[Pythonmac-SIG] intel compiler

Dan White dan at chalkie.org.uk
Mon Aug 28 10:21:17 CEST 2006


>
> Could someone who is using the Intel compiler say something about  
> the performance of resulting binaries compared to binaries created  
> by GCC. I'm obviously mostly interested in the speed of the python  
> interpreter.
>
> Ronald
>

for C++ code which is doing recursive / parallelisable tasks
with processor optimisation flag -xP (for an intel core duo chip)
it is faster than gcc4 with optimisation for auto vectorisation and  
other speed optimisations

At least in my case. Of course mileage might vary...
I suggest its at least worth running some benchmarks
so wee what differece it makes for various kinds of tasks in python.

I suppose what you are really getting with the intel compiler is  
their know how on how best to optimise for their chips,
in this case core duo...
so you benefit most where there are vectorisable loops etc, which can  
be thrown through the SSE processor extensions.
Maybe other benefits too.....
just have to test i suppose...

Dan


Dr. Daniel James White BSc. (Hons.) PhD
Bioimaging Coordinator
Nanoscience Centre and Department of Biological and Environmental  
Sciences
Division of Molecular Recognition
Ambiotica C242
PO Box 35
University of Jyväskylä
Jyväskylä
FIN 40014
Finland

+358 14 260 4183 (work)
+358 468102840 (mobile)
http://www.bioimagexd.org
http://www.chalkie.org.uk
dan at chalkie.org.uk
white at cc.jyu.fi




More information about the Pythonmac-SIG mailing list