Parallella Community

Got it, yes, that sounds about right for 4 workers.

I use foreach on a cluster of Macs and utilize the GPU's through nvBlas. When I look at the parallela, it seems there is no benefit since there is no optimized BLAS utilizing the epiphany co-processor. I can see that I could possibly run 2 foreach forks on the parallela because of the main processors dual core. Is this right? Is there any benefit relative to adding another i7 at 2.6 ghz with 4 cores/8 threads?

I can afford my electric bill, but my GARCH studies take 42 hours each (if nothing gets messed up) to run on 28 intel cores and I need to run 12 of them, so wondering whether I should keep buying more i7 based Macs or could I get more bang for the buck by building a parallela cluster? Right now, w/o epiphany optimization, my guess is stay with the i7.