Eigenvalue decomposition using parallel computing -epiphany

[rajat_1397]

Hi,
I would like some help with eigenvalue decomposition of a random matrix whose computation runs parallelly on 16 cores of the epiphany chip including the data transfers between the cores. I have already written a matlab code over a single core which includes few sub functions which have to be distributed over the 16 cores to be run parallelly keeping in mind the data transfer between the cores. I haven't used the eig() function since it takes more time to compute.

[sebraa]

...what do you want to say?

[rajat_1397]

I would like the zync processor to distribute the work among the 16 epiphany processors in computing the eigenvalue of a matrix

[sebraa]

Then just do it. We won't do your homework, so if you don't ask specific questions, you won't get any answers.

[nickoppen]

One of the standard examples that is probably already on your parallella is matrix multiplication. How about you start there.

[rajat_1397]

Can you give me a sample code so I can understand as to how to access epiphany cores for computing. Thanks in advance!

[nickoppen]

It's all here:

Also, have a look on your parallella for a folder called /home/parallella/parallella-examples. Same stuff, just local.