Epiphany Questions

[louloizides]

I wrote a SUMMA implementation for the Parallella for a term project. Trying to write a report and I have a few general inquires about the Epiphany... if anyone can answer any of these or point me in the write direction I'd appreciate the insight.

1) Why does it get so hot? I can't seem to use passive cooling on the Parallella... Is it the SRAM? And will that hurt the chip's ability to scale up?
2) Why does an MIMD processor use less power than an SIMD one?
3) From most of the benchmarks I've seen here, reading and writing to shared memory is about 150 MB/s. But the off-chip bandwidth is supposed to be 8 GB/s. Is the slow speed because of the FPGA? Could the board be redesigned to have some dedicated shared memory closer to the chip? (something like an L2 cache). My SUMMA implementation is scalable, so that memory transfer to/from shared DRAM is really what's killing the speed.

FYI, these are basically the results I'm getting for the paper... I'd imagine this could be improved quite a bit but it's good enough for a term paper for a single class (not a thesis or anything). All Epiphany results were validated against the naive version.

[sebraa]

[9600]

[louloizides]

Thanks for the info-

The reason I had thought the temperature was high was because when I first tried to run some hello world examples on the Parallella several of the cores weren't running correctly (I posted an earlier message on the board about this) even though I have a newer version with a larger heatsink. After about a week of frustration I was able to solve the problem by placing a fan over the unit... so I assumed the cores were overheating.

Do you believe it was the Zync overheating that could have caused the problem then?

[9600]

If you're putting the Epiphany cores to use they will also be generating heat and so it's difficult to say, yes it's definitely the Zynq that pushed the temperature over the edge! In any case, you should really always use a fan. You can sometimes get away without one — with "light" workloads and the board on its side — but it's just not worth the trouble and if you're experiencing issues, the first things to check are the power supply and cooling.

Regards,

Andrew

[aolofsson]

[louloizides]

Hi Andreas,

So will the newer more optimal implementation just be a software update? If so that'd be awesome.

This last project I did was just a simple term project since I wanted to try the board out. But I have an AI application in mind for a thesis that it could be useful for. I need an embedded device that could search through a decent sized data set & search space in a short period of time (particularly genetic searches). I was hoping I might be able to use the Parallella, but right now the off-memory access is probably too slow. If a software update improves that at some point it'd be awesome. Otherwise for future revs it'd be great if there was some decent bank of memory (even a couple of MB or so) close to the chip and accessible using DMA. It'd make a *huge* difference. The Epiphany really flies without that one bottleneck

Thanks,

Lou

[aolofsson]

Yes, this is an FPGA update (ie, new bit stream for SD card image)...May 30th release date.
(believe it when you see it. )
Andreas

[piotr5]

as I understood, the reason why genetic AI algorithms are popular is because it can be implemented nicely in gpu. have you considered using some other AI method for parallella? for example I was taught an AI for mathematical equations is basically just representing data as a graph, even though the implementation then transforms the graph into a matrix so it can be tackled by the usual mathematical operations...

[louloizides]

Actually what I'm working on is a pseudo-genetic algorithm, designed to search a space for a specific problem.

A GPU is designed to do the same, simple operation on a large data set. So I'm guessing it could be faster if you generated a lot of children at each step. But after you generate children you also have to know which child is the best, which generally means some kind of validation and test on the data. So there is a need to switch between operations. That's why I tend to think that if you want to generate less children and more iterations, instead of more children and less iterations, the Parallella might be a better processor for it.

But honestly I really don't know. It would make for an interesting benchmark. But until the Epiphany has access to faster shared memory the Parallella will probably end up being slower - unless the data set the algorithm is being applied to can fit in the 32kB.