topic Hi qweasd q. in Intel® oneAPI Math Kernel Library

Inplace permutation

asd__asdqwe — Tue, 24 Mar 2015 21:50:08 GMT

Hello,

I need to permute a vector according to an index array jpvt returned by ?geqp3. In the example lapack/source/dgeqpfx.f it is done using an auxiliary array. I would like to do the transformation inplace (or at least with O(1) extra memory), but as you can see from the attached source code, I don’t seem to be able to use ?laswp correctly. Can this be used for this kind of transformation ? If yes, how ? If no, is there another way to do this ?

Thank you in advance.

Hi qweasd q.

Ying_H_Intel — Thu, 26 Mar 2015 05:49:00 GMT

Hi qweasd q.

The ?laswp swaps the row in a cascading way. So it can't do the permutation directly. you may use Array notation like A[:]=A[iptv[:]] in Intel Compiler https://software.intel.com/en-us/blogs/2010/09/03/simd-parallelism-using-array-notation

Best Regards,

Ying.

Hi qweasd q.

Alexander_K_Intel3 — Thu, 26 Mar 2015 08:57:31 GMT

Hi qweasd q.

Please try using dlapmt_, following code does the correct permutation for your example:

int _false = 0, ione = 1;
dlapmt_(&_false,&ione,&n,A,&ione,jpvt);

Best regards,

Alexander

@qweasd q.

Charls_W_ — Wed, 29 Jul 2015 10:10:24 GMT

@qweasd q.

Hello qweasd q.

I really benefited from the function ?lapmt you mentioned above. But now I'm confused with another problem.

I need to permute the 8 bits in a byte in the order specified in parameter pmt.

I wonder if there is a similar function as ?lampt, while permutation is carried out in bits . And inupt char output char would be better. Bits arranged in vector is also OK.

Thanks for you attention.