diff options
| author | Andrew Stubbs <ams@codesourcery.com> | 2020-03-17 12:49:19 +0000 |
|---|---|---|
| committer | Andrew Stubbs <ams@codesourcery.com> | 2020-03-18 12:53:26 +0000 |
| commit | dbde9e2d5952ff1c50c4aeaaabb23cebafb50759 (patch) | |
| tree | 10575e9e6ce584de02602c55d378d2116488b49f /gcc/config/gcn/gcn-valu.md | |
| parent | 5a80a6c3e5f800de63a2eadd8ae3e6822172a718 (diff) | |
| download | gcc-dbde9e2d5952ff1c50c4aeaaabb23cebafb50759.zip gcc-dbde9e2d5952ff1c50c4aeaaabb23cebafb50759.tar.gz gcc-dbde9e2d5952ff1c50c4aeaaabb23cebafb50759.tar.bz2 | |
amdgcn: Fix vector compare modes
The GCN VCC register has 64 CC values in one registers, one bit for each
vector lane.
Previously we avoided problems with invalid optimizations by not declaring
a mode for the comparison operators, but it turns out that causes other
problems (and build warnings).
Instead, the optimization issues can be avoided by setting
STORE_REGISTER_VALUE to -1, meaning that all the bits are significant.
(It would be better if we could set STORE_REGISTER_VALUE according to the
known mask or vector size, but we can't.)
2020-03-18 Andrew Stubbs <ams@codesourcery.com>
gcc/
* config/gcn/gcn-valu.md (vec_cmp<mode>di): Set operand 1 to DImode.
(vec_cmp<mode>di_dup): Likewise.
* config/gcn/gcn.h (STORE_FLAG_VALUE): Set to -1.
Diffstat (limited to 'gcc/config/gcn/gcn-valu.md')
| -rw-r--r-- | gcc/config/gcn/gcn-valu.md | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/gcc/config/gcn/gcn-valu.md b/gcc/config/gcn/gcn-valu.md index 68d89fa..d362068 100644 --- a/gcc/config/gcn/gcn-valu.md +++ b/gcc/config/gcn/gcn-valu.md @@ -2549,7 +2549,7 @@ (define_insn "vec_cmp<mode>di" [(set (match_operand:DI 0 "register_operand" "=cV,cV, e, e,Sg,Sg") - (match_operator 1 "gcn_fp_compare_operator" + (match_operator:DI 1 "gcn_fp_compare_operator" [(match_operand:VCMP_MODE 2 "gcn_alu_operand" "vSv, B,vSv, B, v,vA") (match_operand:VCMP_MODE 3 "gcn_vop3_operand" @@ -2658,7 +2658,7 @@ (define_insn "vec_cmp<mode>di_dup" [(set (match_operand:DI 0 "register_operand" "=cV,cV, e,e,Sg") - (match_operator 1 "gcn_fp_compare_operator" + (match_operator:DI 1 "gcn_fp_compare_operator" [(vec_duplicate:VCMP_MODE (match_operand:<SCALAR_MODE> 2 "gcn_alu_operand" " Sv, B,Sv,B, A")) |