1 |
On 03/20/12 00:03, Mark Knecht wrote: |
2 |
> On Mon, Mar 19, 2012 at 7:05 AM, Andrew Lowe <agl@×××××××.au> wrote: |
3 |
>> On 03/19/12 20:34, Mark Knecht wrote: |
4 |
>>> On Mon, Mar 19, 2012 at 5:32 AM, Mark Knecht <markknecht@×××××.com> wrote: |
5 |
>>>> On Sun, Mar 18, 2012 at 10:26 PM, Andrew Lowe <agl@×××××××.au> wrote: |
6 |
>>>>> Hi all, |
7 |
>>>>> Has anyone played around with the various "better known" compilers on |
8 |
>>>>> Gentoo? By "better known", I'm referring to gcc, Intel, llvm, pathscale. My |
9 |
>>>>> situation is that I've just started my PhD which requires me to do Finite |
10 |
>>>>> Element Analysis, FEA, and Computational Fluid Dynamics, CFD, and I want to |
11 |
>>>>> find the "best" compiler for the job. Before anyone says "Why bother, XXX |
12 |
>>>>> compiler is only 1 - 2% faster than gcc", in the context of the work I'm |
13 |
>>>>> doing this 1 - 2% IS important. |
14 |
>>>>> |
15 |
>>>>> What I'm looking for is any feedback people may have on ability to compile |
16 |
>>>>> the Gentoo environment, the ability to change compilers easily, gcc-config |
17 |
>>>>> or flags in make.conf, as to whether the compiler/linker can use the |
18 |
>>>>> libraries as compiled by gcc on a "standard" gentoo install and so on. |
19 |
>>>>> Obviously there is much web trawling to be done to find what other people |
20 |
>>>>> are saying as well. |
21 |
>>>>> |
22 |
>>>>> Any thoughts, greatly appreciated, |
23 |
>>>>> Andrew Lowe |
24 |
>>>>> |
25 |
>>>>> |
26 |
>>>> |
27 |
>>>> Think CUDA |
28 |
>>>> |
29 |
>>>> Mark |
30 |
>>> |
31 |
>>> Sorry. Meant to include this reference: <$15 on Kindle. Reads great on |
32 |
>>> Kindle for PC. |
33 |
>>> |
34 |
>>> http://www.amazon.com/CUDA-Example-Introduction-General-Purpose-ebook/dp/B003VYBOSE/ref=sr_1_4?ie=UTF8&qid=1332160431&sr=8-4 |
35 |
>>> |
36 |
>>> |
37 |
>> |
38 |
>> I'm sorry but I'm doing a PhD, not creating a career in Academia. The |
39 |
>> concept of writing an FEA or CFD from scratch, with CUDA is laughable, I |
40 |
>> just don't have the time to learn CUDA, research the field, small |
41 |
>> displacement, large displacement, dynamics, material nonlinearities, |
42 |
>> write the code, and then most importantly benchmark it to make sure it's |
43 |
>> actually correct. This is all bearing in mind that I have 20+ years |
44 |
>> experience as a C/C++ technical software developer, including FEA and |
45 |
>> CFD. I'll actually be using Code Aster, an open source FEA code that |
46 |
>> runs under Linux. |
47 |
>> |
48 |
>> Sorry if I sound narky, but compilers is the subject at hand, not how |
49 |
>> to write FEA code. |
50 |
>> |
51 |
>> Anyway, thanks for answering, |
52 |
>> |
53 |
>> Andrew |
54 |
>> |
55 |
> |
56 |
> Nahh, be as snarky as you like as long as you don't really mean it personally. |
57 |
> |
58 |
> My experience with CUDA, and I'm not a programmer, is that there is a |
59 |
> fairly steep learning curve. However changing C compilers will get you |
60 |
> maybe 5%. Changing to CUDA will get you 30,000%, assuming a mid-high |
61 |
> range CUDA card and that you can parallel-ize FEA. I did a little |
62 |
> Googling and it seems that FEA is a pretty common CUDA topic so I |
63 |
> don't think at the outset that you'd find yourself all alone. |
64 |
> |
65 |
> Good luck whatever you do and know that I didn't mind the response at all! :-) |
66 |
> |
67 |
> Cheers, |
68 |
> Mark |
69 |
> |
70 |
> |
71 |
|
72 |
The thing is that I agree that CUDA is the way to go for things like |
73 |
FEA & CFD, in fact the mob that runs the super computer I'm using is |
74 |
installing another one that is top heavy in CUDA cards. But the thing is |
75 |
as I'm using the FEA as a tool, rather than playing around with the |
76 |
innards of the code, I need an established code, one that has |
77 |
verification behind it. My topic looks at the way that steel connections |
78 |
behave so I need an established FEA code that is verified to provide the |
79 |
correct answers, I don't get that if I write my own code. |
80 |
|
81 |
Most likely I'll write a short paper covering a comparison of existing |
82 |
C/C++ compilers and their relative speeds, spend the next 18 months - 2 |
83 |
years doing my research and then to close things off, I'll probably be |
84 |
able to write another short paper concerning CUDA speedups as the FEA |
85 |
code bases will have caught up and been verified. |
86 |
|
87 |
Thanks for peoples replies, |
88 |
|
89 |
Andrew |
90 |
|
91 |
p.s. Writing this I just had a sudden horrific though and checked the |
92 |
FEA code I'm using, Aster. It's written mostly in FORTRAN - fat chance |
93 |
I'm going to be hacking that code........ |