
; VI: v_lshrrev_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
; VI: v_lshrrev_b32_e32 v{{[0-9]+, v[0-9]+, v[0-9]+}}
|
|
|
|
; EG: LSHR {{\*? *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
; EG: LSHR {{\*? *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
; EG: LSHR {{\*? *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
; EG: LSHR {{\*? *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
|
|
define void @lshr_v4i32(<4 x i32> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) {
|
|
%b_ptr = getelementptr <4 x i32>, <4 x i32> addrspace(1)* %in, i32 1
|
|
%a = load <4 x i32>, <4 x i32> addrspace(1)* %in
|
|
%b = load <4 x i32>, <4 x i32> addrspace(1)* %b_ptr
|
|
%result = lshr <4 x i32> %a, %b
|
|
store <4 x i32> %result, <4 x i32> addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}lshr_i64:
|
|
; SI: v_lshr_b64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v[0-9]+}}
|
|
; VI: v_lshrrev_b64 {{v\[[0-9]+:[0-9]+\], v[0-9]+, v\[[0-9]+:[0-9]+\]}}
|
|
|
|
; EG: SUB_INT {{\*? *}}[[COMPSH:T[0-9]+\.[XYZW]]], {{literal.[xy]}}, [[SHIFT:T[0-9]+\.[XYZW]]]
|
|
; EG: LSHL {{\* *}}[[TEMP:T[0-9]+\.[XYZW]]], [[OPHI:T[0-9]+\.[XYZW]]], {{[[COMPSH]]|PV.[XYZW]}}
|
|
; EG-DAG: ADD_INT {{\*? *}}[[BIGSH:T[0-9]+\.[XYZW]]], [[SHIFT]], literal
|
|
; EG-DAG: LSHL {{\*? *}}[[OVERF:T[0-9]+\.[XYZW]]], {{[[TEMP]]|PV.[XYZW]}}, 1
|
|
; EG-DAG: LSHR {{\*? *}}[[LOSMTMP:T[0-9]+\.[XYZW]]], [[OPLO:T[0-9]+\.[XYZW]]], [[SHIFT]]
|
|
; EG-DAG: OR_INT {{\*? *}}[[LOSM:T[0-9]+\.[XYZW]]], {{[[LOSMTMP]]|PV.[XYZW]|PS}}, {{[[OVERF]]|PV.[XYZW]}}
|
|
; EG-DAG: LSHR {{\*? *}}[[HISM:T[0-9]+\.[XYZW]]], [[OPHI]], {{PS|[[SHIFT]]|PV\.[XYZW]}}
|
|
; EG-DAG: SETGT_UINT {{\*? *}}[[RESC:T[0-9]+\.[XYZW]]], [[SHIFT]], literal
|
|
; EG-DAG: CNDE_INT {{\*? *}}[[RESLO:T[0-9]+\.[XYZW]]], {{T[0-9]+\.[XYZW]|PS}}
|
|
; EG-DAG: LSHR {{\*? *}}[[LOBIG:T[0-9]+\.[XYZW]]], [[OPHI]], [[SHIFT]]
|
|
; EG-DAG: CNDE_INT {{\*? *}}[[RESHI:T[0-9]+\.[XYZW]]], {{T[0-9]+\.[XYZW], .*}}, 0.0
|
|
define void @lshr_i64(i64 addrspace(1)* %out, i64 addrspace(1)* %in) {
|
|
%b_ptr = getelementptr i64, i64 addrspace(1)* %in, i64 1
|
|
%a = load i64, i64 addrspace(1)* %in
|
|
%b = load i64, i64 addrspace(1)* %b_ptr
|
|
%result = lshr i64 %a, %b
|
|
store i64 %result, i64 addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}lshr_v2i64:
|
|
; SI: v_lshr_b64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v[0-9]+}}
|
|
; SI: v_lshr_b64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v[0-9]+}}
|
|
|
|
; VI: v_lshrrev_b64 {{v\[[0-9]+:[0-9]+\], v[0-9]+, v\[[0-9]+:[0-9]+\]}}
|
|
; VI: v_lshrrev_b64 {{v\[[0-9]+:[0-9]+\], v[0-9]+, v\[[0-9]+:[0-9]+\]}}
|
|
|
|
; EG-DAG: SUB_INT {{\*? *}}[[COMPSHA:T[0-9]+\.[XYZW]]], {{literal.[xy]}}, [[SHA:T[0-9]+\.[XYZW]]]
|
|
; EG-DAG: SUB_INT {{\*? *}}[[COMPSHB:T[0-9]+\.[XYZW]]], {{literal.[xy]}}, [[SHB:T[0-9]+\.[XYZW]]]
|
|
; EG-DAG: LSHL {{\*? *}}[[COMPSHA]]
|
|
; EG-DAG: LSHL {{\*? *}}[[COMPSHB]]
|
|
; EG-DAG: LSHL {{.*}}, 1
|
|
; EG-DAG: LSHL {{.*}}, 1
|
|
; EG-DAG: LSHR {{.*}}, [[SHA]]
|
|
; EG-DAG: LSHR {{.*}}, [[SHB]]
|
|
; EG-DAG: LSHR {{.*}}, [[SHA]]
|
|
; EG-DAG: LSHR {{.*}}, [[SHB]]
|
|
; EG-DAG: OR_INT
|
|
; EG-DAG: OR_INT
|
|
; EG-DAG: ADD_INT {{\*? *}}[[BIGSHA:T[0-9]+\.[XYZW]]]{{.*}}, literal
|
|
; EG-DAG: ADD_INT {{\*? *}}[[BIGSHB:T[0-9]+\.[XYZW]]]{{.*}}, literal
|
|
; EG-DAG: LSHR
|
|
; EG-DAG: LSHR
|
|
; EG-DAG: SETGT_UINT {{\*? *T[0-9]\.[XYZW]}}, [[SHA]], literal
|
|
; EG-DAG: SETGT_UINT {{\*? *T[0-9]\.[XYZW]}}, [[SHB]], literal
|
|
; EG-DAG: CNDE_INT {{.*}}, 0.0
|
|
; EG-DAG: CNDE_INT {{.*}}, 0.0
|
|
; EG-DAG: CNDE_INT
|
|
; EG-DAG: CNDE_INT
|
|
define void @lshr_v2i64(<2 x i64> addrspace(1)* %out, <2 x i64> addrspace(1)* %in) {
|
|
%b_ptr = getelementptr <2 x i64>, <2 x i64> addrspace(1)* %in, i64 1
|
|
%a = load <2 x i64>, <2 x i64> addrspace(1)* %in
|
|
%b = load <2 x i64>, <2 x i64> addrspace(1)* %b_ptr
|
|
%result = lshr <2 x i64> %a, %b
|
|
store <2 x i64> %result, <2 x i64> addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; FUNC-LABEL: {{^}}lshr_v4i64:
|
|
; SI: v_lshr_b64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v[0-9]+}}
|
|
; SI: v_lshr_b64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v[0-9]+}}
|
|
; SI: v_lshr_b64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v[0-9]+}}
|
|
; SI: v_lshr_b64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v[0-9]+}}
|
|
|
|
; VI: v_lshrrev_b64 {{v\[[0-9]+:[0-9]+\], v[0-9]+, v\[[0-9]+:[0-9]+\]}}
|
|
; VI: v_lshrrev_b64 {{v\[[0-9]+:[0-9]+\], v[0-9]+, v\[[0-9]+:[0-9]+\]}}
|
|
; VI: v_lshrrev_b64 {{v\[[0-9]+:[0-9]+\], v[0-9]+, v\[[0-9]+:[0-9]+\]}}
|
|
; VI: v_lshrrev_b64 {{v\[[0-9]+:[0-9]+\], v[0-9]+, v\[[0-9]+:[0-9]+\]}}
|
|
|
|
; EG-DAG: SUB_INT {{\*? *}}[[COMPSHA:T[0-9]+\.[XYZW]]], {{literal.[xy]}}, [[SHA:T[0-9]+\.[XYZW]]]
|
|
; EG-DAG: SUB_INT {{\*? *}}[[COMPSHB:T[0-9]+\.[XYZW]]], {{literal.[xy]}}, [[SHB:T[0-9]+\.[XYZW]]]
|
|
; EG-DAG: SUB_INT {{\*? *}}[[COMPSHC:T[0-9]+\.[XYZW]]], {{literal.[xy]}}, [[SHC:T[0-9]+\.[XYZW]]]
|
|
; EG-DAG: SUB_INT {{\*? *}}[[COMPSHD:T[0-9]+\.[XYZW]]], {{literal.[xy]}}, [[SHD:T[0-9]+\.[XYZW]]]
|
|
; EG-DAG: LSHL {{\*? *}}[[COMPSHA]]
|
|
; EG-DAG: LSHL {{\*? *}}[[COMPSHB]]
|
|
; EG-DAG: LSHL {{\*? *}}[[COMPSHC]]
|
|
; EG-DAG: LSHL {{\*? *}}[[COMPSHD]]
|
|
; EG-DAG: LSHL {{.*}}, 1
|
|
; EG-DAG: LSHL {{.*}}, 1
|
|
; EG-DAG: LSHL {{.*}}, 1
|
|
; EG-DAG: LSHL {{.*}}, 1
|
|
; EG-DAG: LSHR {{.*}}, [[SHA]]
|
|
; EG-DAG: LSHR {{.*}}, [[SHB]]
|
|
; EG-DAG: LSHR {{.*}}, [[SHC]]
|
|
; EG-DAG: LSHR {{.*}}, [[SHD]]
|
|
; EG-DAG: LSHR {{.*}}, [[SHA]]
|
|
; EG-DAG: LSHR {{.*}}, [[SHB]]
|
|
; EG-DAG: LSHR {{.*}}, [[SHC]]
|
|
; EG-DAG: LSHR {{.*}}, [[SHD]]
|
|
; EG-DAG: OR_INT
|
|
; EG-DAG: OR_INT
|
|
; EG-DAG: OR_INT
|
|
; EG-DAG: OR_INT
|
|
; EG-DAG: ADD_INT {{\*? *}}[[BIGSHA:T[0-9]+\.[XYZW]]]{{.*}}, literal
|
|
; EG-DAG: ADD_INT {{\*? *}}[[BIGSHB:T[0-9]+\.[XYZW]]]{{.*}}, literal
|
|
; EG-DAG: ADD_INT {{\*? *}}[[BIGSHC:T[0-9]+\.[XYZW]]]{{.*}}, literal
|
|
; EG-DAG: ADD_INT {{\*? *}}[[BIGSHD:T[0-9]+\.[XYZW]]]{{.*}}, literal
|
|
; EG-DAG: LSHR
|
|
; EG-DAG: LSHR
|
|
; EG-DAG: LSHR
|
|
; EG-DAG: LSHR
|
|
; EG-DAG: LSHR
|
|
; EG-DAG: LSHR
|
|
; EG-DAG: SETGT_UINT {{\*? *T[0-9]\.[XYZW]}}, [[SHA]], literal
|
|
; EG-DAG: SETGT_UINT {{\*? *T[0-9]\.[XYZW]}}, [[SHB]], literal
|
|
; EG-DAG: SETGT_UINT {{\*? *T[0-9]\.[XYZW]}}, [[SHC]], literal
|
|
; EG-DAG: SETGT_UINT {{\*? *T[0-9]\.[XYZW]}}, [[SHD]], literal
|
|
; EG-DAG: CNDE_INT {{.*}}, 0.0
|
|
; EG-DAG: CNDE_INT {{.*}}, 0.0
|
|
; EG-DAG: CNDE_INT {{.*}}, 0.0
|
|
; EG-DAG: CNDE_INT {{.*}}, 0.0
|
|
; EG-DAG: CNDE_INT
|
|
; EG-DAG: CNDE_INT
|
|
; EG-DAG: CNDE_INT
|
|
; EG-DAG: CNDE_INT
|
|
define void @lshr_v4i64(<4 x i64> addrspace(1)* %out, <4 x i64> addrspace(1)* %in) {
|
|
%b_ptr = getelementptr <4 x i64>, <4 x i64> addrspace(1)* %in, i64 1
|
|
%a = load <4 x i64>, <4 x i64> addrspace(1)* %in
|
|
%b = load <4 x i64>, <4 x i64> addrspace(1)* %b_ptr
|
|
%result = lshr <4 x i64> %a, %b
|
|
store <4 x i64> %result, <4 x i64> addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; Make sure load width gets reduced to i32 load.
|
|
; GCN-LABEL: {{^}}s_lshr_32_i64:
|
|
; GCN-DAG: s_load_dword [[HI_A:s[0-9]+]], s{{\[[0-9]+:[0-9]+\]}}, 0xc{{$}}
|
|
; GCN-DAG: v_mov_b32_e32 v[[VHI:[0-9]+]], 0{{$}}
|
|
; GCN-DAG: v_mov_b32_e32 v[[VLO:[0-9]+]], [[HI_A]]
|
|
; GCN: buffer_store_dwordx2 v{{\[}}[[VLO]]:[[VHI]]{{\]}}
|
|
define void @s_lshr_32_i64(i64 addrspace(1)* %out, i64 %a) {
|
|
%result = lshr i64 %a, 32
|
|
store i64 %result, i64 addrspace(1)* %out
|
|
ret void
|
|
}
|
|
|
|
; GCN-LABEL: {{^}}v_lshr_32_i64:
|
|
; GCN-DAG: buffer_load_dword v[[HI_A:[0-9]+]], v{{\[[0-9]+:[0-9]+\]}}, s{{\[[0-9]+:[0-9]+\]}}, 0 addr64 offset:4
|
|
; GCN-DAG: v_mov_b32_e32 v[[VHI:[0-9]+]], 0{{$}}
|
|
; GCN: buffer_store_dwordx2 v{{\[}}[[HI_A]]:[[VHI]]{{\]}}
|
|
define void @v_lshr_32_i64(i64 addrspace(1)* %out, i64 addrspace(1)* %in) {
|
|
%tid = call i32 @llvm.r600.read.tidig.x() #0
|
|
%gep.in = getelementptr i64, i64 addrspace(1)* %in, i32 %tid
|
|
%gep.out = getelementptr i64, i64 addrspace(1)* %out, i32 %tid
|
|
%a = load i64, i64 addrspace(1)* %gep.in
|
|
%result = lshr i64 %a, 32
|
|
store i64 %result, i64 addrspace(1)* %gep.out
|
|
ret void
|
|
}
|
|
|
|
attributes #0 = { nounwind readnone }
|