| 1 |
On Fri, 02 Dec 2011 20:36:16 -0500 |
| 2 |
Mike Gilbert <floppym@g.o> wrote: |
| 3 |
|
| 4 |
> I've never thought of -mfpmath=sse as a "ricer" flag. If anything, it |
| 5 |
> should make floating point calculations more consistent. |
| 6 |
|
| 7 |
Not really. It's it's extremely target processor dependent and generally the |
| 8 |
defaults are the best settings. It can vary a lot based on the SSE |
| 9 |
instructions and number of registers available and the register allocator and |
| 10 |
scheduler are optimized with that in mind. |
| 11 |
|
| 12 |
> The varying behavior between gcc versions seems like a possibly good |
| 13 |
> reason to not "support" it, but I'm not familiar with this. Has it |
| 14 |
> stabilized recently or is it still in flux? |
| 15 |
|
| 16 |
I'd say it's still in flux. The register allocator and instruction scheduler |
| 17 |
were both rewritten for 4.4 and have seen a lot of work since then. |
| 18 |
|
| 19 |
|
| 20 |
-- |
| 21 |
fonts, gcc-porting, it makes no sense how it makes no sense |
| 22 |
toolchain, wxwidgets but i'll take it free anytime |
| 23 |
@ gentoo.org EFFD 380E 047A 4B51 D2BD C64F 8AA8 8346 F9A4 0662 |
Archives