From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Cc: peter.maydell@linaro.org
Subject: [PULL 11/31] tcg/optimize: Optimize cmp_vec and cmpsel_vec
Date: Sun, 22 Sep 2024 14:00:52 +0200 [thread overview]
Message-ID: <20240922120112.5067-12-richard.henderson@linaro.org> (raw)
In-Reply-To: <20240922120112.5067-1-richard.henderson@linaro.org>
Place immediate values second in the comparison.
Place destination matches first in the true/false values.
All of this mirrors what we do for integer setcond and movcond.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/optimize.c | 36 ++++++++++++++++++++++++++++++++++++
1 file changed, 36 insertions(+)
diff --git a/tcg/optimize.c b/tcg/optimize.c
index cf311790e0..f11f576fd4 100644
--- a/tcg/optimize.c
+++ b/tcg/optimize.c
@@ -2422,6 +2422,36 @@ static bool fold_setcond2(OptContext *ctx, TCGOp *op)
return tcg_opt_gen_movi(ctx, op, op->args[0], i);
}
+static bool fold_cmp_vec(OptContext *ctx, TCGOp *op)
+{
+ /* Canonicalize the comparison to put immediate second. */
+ if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) {
+ op->args[3] = tcg_swap_cond(op->args[3]);
+ }
+ return false;
+}
+
+static bool fold_cmpsel_vec(OptContext *ctx, TCGOp *op)
+{
+ /* If true and false values are the same, eliminate the cmp. */
+ if (args_are_copies(op->args[3], op->args[4])) {
+ return tcg_opt_gen_mov(ctx, op, op->args[0], op->args[3]);
+ }
+
+ /* Canonicalize the comparison to put immediate second. */
+ if (swap_commutative(NO_DEST, &op->args[1], &op->args[2])) {
+ op->args[5] = tcg_swap_cond(op->args[5]);
+ }
+ /*
+ * Canonicalize the "false" input reg to match the destination,
+ * so that the tcg backend can implement "move if true".
+ */
+ if (swap_commutative(op->args[0], &op->args[4], &op->args[3])) {
+ op->args[5] = tcg_invert_cond(op->args[5]);
+ }
+ return false;
+}
+
static bool fold_sextract(OptContext *ctx, TCGOp *op)
{
uint64_t z_mask, s_mask, s_mask_old;
@@ -2928,6 +2958,12 @@ void tcg_optimize(TCGContext *s)
case INDEX_op_setcond2_i32:
done = fold_setcond2(&ctx, op);
break;
+ case INDEX_op_cmp_vec:
+ done = fold_cmp_vec(&ctx, op);
+ break;
+ case INDEX_op_cmpsel_vec:
+ done = fold_cmpsel_vec(&ctx, op);
+ break;
CASE_OP_32_64(sextract):
done = fold_sextract(&ctx, op);
break;
--
2.43.0
next prev parent reply other threads:[~2024-09-22 12:05 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-09-22 12:00 [PULL 00/31] tcg + linux-user patch queue Richard Henderson
2024-09-22 12:00 ` [PULL 01/31] tcg: Return TCGOp from tcg_gen_op[1-6] Richard Henderson
2024-09-22 12:00 ` [PULL 02/31] tcg: Propagate new TCGOp to add_as_label_use Richard Henderson
2024-09-22 12:00 ` [PULL 03/31] tcg: Fix iteration step in 32-bit gvec operation Richard Henderson
2024-09-22 12:00 ` [PULL 04/31] tcg: Export vec_gen_6 Richard Henderson
2024-09-22 12:00 ` [PULL 05/31] tcg/i386: Split out tcg_out_vex_modrm_type Richard Henderson
2024-09-22 12:00 ` [PULL 06/31] tcg/i386: Do not expand cmp_vec early Richard Henderson
2024-09-22 12:00 ` [PULL 07/31] tcg/i386: Do not expand cmpsel_vec early Richard Henderson
2024-09-22 12:00 ` [PULL 08/31] tcg/ppc: Do not expand cmp_vec early Richard Henderson
2024-09-22 12:00 ` [PULL 09/31] tcg/s390x: " Richard Henderson
2024-09-22 12:00 ` [PULL 10/31] tcg/optimize: Fold movcond with true and false values identical Richard Henderson
2024-09-22 12:00 ` Richard Henderson [this message]
2024-09-22 12:00 ` [PULL 12/31] tcg/optimize: Optimize bitsel_vec Richard Henderson
2024-09-22 12:00 ` [PULL 13/31] tcg/i386: Optimize cmpsel with constant 0 operand 3 Richard Henderson
2024-09-22 12:00 ` [PULL 14/31] tcg/i386: Implement cmp_vec with avx512 insns Richard Henderson
2024-09-22 12:00 ` [PULL 15/31] tcg/i386: Add predicate parameters to tcg_out_evex_opc Richard Henderson
2024-09-22 12:00 ` [PULL 16/31] tcg/i386: Implement cmpsel_vec with avx512 insns Richard Henderson
2024-09-22 12:00 ` [PULL 17/31] tcg/i386: Implement vector TST{EQ,NE} for avx512 Richard Henderson
2024-09-22 12:00 ` [PULL 18/31] tcg/ppc: Implement cmpsel_vec Richard Henderson
2024-09-22 12:01 ` [PULL 19/31] tcg/ppc: Optimize cmpsel with constant 0/-1 arguments Richard Henderson
2024-09-22 12:01 ` [PULL 20/31] tcg/s390x: Implement cmpsel_vec Richard Henderson
2024-09-22 12:01 ` [PULL 21/31] tcg/s390x: Optimize cmpsel with constant 0/-1 arguments Richard Henderson
2024-09-22 12:01 ` [PULL 22/31] target/ppc: Fix lxvx/stxvx facility check Richard Henderson
2024-09-22 12:01 ` [PULL 23/31] linux-user: update syscall_nr.h to Linux v6.10 Richard Henderson
2024-09-22 12:01 ` [PULL 24/31] linux-user, mips: update syscall-args-o32.c.inc " Richard Henderson
2024-09-22 12:01 ` [PULL 25/31] linux-user: update syscall.tbl " Richard Henderson
2024-09-22 12:01 ` [PULL 26/31] linux-user,aarch64: move to syscalltbl file Richard Henderson
2024-09-22 12:01 ` [PULL 27/31] linux-user,openrisc: " Richard Henderson
2024-09-22 12:01 ` [PULL 28/31] linux-user,riscv: " Richard Henderson
2024-09-22 12:01 ` [PULL 29/31] linux-user,hexagon: " Richard Henderson
2024-09-22 12:01 ` [PULL 30/31] linux-user,loongarch: " Richard Henderson
2024-09-22 12:01 ` [PULL 31/31] linux-user: update syscall.tbl to Linux v6.11 Richard Henderson
2024-09-27 15:18 ` [PULL 00/31] tcg + linux-user patch queue Peter Maydell
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=20240922120112.5067-12-richard.henderson@linaro.org \
--to=richard.henderson@linaro.org \
--cc=peter.maydell@linaro.org \
--cc=qemu-devel@nongnu.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: link
Be 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 NNTP newsgroup(s).