diff --git a/lib/Target/R600/SIInstructions.td b/lib/Target/R600/SIInstructions.td index 188c2cdae19..75764a430c4 100644 --- a/lib/Target/R600/SIInstructions.td +++ b/lib/Target/R600/SIInstructions.td @@ -3230,6 +3230,12 @@ def : Pat < (V_CMP_EQ_I32_e64 (V_AND_B32_e64 (i32 1), $a), 1) >; +def : Pat < + (i1 (trunc i64:$a)), + (V_CMP_EQ_I32_e64 (V_AND_B32_e64 (i32 1), + (EXTRACT_SUBREG $a, sub0)), 1) +>; + def : Pat < (i32 (bswap i32:$a)), (V_BFI_B32 (S_MOV_B32 0x00ff00ff), diff --git a/test/CodeGen/R600/trunc.ll b/test/CodeGen/R600/trunc.ll index eb7e4ed8f9f..fa44264e365 100644 --- a/test/CodeGen/R600/trunc.ll +++ b/test/CodeGen/R600/trunc.ll @@ -1,6 +1,8 @@ ; RUN: llc -march=amdgcn -mcpu=SI -verify-machineinstrs< %s | FileCheck -check-prefix=SI %s ; RUN: llc -march=r600 -mcpu=cypress < %s | FileCheck -check-prefix=EG %s +declare i32 @llvm.r600.read.tidig.x() nounwind readnone + define void @trunc_i64_to_i32_store(i32 addrspace(1)* %out, i64 %in) { ; SI-LABEL: {{^}}trunc_i64_to_i32_store: ; SI: s_load_dword [[SLOAD:s[0-9]+]], s[0:1], 0xb @@ -67,3 +69,32 @@ define void @sgpr_trunc_i32_to_i1(i32 addrspace(1)* %out, i32 %a) { store i32 %result, i32 addrspace(1)* %out, align 4 ret void } + +; SI-LABEL: {{^}}s_trunc_i64_to_i1: +; SI: s_load_dwordx2 s{{\[}}[[SLO:[0-9]+]]:{{[0-9]+\]}}, {{s\[[0-9]+:[0-9]+\]}}, 0xb +; SI: v_and_b32_e64 [[MASKED:v[0-9]+]], 1, s[[SLO]] +; SI: v_cmp_eq_i32_e64 [[CMP:s\[[0-9]+:[0-9]+\]]], [[MASKED]], 1 +; SI: v_cndmask_b32_e64 {{v[0-9]+}}, -12, 63, [[CMP]] +define void @s_trunc_i64_to_i1(i32 addrspace(1)* %out, i64 %x) { + %trunc = trunc i64 %x to i1 + %sel = select i1 %trunc, i32 63, i32 -12 + store i32 %sel, i32 addrspace(1)* %out + ret void +} + +; SI-LABEL: {{^}}v_trunc_i64_to_i1: +; SI: buffer_load_dwordx2 v{{\[}}[[VLO:[0-9]+]]:{{[0-9]+\]}} +; SI: v_and_b32_e32 [[MASKED:v[0-9]+]], 1, v[[VLO]] +; SI: v_cmp_eq_i32_e64 [[CMP:s\[[0-9]+:[0-9]+\]]], [[MASKED]], 1 +; SI: v_cndmask_b32_e64 {{v[0-9]+}}, -12, 63, [[CMP]] +define void @v_trunc_i64_to_i1(i32 addrspace(1)* %out, i64 addrspace(1)* %in) { + %tid = call i32 @llvm.r600.read.tidig.x() nounwind readnone + %gep = getelementptr i64 addrspace(1)* %in, i32 %tid + %out.gep = getelementptr i32 addrspace(1)* %out, i32 %tid + %x = load i64 addrspace(1)* %gep + + %trunc = trunc i64 %x to i1 + %sel = select i1 %trunc, i32 63, i32 -12 + store i32 %sel, i32 addrspace(1)* %out.gep + ret void +}