projects / gcc.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 70fef63) | patch
re PR rtl-optimization/34522 (inefficient code for long long multiply when only low...
authorPaolo Bonzini <bonzini@gnu.org>
Wed, 12 Mar 2008 15:33:45 +0000 (15:33 +0000)
committerPaolo Bonzini <bonzini@gcc.gnu.org>
Wed, 12 Mar 2008 15:33:45 +0000 (15:33 +0000)
commitc83bd37c9d8859d4e8dc6d1f021c6e8b8ffabb99
treec1959cceb74dfe8a50fb7fa66a39bd956922bdbbtree
parent70fef63aef8a841f20cfa08c88da9ca871a3c8f3commit | diff
re PR rtl-optimization/34522 (inefficient code for long long multiply when only low bits are needed)

2008-03-12  Paolo Bonzini  <bonzini@gnu.org>

PR tree-opt/35422
* fold-const.c (fold_unary) <NOP_EXPR>: Distribute a narrowing
conversion to the operands of a multiplication.

testsuite:
2008-03-12  Paolo Bonzini  <bonzini@gnu.org>

PR tree-opt/35422
* gcc.dg/vect/slp-7.c: Change target keywords required for vectorizing
third loop.
* gcc.target/i386/pr35422.c: New.

From-SVN: r133144
gcc/ChangeLog diff | blob | history
gcc/fold-const.c diff | blob | history
gcc/testsuite/ChangeLog diff | blob | history
gcc/testsuite/gcc.dg/vect/slp-7.c diff | blob | history
gcc/testsuite/gcc.target/i386/pr34522.c [new file with mode: 0644] blob