1 |
On Wed, Nov 16, 2011 at 2:58 PM, Pandu Poluan <pandu@××××××.info> wrote: |
2 |
> |
3 |
> On Nov 16, 2011 2:26 PM, "Michael Mol" <mikemol@×××××.com> wrote: |
4 |
>> |
5 |
>> On Wed, Nov 16, 2011 at 2:11 AM, Stéphane Guedon <stephane@××××××××××.eu> |
6 |
>> wrote: |
7 |
>> > On Wednesday 16 November 2011 02:07:12 Pandu Poluan wrote: |
8 |
>> >> And if you're adventurous, add USE "graphite", reemerge gcc, and |
9 |
>> >> reemerge |
10 |
>> >> world :) |
11 |
>> > |
12 |
>> > what does "graphite" add ? |
13 |
>> |
14 |
>> Thanks for reminding me; I meant to look it up when I got home. |
15 |
>> |
16 |
>> shortcircuit:1@serenity~ |
17 |
>> Wed Nov 16 02:16 AM |
18 |
>> !501 #1 j0 ?0 $ euse -i graphite |
19 |
>> global use flags (searching: graphite) |
20 |
>> ************************************************************ |
21 |
>> no matching entries found |
22 |
>> |
23 |
>> local use flags (searching: graphite) |
24 |
>> ************************************************************ |
25 |
>> |
26 |
>> [snip] |
27 |
>> |
28 |
>> [- ] graphite |
29 |
>> sys-devel/gcc: Add support for the framework for loop optimizations |
30 |
>> based on a polyhedral intermediate representation |
31 |
>> |
32 |
>> So, a new, experimental optimization model and framework inside your |
33 |
>> compiler. If it's specifically for optimizing on loops, I'll venture a |
34 |
>> guess it's going to be mostly effective for graphics libraries and |
35 |
>> apps. I've got some slightly riskier educated guesses on how it works |
36 |
>> and what some numeric side effects and consequences might be, but they |
37 |
>> scare me, so I think I'll leave it to someone who actually knows more |
38 |
>> about it... |
39 |
>> |
40 |
> |
41 |
> I've been using USE "graphite" since gcc-4.5.3-r1 appeared. Upstream says |
42 |
> that graphite is stable, feature-complete, and production-ready since 4.5.3. |
43 |
> |
44 |
> To fully taste the effect of graphite, I even went the torturous route of |
45 |
> emerging gcc + libtool + binutils (in that order) twice, followed by a |
46 |
> wholesale-rebuild of everything (emerge --emptytree), then tarballed the |
47 |
> result to my own "stage3.1" tarball to spare me the *huge* amount of time |
48 |
> required. |
49 |
> |
50 |
> I've deployed 3 systems with USE "graphite", and they *felt* snappier. |
51 |
> emerge's *felt* slower, though. (no objective tests, I know). |
52 |
> |
53 |
> I use Gentoo as a gatewall, and there I did a wholesale-rebuild one more |
54 |
> time, this time specifying CFLAGS "-march=native"... and I just couldn't be |
55 |
> happier with the resulting performance :-) |
56 |
> |
57 |
> Rgds, |
58 |
> |
59 |
|
60 |
I might be wrong but don't you need to have the gcc's options for |
61 |
graphite enabled to actually make use of the graphite framework? (You |
62 |
might be using them but you haven't mentioned it.) |
63 |
|
64 |
//Fredric |
65 |
|
66 |
//Fredric |