diff options
| author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2019-06-20 21:58:24 +0000 |
|---|---|---|
| committer | Matt Arsenault <Matthew.Arsenault@amd.com> | 2019-06-20 21:58:24 +0000 |
| commit | d88db6d7fc942947ad4a068b38c5b5af7d5d1751 (patch) | |
| tree | 670dc13f6ccde2424eded38229f3b60061f5e6a2 /llvm/test/CodeGen/AMDGPU/build-vector-packed-partial-undef.ll | |
| parent | 25f08a17c318e8ffbd30ecbab4d3ea5c5105ddbc (diff) | |
| download | bcm5719-llvm-d88db6d7fc942947ad4a068b38c5b5af7d5d1751.tar.gz bcm5719-llvm-d88db6d7fc942947ad4a068b38c5b5af7d5d1751.zip | |
AMDGPU: Always use s33 for global scratch wave offset
Every called function could possibly need this to calculate the
absolute address of stack objectst, and this avoids inserting a copy
around every call site in the kernel. It's also somewhat cleaner to
keep this in a callee saved SGPR.
llvm-svn: 363990
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU/build-vector-packed-partial-undef.ll')
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/build-vector-packed-partial-undef.ll | 8 |
1 files changed, 4 insertions, 4 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/build-vector-packed-partial-undef.ll b/llvm/test/CodeGen/AMDGPU/build-vector-packed-partial-undef.ll index b78dced5bcf..b160b512b86 100644 --- a/llvm/test/CodeGen/AMDGPU/build-vector-packed-partial-undef.ll +++ b/llvm/test/CodeGen/AMDGPU/build-vector-packed-partial-undef.ll @@ -80,8 +80,8 @@ define void @undef_lo_op_v2i16(i16 %arg0) { ; GFX9: ; %bb.0: ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) ; GFX9-NEXT: v_lshlrev_b32_e32 v0, 16, v0 -; GFX9-NEXT: s_movk_i32 s6, 0x63 -; GFX9-NEXT: v_pk_add_u16 v0, v0, s6 op_sel_hi:[1,0] +; GFX9-NEXT: s_movk_i32 s4, 0x63 +; GFX9-NEXT: v_pk_add_u16 v0, v0, s4 op_sel_hi:[1,0] ; GFX9-NEXT: ;;#ASMSTART ; GFX9-NEXT: ; use v0 ; GFX9-NEXT: ;;#ASMEND @@ -273,8 +273,8 @@ define void @undef_hi_op_v2i16(i16 %arg0) { ; GFX9-LABEL: undef_hi_op_v2i16: ; GFX9: ; %bb.0: ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) -; GFX9-NEXT: s_movk_i32 s6, 0x63 -; GFX9-NEXT: v_pk_add_u16 v0, v0, s6 op_sel_hi:[1,0] +; GFX9-NEXT: s_movk_i32 s4, 0x63 +; GFX9-NEXT: v_pk_add_u16 v0, v0, s4 op_sel_hi:[1,0] ; GFX9-NEXT: ;;#ASMSTART ; GFX9-NEXT: ; use v0 ; GFX9-NEXT: ;;#ASMEND |
