g******n 发帖数: 339 | 1 Hi,folks:
I am running a very computationally intensive Monte-Carlo Markov Chain
algorithm using R, but it seemed extremely slow. I know if I rewrite the
program in C, the efficiency will be improved a gread deal, however, it
might be easier for me to rewrite in matlab than C. Is there a significant
difference in efficiency between C and matlab for running MCMC program?
Thanks. |
L*******g 发帖数: 913 | 2 I think it'll be faster in C.
【在 g******n 的大作中提到】 : Hi,folks: : I am running a very computationally intensive Monte-Carlo Markov Chain : algorithm using R, but it seemed extremely slow. I know if I rewrite the : program in C, the efficiency will be improved a gread deal, however, it : might be easier for me to rewrite in matlab than C. Is there a significant : difference in efficiency between C and matlab for running MCMC program? : Thanks.
|
m**********e 发帖数: 24 | 3 Unless you have exactly what is needed to be done and if you have to check
and debug your C code. Matlab is very efficient already and has plenty of
built-in functions. Getting a new CPU would help too. |
L*******g 发帖数: 913 | 4 If you want to run your code on a cluster, you definitely want to use C. |
k******n 发帖数: 35 | 5 For matrix computations intensive codes, Matlab could be as fast as C. |
s**b 发帖数: 169 | 6 agree,by the way, how fast can atlas provide gcc in general?
【在 k******n 的大作中提到】 : For matrix computations intensive codes, Matlab could be as fast as C.
|
k******n 发帖数: 35 | 7 Sometimes, ATLAS could achieve more than 10x performance than naively
compiled BLAS.
【在 s**b 的大作中提到】 : agree,by the way, how fast can atlas provide gcc in general?
|
s**b 发帖数: 169 | 8 that's great, worth take a try
【在 k******n 的大作中提到】 : Sometimes, ATLAS could achieve more than 10x performance than naively : compiled BLAS.
|
M******e 发帖数: 48 | 9 http://www.interactivesupercomputing.com/products/
Star-P software is a client-server parallel-computing platform that’s been
designed to work with multiple Very High Level Language (VHLL) client
applications such as MATLAB®, Python, or R, and has built-in tools to
expand VHLL computing capability through addition of libraries and hardware-
based accelerators.
某一天看到了,觉得如果有条件可以看看 |
s**i 发帖数: 381 | 10 my experience is that Matlab won't be as fast after the size has increased
to some very large point
【在 k******n 的大作中提到】 : For matrix computations intensive codes, Matlab could be as fast as C.
|
k******n 发帖数: 35 | 11 Is there any reason?
On some platforms, Matlab calls ATLAS. I guess your C codes also call ATLAS.
Why does it get slow?
【在 s**i 的大作中提到】 : my experience is that Matlab won't be as fast after the size has increased : to some very large point
|