public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc(refs/users/meissner/heads/work056)] Split loading DImode integers early.
@ 2021-06-23 1:23 Michael Meissner
0 siblings, 0 replies; only message in thread
From: Michael Meissner @ 2021-06-23 1:23 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:1f597e0024c1b30a5d3a0c77491f464d7235d53a
commit 1f597e0024c1b30a5d3a0c77491f464d7235d53a
Author: Michael Meissner <meissner@linux.ibm.com>
Date: Tue Jun 22 21:23:17 2021 -0400
Split loading DImode integers early.
gcc/
2021-06-22 Michael Meissner <meissner@linux.ibm.com>
* config/rs6000/rs6000.md (load DImode constant spliters): Use
int_reg_operand instead of int_reg_operand_not_pseudo so loading
up large DImode constants happens before register allocation.
Diff:
---
gcc/config/rs6000/rs6000.md | 22 ++++++++++++----------
1 file changed, 12 insertions(+), 10 deletions(-)
diff --git a/gcc/config/rs6000/rs6000.md b/gcc/config/rs6000/rs6000.md
index 3ba0beaee32..4160efa36cc 100644
--- a/gcc/config/rs6000/rs6000.md
+++ b/gcc/config/rs6000/rs6000.md
@@ -9404,29 +9404,32 @@
; Some DImode loads are best done as a load of -1 followed by a mask
; instruction.
(define_split
- [(set (match_operand:DI 0 "int_reg_operand_not_pseudo")
+ [(set (match_operand:DI 0 "int_reg_operand")
(match_operand:DI 1 "const_int_operand"))]
"TARGET_POWERPC64
&& num_insns_constant (operands[1], DImode) > 1
&& !IN_RANGE (INTVAL (operands[1]), -0x80000000, 0xffffffff)
&& rs6000_is_valid_and_mask (operands[1], DImode)"
- [(set (match_dup 0)
+ [(set (match_dup 2)
(const_int -1))
(set (match_dup 0)
- (and:DI (match_dup 0)
+ (and:DI (match_dup 2)
(match_dup 1)))]
- "")
+{
+ operands[2] = ((can_create_pseudo_p ())
+ ? gen_reg_rtx (DImode)
+ : operands[0]);
+})
;; Split a load of a large constant into the appropriate five-instruction
;; sequence. Handle anything in a constant number of insns.
;; When non-easy constants can go in the TOC, this should use
;; easy_fp_constant predicate.
(define_split
- [(set (match_operand:DI 0 "int_reg_operand_not_pseudo")
+ [(set (match_operand:DI 0 "int_reg_operand")
(match_operand:DI 1 "const_int_operand"))]
"TARGET_POWERPC64 && num_insns_constant (operands[1], DImode) > 1"
- [(set (match_dup 0) (match_dup 2))
- (set (match_dup 0) (plus:DI (match_dup 0) (match_dup 3)))]
+ [(pc)]
{
if (rs6000_emit_set_const (operands[0], operands[1]))
DONE;
@@ -9435,11 +9438,10 @@
})
(define_split
- [(set (match_operand:DI 0 "int_reg_operand_not_pseudo")
+ [(set (match_operand:DI 0 "int_reg_operand")
(match_operand:DI 1 "const_scalar_int_operand"))]
"TARGET_POWERPC64 && num_insns_constant (operands[1], DImode) > 1"
- [(set (match_dup 0) (match_dup 2))
- (set (match_dup 0) (plus:DI (match_dup 0) (match_dup 3)))]
+ [(pc)]
{
if (rs6000_emit_set_const (operands[0], operands[1]))
DONE;
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2021-06-23 1:23 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-06-23 1:23 [gcc(refs/users/meissner/heads/work056)] Split loading DImode integers early Michael Meissner
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).