public inbox for gentoo-dev@lists.gentoo.org
 help / color / mirror / Atom feed
* Fw: Re: [gentoo-dev] -fbranch-probabilities optimisation
@ 2003-05-17 15:43 Priit Laes
  2003-05-17 17:21 ` Chris Davies
  0 siblings, 1 reply; 3+ messages in thread
From: Priit Laes @ 2003-05-17 15:43 UTC (permalink / raw
  To: gentoo-dev

What about implementing it as a USE flag? 
PS. could i try this modified portage?
-- 
Priit Laes <amdATttDOTee>                  _o)
http://amd-core.tk                         /\\  _o)  _o)
I exist, therefore i am paid.             _\_V _( ) _( )

--
gentoo-dev@gentoo.org mailing list


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Fw: Re: [gentoo-dev] -fbranch-probabilities optimisation
  2003-05-17 15:43 Fw: Re: [gentoo-dev] -fbranch-probabilities optimisation Priit Laes
@ 2003-05-17 17:21 ` Chris Davies
  2003-05-18  7:50   ` Amiel Martin
  0 siblings, 1 reply; 3+ messages in thread
From: Chris Davies @ 2003-05-17 17:21 UTC (permalink / raw
  To: gentoo-dev

* Priit Laes (amd@tt.ee) wrote:
> What about implementing it as a USE flag? 
> PS. could i try this modified portage?
> -- 
> Priit Laes <amdATttDOTee>                  _o)
> http://amd-core.tk                         /\\  _o)  _o)
> I exist, therefore i am paid.             _\_V _( ) _( )
> 
> --
> gentoo-dev@gentoo.org mailing list

Yes, a use flag would be quite helpful to possible implementation ideas, it means you I could optionally depend on a utility for determining CFLAGS and fetching the correct profile patch.

I'm provisionally convinced having the user's machine run the profiling is a bad idea for most applications. I've fought mplayer and got some valid profile data from it, but it wasn't easy. This means any implementation must centre around common CFLAGS and maintaining a cache of profile data for applications on a server somewhere.

Thanks,
C.Davies

--
gentoo-dev@gentoo.org mailing list


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Fw: Re: [gentoo-dev] -fbranch-probabilities optimisation
  2003-05-17 17:21 ` Chris Davies
@ 2003-05-18  7:50   ` Amiel Martin
  0 siblings, 0 replies; 3+ messages in thread
From: Amiel Martin @ 2003-05-18  7:50 UTC (permalink / raw
  To: Chris Davies; +Cc: gentoo-dev

wouldnt it make sense not to use a USE flag, but maybe a whole new
variable? would every package have to add this option to their IUSE flags

On Sat, 17 May 2003, Chris Davies wrote:

* Priit Laes (amd@tt.ee) wrote:
> What about implementing it as a USE flag? 
> PS. could i try this modified portage?
> -- 
> Priit Laes <amdATttDOTee>                  _o)
> http://amd-core.tk                         /\\  _o)  _o)
> I exist, therefore i am paid.             _\_V _( ) _( )
> 
> --
> gentoo-dev@gentoo.org mailing list

Yes, a use flag would be quite helpful to possible implementation ideas, it means you I could optionally depend on a utility for determining CFLAGS and fetching the correct profile patch.

I'm provisionally convinced having the user's machine run the profiling is a bad idea for most applications. I've fought mplayer and got some valid profile data from it, but it wasn't easy. This means any implementation must centre around common CFLAGS and maintaining a cache of profile data for applications on a server somewhere.

Thanks,
C.Davies

--
gentoo-dev@gentoo.org mailing list


--
gentoo-dev@gentoo.org mailing list


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2003-05-18  7:50 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2003-05-17 15:43 Fw: Re: [gentoo-dev] -fbranch-probabilities optimisation Priit Laes
2003-05-17 17:21 ` Chris Davies
2003-05-18  7:50   ` Amiel Martin

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox