https://gcc.gnu.org/g:ce4a8b7f47aa091b61ae678a65eb2c3187dadedf
commit ce4a8b7f47aa091b61ae678a65eb2c3187dadedf Author: Michael Meissner <[email protected]> Date: Thu Jun 25 05:06:38 2026 -0400 PR target/117251: Improve vector fusion #4 See the following post for a complete explanation of what the patches for PR target/117251: * https://gcc.gnu.org/pipermail/gcc-patches/2025-June/686474.html This is patch #2 of 45 to generate the 'XXEVAL' instruction on power10 and power11 instead of using the Altivec 'VANDC' instruction feeding into 'VAND'. The 'XXEVAL' instruction can use all 64 vector registers, instead of the 32 registers that traditional Altivec vector instructions use. By allowing all of the vector registers to be used, it reduces the amount of spilling that a large benchmark generated. For vectors such as: vector int a, b, c, d; This patch will generate a call to xxeval if any of the registers are allocated to traditional floating point registers: Code: Old: New: ===== ==== ==== a = ~ ((~ (c ^ d)) | b); veqv t,c,d xxeval a,b,c,d,96 vnor a,t,b a = (c ^ d) ^ b; vxor t,c,d xxeval a,b,c,d,105 vxor a,t,b a = (c ^ d) | b; vxor t,c,d xxeval a,b,c,d,111 vor a,t,b a = ~ ((~ (c | d)) | b); vnor t,c,d xxeval a,b,c,d,112 vnor a,t,b a = (c | d) ^ b; vor t,c,d xxeval a,b,c,d,120 vxor a,t,b a = (c | d) | b; vor t,c,d xxeval a,b,c,d,127 vor a,t,b Since fusion using 2 Altivec instructions is slightly faster than using the 'XXEVAL' instruction we prefer to generate the Altivec instructions if we can. In addition, because 'XXEVAL' is a prefixed instruction, it possibly might generate an extra NOP instruction to align the 'XXEVAL' instruction. I have tested these patches on both big endian and little endian PowerPC servers, with no regressions. Can I check these patchs into the trunk? 2026-06-25 Michael Meissner <[email protected]> gcc/ PR target/117251 * config/rs6000/fusion.md: Regenerate. * config/rs6000/genfusion.pl (gen_logical_addsubf): Add support to generate vector/vector fusion if XXEVAL is supported. Diff: --- gcc/config/rs6000/fusion.md | 66 +++++++++++++++++++++++------------------- gcc/config/rs6000/genfusion.pl | 6 ++++ 2 files changed, 42 insertions(+), 30 deletions(-) diff --git a/gcc/config/rs6000/fusion.md b/gcc/config/rs6000/fusion.md index 9c64893fda91..d18904734b70 100644 --- a/gcc/config/rs6000/fusion.md +++ b/gcc/config/rs6000/fusion.md @@ -2535,16 +2535,17 @@ ;; logical-logical fusion pattern generated by gen_logical_addsubf ;; vector veqv -> vnor (define_insn "*fuse_veqv_vnor" - [(set (match_operand:VM 3 "altivec_register_operand" "=&0,&1,&v,v") - (and:VM (not:VM (not:VM (xor:VM (match_operand:VM 0 "altivec_register_operand" "v,v,v,v") - (match_operand:VM 1 "altivec_register_operand" "v,v,v,v")))) - (not:VM (match_operand:VM 2 "altivec_register_operand" "v,v,v,v")))) - (clobber (match_scratch:VM 4 "=X,X,X,&v"))] + [(set (match_operand:VM 3 "vector_fusion_operand" "=&0,&1,&v,wa,v") + (and:VM (not:VM (not:VM (xor:VM (match_operand:VM 0 "vector_fusion_operand" "v,v,v,wa,v") + (match_operand:VM 1 "vector_fusion_operand" "v,v,v,wa,v")))) + (not:VM (match_operand:VM 2 "vector_fusion_operand" "v,v,v,wa,v")))) + (clobber (match_scratch:VM 4 "=X,X,X,X,&v"))] "(TARGET_P10_FUSION)" "@ veqv %3,%1,%0\;vnor %3,%3,%2 veqv %3,%1,%0\;vnor %3,%3,%2 veqv %3,%1,%0\;vnor %3,%3,%2 + xxeval %x3,%x2,%x1,%x0,96 veqv %4,%1,%0\;vnor %3,%4,%2" [(set_attr "type" "fused_vector") (set_attr "cost" "6") @@ -2576,16 +2577,17 @@ ;; logical-logical fusion pattern generated by gen_logical_addsubf ;; vector vnor -> vnor (define_insn "*fuse_vnor_vnor" - [(set (match_operand:VM 3 "altivec_register_operand" "=&0,&1,&v,v") - (and:VM (not:VM (and:VM (not:VM (match_operand:VM 0 "altivec_register_operand" "v,v,v,v")) - (not:VM (match_operand:VM 1 "altivec_register_operand" "v,v,v,v")))) - (not:VM (match_operand:VM 2 "altivec_register_operand" "v,v,v,v")))) - (clobber (match_scratch:VM 4 "=X,X,X,&v"))] + [(set (match_operand:VM 3 "vector_fusion_operand" "=&0,&1,&v,wa,v") + (and:VM (not:VM (and:VM (not:VM (match_operand:VM 0 "vector_fusion_operand" "v,v,v,wa,v")) + (not:VM (match_operand:VM 1 "vector_fusion_operand" "v,v,v,wa,v")))) + (not:VM (match_operand:VM 2 "vector_fusion_operand" "v,v,v,wa,v")))) + (clobber (match_scratch:VM 4 "=X,X,X,X,&v"))] "(TARGET_P10_FUSION)" "@ vnor %3,%1,%0\;vnor %3,%3,%2 vnor %3,%1,%0\;vnor %3,%3,%2 vnor %3,%1,%0\;vnor %3,%3,%2 + xxeval %x3,%x2,%x1,%x0,112 vnor %4,%1,%0\;vnor %3,%4,%2" [(set_attr "type" "fused_vector") (set_attr "cost" "6") @@ -2759,16 +2761,17 @@ ;; logical-logical fusion pattern generated by gen_logical_addsubf ;; vector vor -> vor (define_insn "*fuse_vor_vor" - [(set (match_operand:VM 3 "altivec_register_operand" "=&0,&1,&v,v") - (ior:VM (ior:VM (match_operand:VM 0 "altivec_register_operand" "v,v,v,v") - (match_operand:VM 1 "altivec_register_operand" "%v,v,v,v")) - (match_operand:VM 2 "altivec_register_operand" "v,v,v,v"))) - (clobber (match_scratch:VM 4 "=X,X,X,&v"))] + [(set (match_operand:VM 3 "vector_fusion_operand" "=&0,&1,&v,wa,v") + (ior:VM (ior:VM (match_operand:VM 0 "vector_fusion_operand" "v,v,v,wa,v") + (match_operand:VM 1 "vector_fusion_operand" "%v,v,v,wa,v")) + (match_operand:VM 2 "vector_fusion_operand" "v,v,v,wa,v"))) + (clobber (match_scratch:VM 4 "=X,X,X,X,&v"))] "(TARGET_P10_FUSION)" "@ vor %3,%1,%0\;vor %3,%3,%2 vor %3,%1,%0\;vor %3,%3,%2 vor %3,%1,%0\;vor %3,%3,%2 + xxeval %x3,%x2,%x1,%x0,127 vor %4,%1,%0\;vor %3,%4,%2" [(set_attr "type" "fused_vector") (set_attr "cost" "6") @@ -2799,16 +2802,17 @@ ;; logical-logical fusion pattern generated by gen_logical_addsubf ;; vector vxor -> vor (define_insn "*fuse_vxor_vor" - [(set (match_operand:VM 3 "altivec_register_operand" "=&0,&1,&v,v") - (ior:VM (xor:VM (match_operand:VM 0 "altivec_register_operand" "v,v,v,v") - (match_operand:VM 1 "altivec_register_operand" "v,v,v,v")) - (match_operand:VM 2 "altivec_register_operand" "v,v,v,v"))) - (clobber (match_scratch:VM 4 "=X,X,X,&v"))] + [(set (match_operand:VM 3 "vector_fusion_operand" "=&0,&1,&v,wa,v") + (ior:VM (xor:VM (match_operand:VM 0 "vector_fusion_operand" "v,v,v,wa,v") + (match_operand:VM 1 "vector_fusion_operand" "v,v,v,wa,v")) + (match_operand:VM 2 "vector_fusion_operand" "v,v,v,wa,v"))) + (clobber (match_scratch:VM 4 "=X,X,X,X,&v"))] "(TARGET_P10_FUSION)" "@ vxor %3,%1,%0\;vor %3,%3,%2 vxor %3,%1,%0\;vor %3,%3,%2 vxor %3,%1,%0\;vor %3,%3,%2 + xxeval %x3,%x2,%x1,%x0,111 vxor %4,%1,%0\;vor %3,%4,%2" [(set_attr "type" "fused_vector") (set_attr "cost" "6") @@ -3068,16 +3072,17 @@ ;; logical-logical fusion pattern generated by gen_logical_addsubf ;; vector vor -> vxor (define_insn "*fuse_vor_vxor" - [(set (match_operand:VM 3 "altivec_register_operand" "=&0,&1,&v,v") - (xor:VM (ior:VM (match_operand:VM 0 "altivec_register_operand" "v,v,v,v") - (match_operand:VM 1 "altivec_register_operand" "v,v,v,v")) - (match_operand:VM 2 "altivec_register_operand" "v,v,v,v"))) - (clobber (match_scratch:VM 4 "=X,X,X,&v"))] + [(set (match_operand:VM 3 "vector_fusion_operand" "=&0,&1,&v,wa,v") + (xor:VM (ior:VM (match_operand:VM 0 "vector_fusion_operand" "v,v,v,wa,v") + (match_operand:VM 1 "vector_fusion_operand" "v,v,v,wa,v")) + (match_operand:VM 2 "vector_fusion_operand" "v,v,v,wa,v"))) + (clobber (match_scratch:VM 4 "=X,X,X,X,&v"))] "(TARGET_P10_FUSION)" "@ vor %3,%1,%0\;vxor %3,%3,%2 vor %3,%1,%0\;vxor %3,%3,%2 vor %3,%1,%0\;vxor %3,%3,%2 + xxeval %x3,%x2,%x1,%x0,120 vor %4,%1,%0\;vxor %3,%4,%2" [(set_attr "type" "fused_vector") (set_attr "cost" "6") @@ -3108,16 +3113,17 @@ ;; logical-logical fusion pattern generated by gen_logical_addsubf ;; vector vxor -> vxor (define_insn "*fuse_vxor_vxor" - [(set (match_operand:VM 3 "altivec_register_operand" "=&0,&1,&v,v") - (xor:VM (xor:VM (match_operand:VM 0 "altivec_register_operand" "v,v,v,v") - (match_operand:VM 1 "altivec_register_operand" "%v,v,v,v")) - (match_operand:VM 2 "altivec_register_operand" "v,v,v,v"))) - (clobber (match_scratch:VM 4 "=X,X,X,&v"))] + [(set (match_operand:VM 3 "vector_fusion_operand" "=&0,&1,&v,wa,v") + (xor:VM (xor:VM (match_operand:VM 0 "vector_fusion_operand" "v,v,v,wa,v") + (match_operand:VM 1 "vector_fusion_operand" "%v,v,v,wa,v")) + (match_operand:VM 2 "vector_fusion_operand" "v,v,v,wa,v"))) + (clobber (match_scratch:VM 4 "=X,X,X,X,&v"))] "(TARGET_P10_FUSION)" "@ vxor %3,%1,%0\;vxor %3,%3,%2 vxor %3,%1,%0\;vxor %3,%3,%2 vxor %3,%1,%0\;vxor %3,%3,%2 + xxeval %x3,%x2,%x1,%x0,105 vxor %4,%1,%0\;vxor %3,%4,%2" [(set_attr "type" "fused_vector") (set_attr "cost" "6") diff --git a/gcc/config/rs6000/genfusion.pl b/gcc/config/rs6000/genfusion.pl index d983144b0843..5fec485cdaa0 100755 --- a/gcc/config/rs6000/genfusion.pl +++ b/gcc/config/rs6000/genfusion.pl @@ -232,6 +232,12 @@ sub gen_logical_addsubf "vorc_vnor" => 64, "vorc_veqv" => 75, "vorc_vorc" => 79, + "veqv_vnor" => 96, + "vxor_vxor" => 105, + "vxor_vor" => 111, + "vnor_vnor" => 112, + "vor_vxor" => 120, + "vor_vor" => 127, ); KIND: foreach $kind ('scalar','vector') {
