public inbox for gcc-cvs@sourceware.org help / color / mirror / Atom feed
From: Maciej W. Rozycki <macro@gcc.gnu.org> To: gcc-cvs@gcc.gnu.org Subject: [gcc r14-5734] RISC-V: Avoid extraneous integer comparison for FP comparisons Date: Wed, 22 Nov 2023 01:24:00 +0000 (GMT) [thread overview] Message-ID: <20231122012400.A02363858C3A@sourceware.org> (raw) https://gcc.gnu.org/g:2f0c6252f4e697e187a35427b4b78ba55a830b3d commit r14-5734-g2f0c6252f4e697e187a35427b4b78ba55a830b3d Author: Maciej W. Rozycki <macro@embecosm.com> Date: Wed Nov 22 01:18:30 2023 +0000 RISC-V: Avoid extraneous integer comparison for FP comparisons We have floating-point coditional-set machine instructions for a subset of FP comparisons, so avoid going through a comparison against constant zero in `riscv_expand_float_scc' where not necessary, preventing an extraneous RTL instruction from being produced that counts against the cost of the replacement branchless code sequence in if-conversion, e.g.: (insn 29 6 30 2 (set (reg:DI 142) (ge:DI (reg/v:DF 135 [ w ]) (reg/v:DF 136 [ x ]))) 297 {*cstoredfdi4} (nil)) (insn 30 29 31 2 (set (reg:DI 143) (ne:DI (reg:DI 142) (const_int 0 [0]))) 319 {*sne_zero_didi} (nil)) (insn 31 30 32 2 (set (reg:DI 141) (reg:DI 143)) 206 {*movdi_64bit} (nil)) (insn 32 31 33 2 (set (reg:DI 144) (neg:DI (reg:DI 141))) 15 {negdi2} (nil)) (insn 33 32 34 2 (set (reg:DI 145) (and:DI (reg:DI 144) (reg/v:DI 137 [ y ]))) 102 {*anddi3} (nil)) (insn 34 33 35 2 (set (reg:DI 146) (not:DI (reg:DI 144))) 111 {one_cmpldi2} (nil)) (insn 35 34 36 2 (set (reg:DI 147) (and:DI (reg:DI 146) (reg/v:DI 138 [ z ]))) 102 {*anddi3} (nil)) (insn 36 35 21 2 (set (reg/v:DI 138 [ z ]) (ior:DI (reg:DI 145) (reg:DI 147))) 105 {iordi3} (nil)) where the second insn effectively just copies its input. This now gets simplified to: (insn 29 6 30 2 (set (reg:DI 141) (ge:DI (reg/v:DF 135 [ w ]) (reg/v:DF 136 [ x ]))) 297 {*cstoredfdi4} (nil)) (insn 30 29 31 2 (set (reg:DI 142) (neg:DI (reg:DI 141))) 15 {negdi2} (nil)) (insn 31 30 32 2 (set (reg:DI 143) (and:DI (reg:DI 142) (reg/v:DI 137 [ y ]))) 102 {*anddi3} (nil)) (insn 32 31 33 2 (set (reg:DI 144) (not:DI (reg:DI 142))) 111 {one_cmpldi2} (nil)) (insn 33 32 34 2 (set (reg:DI 145) (and:DI (reg:DI 144) (reg/v:DI 138 [ z ]))) 102 {*anddi3} (nil)) (insn 34 33 21 2 (set (reg/v:DI 138 [ z ]) (ior:DI (reg:DI 143) (reg:DI 145))) 105 {iordi3} (nil)) lowering the cost of the code sequence produced (even though combine would swallow the second insn anyway). We still need to produce a comparison against constant zero where the instruction following a floating-point coditional-set operation is a branch, so add canonicalization to `riscv_expand_conditional_branch' instead. gcc/ * config/riscv/riscv.cc (riscv_emit_float_compare) <NE>: Handle separately. <EQ, LE, LT, GE, GT>: Return operands supplied as is. (riscv_emit_binary): Call `riscv_emit_binary' directly rather than going through a temporary register for word-mode targets. (riscv_expand_conditional_branch): Canonicalize the comparison if not against constant zero. Diff: --- gcc/config/riscv/riscv.cc | 29 +++++++++++++++++++++-------- 1 file changed, 21 insertions(+), 8 deletions(-) diff --git a/gcc/config/riscv/riscv.cc b/gcc/config/riscv/riscv.cc index 998984cfbe6..dc8fc3e719b 100644 --- a/gcc/config/riscv/riscv.cc +++ b/gcc/config/riscv/riscv.cc @@ -4058,9 +4058,10 @@ riscv_emit_float_compare (enum rtx_code *code, rtx *op0, rtx *op1) #undef UNORDERED_COMPARISON case NE: - fp_code = EQ; *code = EQ; - /* Fall through. */ + *op0 = riscv_force_binary (word_mode, EQ, cmp_op0, cmp_op1); + *op1 = const0_rtx; + break; case EQ: case LE: @@ -4068,8 +4069,9 @@ riscv_emit_float_compare (enum rtx_code *code, rtx *op0, rtx *op1) case GE: case GT: /* We have instructions for these cases. */ - *op0 = riscv_force_binary (word_mode, fp_code, cmp_op0, cmp_op1); - *op1 = const0_rtx; + *code = fp_code; + *op0 = cmp_op0; + *op1 = cmp_op1; break; case LTGT: @@ -4109,10 +4111,14 @@ riscv_expand_float_scc (rtx target, enum rtx_code code, rtx op0, rtx op1) { riscv_emit_float_compare (&code, &op0, &op1); - rtx cmp = riscv_force_binary (word_mode, code, op0, op1); - if (GET_MODE (target) != word_mode) - cmp = lowpart_subreg (GET_MODE (target), cmp, word_mode); - riscv_emit_set (target, cmp); + machine_mode mode = GET_MODE (target); + if (mode != word_mode) + { + rtx cmp = riscv_force_binary (word_mode, code, op0, op1); + riscv_emit_set (target, lowpart_subreg (mode, cmp, word_mode)); + } + else + riscv_emit_binary (code, target, op0, op1); } /* Jump to LABEL if (CODE OP0 OP1) holds. */ @@ -4125,6 +4131,13 @@ riscv_expand_conditional_branch (rtx label, rtx_code code, rtx op0, rtx op1) else riscv_emit_int_compare (&code, &op0, &op1); + if (FLOAT_MODE_P (GET_MODE (op0))) + { + op0 = riscv_force_binary (word_mode, code, op0, op1); + op1 = const0_rtx; + code = NE; + } + rtx condition = gen_rtx_fmt_ee (code, VOIDmode, op0, op1); emit_jump_insn (gen_condjump (condition, label)); }
reply other threads:[~2023-11-22 1:24 UTC|newest] Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20231122012400.A02363858C3A@sourceware.org \ --to=macro@gcc.gnu.org \ --cc=gcc-cvs@gcc.gnu.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).