Summary: | gcc losing optimization with -O3 | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Bret Towe <magnade> |
Component: | New packages | Assignee: | Please assign to toolchain <gcc-porting> |
Status: | RESOLVED INVALID | ||
Severity: | normal | ||
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Bret Towe
2003-08-05 11:02:53 UTC
i ran some tests on a dual ppro2 200mhz you get about 10 minute compile time difference for compiling a 2.5 kernel oh gcc version is gcc-3.2.3-r1 Resolving as INVALID as I really don't think there is much we can do, unless you can provide some more information. From what I've heard, gcc has gotten progressively slower in code compiliation time since gcc-2.95.3, and it also produces larger executables. I believe the GCC Steering Committee is planning on addressing this issue in gcc-3.4 to speed up compiliation time as well as reduce generated code size (or something to that effect). It is also becoming common knowledge that -O3 really doesn't offer much more optimization because I believe it is just -O2 with -funroll-loops tacked on (and maybe one or two other optimizations), so people are likely to stray away from using -O3 constantly, and only in certain situations. |