Page 2 of 2

Re: Harris corner detector and openCL

PostPosted: Thu Apr 16, 2015 12:25 pm
by Nader_Jendoubi
Do you think that is a memory problem i think the epiphany memory is not enough what do you think??

Re: Harris corner detector and openCL

PostPosted: Fri Apr 17, 2015 1:17 am
by nickoppen
At this point we have not come anywhere near the epiphany. clcc operates entirely on the ARM host cores.

Get an executable and then figure out if the algorithm works. I assume that you need to pass the routine an image. Find one that is around 20k (i.e. would easily fit onto one core) and run it. That will eliminate any concern about the epiphanies memory getting in the way. Once you have the algorithm working then use bigger and bigger images to test if your partitioning is working.

The only way to operate in the absence of documentation is to solve one problem at a time.

Re: Harris corner detector and openCL

PostPosted: Fri Apr 17, 2015 2:30 pm
by Nader_Jendoubi
HI

i'm trying to optimize the sobel application https://github.com/sclukey/parallella-o ... ter/opencl

I think the transmission time between epiphany(kernel) and ARM is too long, when i took the initial sobel program and delete all the instructions and i tried gbb[i] = gaa[i]

the time measured is 0.33 second .

Is there any solution to minimize the transmission time between ARM and epiphany ???