Re: [x86, PATCH] operand reordering for commutative operations

2015-01-12 Thread Richard Biener
On Mon, Jan 12, 2015 at 4:00 PM, Yuri Rumyantsev wrote: > Hi All, > > Thanks a lot for your comments. > I've re-written reorder_operands as you proposed, but I'd like to know > if we should apply this reordering at -O0? No, I think we can spare those cycles there. Richard. > I will re-send the

Re: [x86, PATCH] operand reordering for commutative operations

2015-01-12 Thread Yuri Rumyantsev
Hi All, Thanks a lot for your comments. I've re-written reorder_operands as you proposed, but I'd like to know if we should apply this reordering at -O0? I will re-send the patch after testing completion. Thanks. Yuri. 2015-01-09 13:13 GMT+03:00 Richard Biener : > On Mon, Jan 5, 2015 at 9:26 PM,

Re: [x86, PATCH] operand reordering for commutative operations

2015-01-09 Thread Richard Biener
On Mon, Jan 5, 2015 at 9:26 PM, Jeff Law wrote: > On 12/29/14 06:30, Yuri Rumyantsev wrote: >> >> Hi All, >> >> Here is a patch which fixed several performance degradation after >> operand canonicalization (r216728). Very simple approach is used - if >> operation is commutative and its second oper

Re: [x86, PATCH] operand reordering for commutative operations

2015-01-05 Thread Jeff Law
On 12/29/14 06:30, Yuri Rumyantsev wrote: Hi All, Here is a patch which fixed several performance degradation after operand canonicalization (r216728). Very simple approach is used - if operation is commutative and its second operand required more operations (statements) for computation, swap op

Re: [x86, PATCH] operand reordering for commutative operations

2014-12-29 Thread Yuri Rumyantsev
Hi H.J. Bug with reproducer was submitted: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64434 2014-12-29 16:53 GMT+03:00 H.J. Lu : > On Mon, Dec 29, 2014 at 5:30 AM, Yuri Rumyantsev wrote: >> Hi All, >> >> Here is a patch which fixed several performance degradation after >> operand canonicaliz

Re: [x86, PATCH] operand reordering for commutative operations

2014-12-29 Thread H.J. Lu
On Mon, Dec 29, 2014 at 5:30 AM, Yuri Rumyantsev wrote: > Hi All, > > Here is a patch which fixed several performance degradation after > operand canonicalization (r216728). Very simple approach is used - if > operation is commutative and its second operand required more > operations (statements)

[x86, PATCH] operand reordering for commutative operations

2014-12-29 Thread Yuri Rumyantsev
Hi All, Here is a patch which fixed several performance degradation after operand canonicalization (r216728). Very simple approach is used - if operation is commutative and its second operand required more operations (statements) for computation, swap operands. Currently this is done under special