diff options
| author | H.J. Lu <hjl.tools@gmail.com> | 2025-05-01 06:30:41 +0800 |
|---|---|---|
| committer | H.J. Lu <hjl.tools@gmail.com> | 2025-05-07 16:34:08 +0800 |
| commit | b8c4b6aa8e0521770c6f9fd48dd13dd85e3a2fc9 (patch) | |
| tree | 264d8fba621ce27c7ce66aef16a52483f4381ae1 /gcc/testsuite/gcc.target/i386 | |
| parent | ae987baeb5fb70877fd007db80b77e374f676c76 (diff) | |
| download | gcc-master.zip gcc-master.tar.gz gcc-master.tar.bz2 | |
When generating a SUBREG from V16QI to V2HF, validate_subreg fails since
V2HF is a floating point vector and its size (4 bytes) is smaller than its
natural size (word size). Insert an extra move with a QI vector SUBREG of
the same size to avoid validate_subreg failure.
gcc/
PR target/120036
* config/i386/i386-features.cc (ix86_get_vector_load_mode):
Handle 8/4/2 bytes.
(remove_redundant_vector_load): If the mode size is smaller than
its natural size, first insert an extra move with a QI vector
SUBREG of the same size to avoid validate_subreg failure.
gcc/testsuite/
PR target/120036
* g++.target/i386/pr120036.C: New test.
* gcc.target/i386/pr117839-3a.c: Likewise.
* gcc.target/i386/pr117839-3b.c: Likewise.
Signed-off-by: H.J. Lu <hjl.tools@gmail.com>
Diffstat (limited to 'gcc/testsuite/gcc.target/i386')
| -rw-r--r-- | gcc/testsuite/gcc.target/i386/pr117839-3a.c | 22 | ||||
| -rw-r--r-- | gcc/testsuite/gcc.target/i386/pr117839-3b.c | 5 |
2 files changed, 27 insertions, 0 deletions
diff --git a/gcc/testsuite/gcc.target/i386/pr117839-3a.c b/gcc/testsuite/gcc.target/i386/pr117839-3a.c new file mode 100644 index 0000000..81afa9d --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/pr117839-3a.c @@ -0,0 +1,22 @@ +/* { dg-do compile } */ +/* { dg-options "-O2 -mno-avx -msse2 -mtune=generic" } */ +/* { dg-final { scan-assembler-times "xor\[a-z\]*\[\t \]*%xmm\[0-9\]\+,\[^,\]*" 1 } } */ + +typedef char v4qi __attribute__((vector_size(4))); +typedef char v16qi __attribute__((vector_size(16))); + +v4qi a; +v16qi b; +void +foo (v4qi* c, v16qi* d) +{ + v4qi sum = __extension__(v4qi){0, 0, 0, 0}; + v16qi sum2 = __extension__(v16qi){0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, +0, 0, 0, 0, 0}; + for (int i = 0; i != 100; i++) + sum += c[i]; + for (int i = 0 ; i != 100; i++) + sum2 += d[i]; + a = sum; + b = sum2; +} diff --git a/gcc/testsuite/gcc.target/i386/pr117839-3b.c b/gcc/testsuite/gcc.target/i386/pr117839-3b.c new file mode 100644 index 0000000..a599c28 --- /dev/null +++ b/gcc/testsuite/gcc.target/i386/pr117839-3b.c @@ -0,0 +1,5 @@ +/* { dg-do compile } */ +/* { dg-options "-O2 -march=x86-64-v3" } */ +/* { dg-final { scan-assembler-times "xor\[a-z\]*\[\t \]*%xmm\[0-9\]\+,\[^,\]*" 1 } } */ + +#include "pr117839-3a.c" |