public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r14-179] aarch64: Annotate fcvtn pattern for vec_concat with zeroes
@ 2023-04-23 13:44 Kyrylo Tkachov
  0 siblings, 0 replies; only message in thread
From: Kyrylo Tkachov @ 2023-04-23 13:44 UTC (permalink / raw)
  To: gcc-cvs

https://gcc.gnu.org/g:7e26fd6bcd39f53bc917f55f8cce6101180c1dcd

commit r14-179-g7e26fd6bcd39f53bc917f55f8cce6101180c1dcd
Author: Kyrylo Tkachov <kyrylo.tkachov@arm.com>
Date:   Sun Apr 23 14:44:13 2023 +0100

    aarch64: Annotate fcvtn pattern for vec_concat with zeroes
    
    Using the define_substs in aarch64-simd.md this is a straightforward annotation to remove
    a redundant fmov insn.
    
    So the codegen goes from:
    foo_d:
            fcvtn   v0.2s, v0.2d
            fmov    d0, d0
            ret
    
    to the simple:
    foo_d:
            fcvtn   v0.2s, v0.2d
            ret
    
    Bootstrapped and tested on aarch64-none-linux-gnu.
    
    gcc/ChangeLog:
    
            * config/aarch64/aarch64-simd.md (aarch64_float_truncate_lo_): Rename to...
            (aarch64_float_truncate_lo_<mode><vczle><vczbe>): ... This.
    
    gcc/testsuite/ChangeLog:
    
            * gcc.target/aarch64/float_truncate_zero.c: New test.

Diff:
---
 gcc/config/aarch64/aarch64-simd.md                 |  2 +-
 .../gcc.target/aarch64/float_truncate_zero.c       | 32 ++++++++++++++++++++++
 2 files changed, 33 insertions(+), 1 deletion(-)

diff --git a/gcc/config/aarch64/aarch64-simd.md b/gcc/config/aarch64/aarch64-simd.md
index 4a1ec71995d..7bd4362318b 100644
--- a/gcc/config/aarch64/aarch64-simd.md
+++ b/gcc/config/aarch64/aarch64-simd.md
@@ -3331,7 +3331,7 @@
 }
 )
 
-(define_insn "aarch64_float_truncate_lo_<mode>"
+(define_insn "aarch64_float_truncate_lo_<mode><vczle><vczbe>"
   [(set (match_operand:VDF 0 "register_operand" "=w")
       (float_truncate:VDF
 	(match_operand:<VWIDE> 1 "register_operand" "w")))]
diff --git a/gcc/testsuite/gcc.target/aarch64/float_truncate_zero.c b/gcc/testsuite/gcc.target/aarch64/float_truncate_zero.c
new file mode 100644
index 00000000000..41775d164a5
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/float_truncate_zero.c
@@ -0,0 +1,32 @@
+/* { dg-do compile } */
+/* { dg-additional-options "--save-temps -O1" } */
+/* { dg-final { check-function-bodies "**" "" "" } } */
+
+#include <arm_neon.h>
+
+/*
+** foo:
+**      fcvtn	v0.4h, v0.4s
+**      ret
+*/
+
+float16x8_t
+foo (float32x4_t a)
+{
+  float16x4_t b = vcvt_f16_f32 (a);
+  return vcombine_f16 (b, vdup_n_f16 (0.0));
+}
+
+/*
+** foo_d:
+**      fcvtn	v0.2s, v0.2d
+**      ret
+*/
+
+float32x4_t
+foo_d (float64x2_t a)
+{
+  float32x2_t b = vcvt_f32_f64 (a);
+  return vcombine_f32 (b, vdup_n_f32 (0.0));
+}
+

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2023-04-23 13:44 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-04-23 13:44 [gcc r14-179] aarch64: Annotate fcvtn pattern for vec_concat with zeroes Kyrylo Tkachov

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).