1 |
On Fri, 16 Sep 2011 08:36:32 -0500 |
2 |
Donnie Berkholz <dberkholz@g.o> wrote: |
3 |
|
4 |
> On 15:34 Thu 15 Sep , Mike Frysinger wrote: |
5 |
> > ive converted my system over to x86/amd64/x32 multilib for funs. |
6 |
> > but i can see how some people wont want all three all the time. so |
7 |
> > the question is how we want to make this available to users at the |
8 |
> > release/profile level. |
9 |
> > |
10 |
> > background: x32 is a new ABI that runs on 64bit x86_64 processors. |
11 |
> > see [1]. you'll need gcc-4.7+, binutils-2.21.50+, glibc-2.15+, and |
12 |
> > linux-3.2+. |
13 |
> |
14 |
> For anyone interested how the performance compares to amd64 in more |
15 |
> comprehensive tests, check out the slides from the Linux Plumbers |
16 |
> Conference (particularly 14-21): |
17 |
> |
18 |
> http://linuxplumbersconf.org/2011/ocw/proposals/531 |
19 |
> |
20 |
> In summary, on those benchmarks it looks like a small global win |
21 |
> (maybe 5%) on integer calculations with a few huge wins of ≥25%, but |
22 |
> a net loss around 5% pretty much globally for floating-point |
23 |
> calculations. |
24 |
> |
25 |
> Most people probably do a lot more integer calculations unless |
26 |
> they're science geeks like me, plus it should have lower memory use, |
27 |
> so my understanding is that it probably makes sense to switch to x32 |
28 |
> no matter what you're using now (x86 or amd64). |
29 |
> |
30 |
> Mike, would you agree? |
31 |
|
32 |
But doesn't switching mean we're going to hit LFS PITA once again? |
33 |
|
34 |
-- |
35 |
Best regards, |
36 |
Michał Górny |