[Bug regression/33928] 33% performance slowdown from 4.2.2 to 4.3.0 in floating-point code with computed gotos

2007-10-28 Thread rguenth at gcc dot gnu dot org
--- Comment #9 from rguenth at gcc dot gnu dot org 2007-10-28 16:38 --- The main difference I see is that 4.2 avoids re-use of %eax as index register: .L34: movq%r11, %rdi addq8(%r10), %rdi movq8(%r10), %rsi movq8(%r10), %rdx movq

[Bug regression/33928] 33% performance slowdown from 4.2.2 to 4.3.0 in floating-point code with computed gotos

2007-10-28 Thread lucier at math dot purdue dot edu
--- Comment #8 from lucier at math dot purdue dot edu 2007-10-28 16:08 --- Subject: Re: 33% performance slowdown from 4.2.2 to 4.3.0 in floating-point code On Oct 28, 2007, at 8:05 AM, rguenth at gcc dot gnu dot org wrote: > --- Comment #2 from rguenth at gcc dot gnu dot org 20

[Bug regression/33928] 33% performance slowdown from 4.2.2 to 4.3.0 in floating-point code with computed gotos

2007-10-28 Thread lucier at math dot purdue dot edu
--- Comment #7 from lucier at math dot purdue dot edu 2007-10-28 16:05 --- time with -O2 instead of -O1: with 4.2.2: (time (direct-fft-recursive-4 a table)) 426 ms real time 426 ms cpu time (425 user, 1 system) no collections 64 bytes allocated no minor faults

[Bug regression/33928] 33% performance slowdown from 4.2.2 to 4.3.0 in floating-point code with computed gotos

2007-10-28 Thread lucier at math dot purdue dot edu
--- Comment #6 from lucier at math dot purdue dot edu 2007-10-28 15:45 --- Created an attachment (id=14426) --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14426&action=view) assembly after replacing -O1 with -O2 -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33928

[Bug regression/33928] 33% performance slowdown from 4.2.2 to 4.3.0 in floating-point code with computed gotos

2007-10-28 Thread lucier at math dot purdue dot edu
--- Comment #5 from lucier at math dot purdue dot edu 2007-10-28 15:45 --- Created an attachment (id=14425) --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14425&action=view) assembly after replacing -O1 with -O2 -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33928

[Bug regression/33928] 33% performance slowdown from 4.2.2 to 4.3.0 in floating-point code with computed gotos

2007-10-28 Thread lucier at math dot purdue dot edu
--- Comment #4 from lucier at math dot purdue dot edu 2007-10-28 15:42 --- Created an attachment (id=14424) --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14424&action=view) assembly from 4.3.0 I had to remove the "static" from the declaration of direct-fft-recursive to get assemb

[Bug regression/33928] 33% performance slowdown from 4.2.2 to 4.3.0 in floating-point code with computed gotos

2007-10-28 Thread lucier at math dot purdue dot edu
--- Comment #3 from lucier at math dot purdue dot edu 2007-10-28 15:41 --- Created an attachment (id=14423) --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14423&action=view) Assembly from 4.2.2 -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33928

[Bug regression/33928] 33% performance slowdown from 4.2.2 to 4.3.0 in floating-point code

2007-10-28 Thread rguenth at gcc dot gnu dot org
--- Comment #2 from rguenth at gcc dot gnu dot org 2007-10-28 12:05 --- Can you attach assembler files? What happens if you use -O2? Why do you need -fno-strict-aliasing? Does -fno-ivopts help? -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33928