Gentoo Archives: gentoo-user

From: Michael Mol <mikemol@×××××.com>
To: gentoo-user@l.g.o
Subject: Re: [gentoo-user] Changing compilers
Date: Mon, 19 Mar 2012 14:25:32
Message-Id: CA+czFiDjMme-Lb6a0GWx9Uvt1=VMx_3XCj2dOUTc=5OPJm=iQA@mail.gmail.com
In Reply to: Re: [gentoo-user] Changing compilers by Andrew Lowe
1 On Mon, Mar 19, 2012 at 10:05 AM, Andrew Lowe <agl@×××××××.au> wrote:
2 > On 03/19/12 20:34, Mark Knecht wrote:
3 >> On Mon, Mar 19, 2012 at 5:32 AM, Mark Knecht <markknecht@×××××.com> wrote:
4 >>> On Sun, Mar 18, 2012 at 10:26 PM, Andrew Lowe <agl@×××××××.au> wrote:
5 >>>> Hi all,
6 >>>>    Has anyone played around with the various "better known" compilers on
7 >>>> Gentoo? By "better known", I'm referring to gcc, Intel, llvm, pathscale. My
8 >>>> situation is that I've just started my PhD which requires me to do Finite
9 >>>> Element Analysis, FEA, and Computational Fluid Dynamics, CFD, and I want to
10 >>>> find the "best" compiler for the job. Before anyone says "Why bother, XXX
11 >>>> compiler is only 1 - 2% faster than gcc", in the context of the work I'm
12 >>>> doing this 1 - 2% IS important.
13 >>>>
14 >>>> What I'm looking for is any feedback people may have on ability to compile
15 >>>> the Gentoo environment, the ability to change compilers easily, gcc-config
16 >>>> or flags in make.conf, as to whether the compiler/linker can use the
17 >>>> libraries as compiled by gcc on a "standard" gentoo install and so on.
18 >>>> Obviously there is much web trawling to be done to find what other people
19 >>>> are saying as well.
20 >>>>
21 >>>> Any thoughts, greatly appreciated,
22 >>>>       Andrew Lowe
23 >>>>
24 >>>>
25 >>>
26 >>> Think CUDA
27 >>>
28 >>> Mark
29 >>
30 >> Sorry. Meant to include this reference: <$15 on Kindle. Reads great on
31 >> Kindle for PC.
32 >>
33 >> http://www.amazon.com/CUDA-Example-Introduction-General-Purpose-ebook/dp/B003VYBOSE/ref=sr_1_4?ie=UTF8&qid=1332160431&sr=8-4
34 >>
35 >>
36 >
37 >        I'm sorry but I'm doing a PhD, not creating a career in Academia. The
38 > concept of writing an FEA or CFD from scratch, with CUDA is laughable, I
39 > just don't have the time to learn CUDA, research the field, small
40 > displacement, large displacement, dynamics, material nonlinearities,
41 > write the code, and then most importantly benchmark it to make sure it's
42 > actually correct. This is all bearing in mind that I have 20+ years
43 > experience as a C/C++ technical software developer, including FEA and
44 > CFD. I'll actually be using Code Aster, an open source FEA code that
45 > runs under Linux.
46 >
47 >        Sorry if I sound narky, but compilers is the subject at hand, not how
48 > to write FEA code.
49
50 If you really care about a 1-2% difference, you should not be
51 dismissing GPGPU-accelerated code so easily! If the tools you seem to
52 have already settled on don't support it, you should either use
53 different tools, or correct the ones you're working with.
54
55 The lead Python guy had an astute observation (which I'll generalize)
56 the other day; for 99% of your program, it doesn't matter what
57 programming language you use. For the 1% where you need speed, you
58 should call out into the faster language.
59
60 --
61 :wq