1 |
Scott Tiret wrote: |
2 |
> On Thu, 2005-10-20 at 20:07 -0400, Sean wrote: |
3 |
> |
4 |
>>I have a dual opteron here and I am thinking of putting Gentoo on it. I |
5 |
>>am trying to decide to go with either the amd64 or i386 version. |
6 |
>>So I am asking some Gentoo amd64 users, are you happy with the version |
7 |
>>or would you have gained more with i386? |
8 |
>>Do most applications work on amd64 or are there some important ones missing? |
9 |
> |
10 |
> |
11 |
> I have been running an x86_64 (amd64) system for a few months now. The |
12 |
> only thing I have been missing is a 64bit version of Macromedia |
13 |
> Shockwave plugin. Apparently, there is no 64bit version for this |
14 |
> proprietary software. |
15 |
> |
16 |
> Everything else is fine. I have all I need on my desktop. x86_64 |
17 |
> version of Openoffice-bin (rc3) takes a long time to open, but is |
18 |
> promising. |
19 |
> |
20 |
> Good luck, |
21 |
> |
22 |
I thought the email might be a good place to ask for some ideas: |
23 |
|
24 |
I don't want to start a 64bit vs 32 bit war, or a Windows versus *nix |
25 |
war, but it has been my experience so far that the fastest benchmarks |
26 |
for a highly computation intensive program written in Numeric Python |
27 |
came on my 3.5Ghz P4 laptop with hyperthreading- on Windows. Also, |
28 |
running the same program on an AMD Opteron gave me a slower speed no |
29 |
matter what OS I was using. I performed the experiments when the |
30 |
Opteron was first introduced. I paid a high price for the fastest chip |
31 |
I could find- I don't remember the exact speed. I haven't tried the |
32 |
test lately though. Maybe it has gotten much better. |
33 |
|
34 |
Do not ask me why it happened, I have no idea. But even now, Windows+P4 |
35 |
has consistently been 3x faster in execution time than any Python on 32 |
36 |
bit *nix systems. The specific program is a Numeric Python port of the |
37 |
NEC2 EM Simulator program which calculates the Norton-Summerfield ground |
38 |
coefficients under an antenna. It makes much use of Complex-64 |
39 |
variables. I ported it from FORTRAN so I could more easily see how the |
40 |
program worked. |
41 |
|
42 |
I am baffled by the behavior. The only thing I can figure might be |
43 |
occuring would be that the *nix 64 bit toolchains are much younger than |
44 |
the 32 bit ones. But as the 32 bit Numeric Python on Windows is still |
45 |
3x faster than the *nix equivalents, I have asked Activestate, the |
46 |
Windows Python provider, if they do anything special when compiling the |
47 |
code and they say no. I think they said that they use some ordinary MS |
48 |
comiler. |
49 |
|
50 |
Any ideas would help me to put to rest the problem. I say it is a |
51 |
problem as I really don't want to boot into Windows XP to run scientific |
52 |
programs in Numeric Python. |
53 |
|
54 |
Thanks, |
55 |
|
56 |
Rob. |
57 |
-- |
58 |
gentoo-user@g.o mailing list |