using pdsyev(d)
Hello, I new to scalapack and have some basic questions.
From the example at the website below I am trying to understand how to properly make a call to the PDSYEV/PDSYEVD routine.
http://www.netlib.org/scalapack/examples/sample_pdsyev_call.f
I compactified the code a bit and modified it to be a able to use different number of processors as well as larger matrices.
Everything seems fine. The eigenvalues are correct etc. Except, I don't find any speed-up in using a larger number of processors.
So I am wondering if the calculation is done in parallel at all.
Here is my code
http://pastecode.org/index.php/view/9070020
Can someone tell if this is actually distributing the matrix?
I expect this to be done by the BLACS calls to set-up the grid.
Thanks
From the example at the website below I am trying to understand how to properly make a call to the PDSYEV/PDSYEVD routine.
http://www.netlib.org/scalapack/examples/sample_pdsyev_call.f
I compactified the code a bit and modified it to be a able to use different number of processors as well as larger matrices.
Everything seems fine. The eigenvalues are correct etc. Except, I don't find any speed-up in using a larger number of processors.
So I am wondering if the calculation is done in parallel at all.
Here is my code
http://pastecode.org/index.php/view/9070020
Can someone tell if this is actually distributing the matrix?
I expect this to be done by the BLACS calls to set-up the grid.
Thanks