| 1 |
Klavs Klavsen wrote: |
| 2 |
|
| 3 |
>Chris Gianelloni said: |
| 4 |
> |
| 5 |
> |
| 6 |
>> While I agree that there can be great performance increases, I believe |
| 7 |
>> |
| 8 |
>>that there is a definite trade-off between performance and |
| 9 |
>>manageability. This would be wholly unmanageable without an army of |
| 10 |
>>testers working around the clock until Gentoo ceased to be... *grin* |
| 11 |
>> |
| 12 |
>> |
| 13 |
>> |
| 14 |
>The idea would ofcourse be that, only the "obvious" programs would be |
| 15 |
>tested - but if profiling were implemented/possible with gcc-3.5 and |
| 16 |
>portage easily - I'm fairly certain that would be of more value (would |
| 17 |
>that also help select the right CFLAGS ?) |
| 18 |
> |
| 19 |
> |
| 20 |
> |
| 21 |
I would tend to agree here - if GCC 3.5 has features that can |
| 22 |
automatically profile applications and use the correct optimisations |
| 23 |
then there would be little point in spending the time doing this by hand |
| 24 |
(even if using some automated test scripts). |
| 25 |
|
| 26 |
I personally saw it being of use for larger applications in the tree, |
| 27 |
such as Apache, MySQL, PostgreSQL, PHP, Postfix, bincimap, Courier, KDE, |
| 28 |
GNOME, GCC, GlibC etc. It would certainly be an impossible task to |
| 29 |
perform this work on all of the tree for all archs :) |
| 30 |
|
| 31 |
I personally use amd64 platform with fairly modest CFLAGS="-march=k8 -O2 |
| 32 |
-pipe" at the moment - going for best all round performance with |
| 33 |
stability. There are things such as fftw where I do try to optimise if |
| 34 |
possible, and I would be interested in a 10% speed gain on the fftw |
| 35 |
library for a 5 hour simulation run! |
| 36 |
|
| 37 |
-- |
| 38 |
gentoo-dev@g.o mailing list |
Archives