projects / gcc.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 8f31650) | patch
re PR tree-optimization/92645 (Hand written vector code is 450 times slower when...
authorRichard Biener <rguenther@suse.de>
Tue, 3 Dec 2019 11:59:13 +0000 (11:59 +0000)
committerRichard Biener <rguenth@gcc.gnu.org>
Tue, 3 Dec 2019 11:59:13 +0000 (11:59 +0000)
commit5105b576dfdcae0c6cc30a89a68b4ef3cbf56a1c
tree5c84bcbb4ae4e3a1b2b326b0cbf5f37160d5b143tree
parent8f316505da1348cca2eb7aeac7919710c255d396commit | diff
re PR tree-optimization/92645 (Hand written vector code is 450 times slower when compiled with GCC compared to Clang)

2019-12-03  Richard Biener  <rguenther@suse.de>

PR tree-optimization/92645
* gimple-fold.c (gimple_fold_builtin_memory_op): Fold memcpy
from or to a properly aligned register variable.

* gcc.target/i386/pr92645-5.c: New testcase.

From-SVN: r278934
gcc/ChangeLog diff | blob | history
gcc/gimple-fold.c diff | blob | history
gcc/testsuite/ChangeLog diff | blob | history
gcc/testsuite/gcc.target/i386/pr92645-5.c [new file with mode: 0644] blob