Epiphany Questions

Forum for anything not suitable for the other forums.

Epiphany Questions

Postby louloizides » Sun May 03, 2015 4:14 pm

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.

























Core Grid Side SizeMatrix Side SizeARM Naive Time (s)Epiphany Time (s)Speedup vs. ARM
11000.130.320.42
12001.042.540.41
14008.3120.30.41
160027.9867.690.41
180066.39159.990.41
11000129.72317.20.41
21000.130.081.64
22001.040.651.61
24008.315.181.6
260028.0317.231.63
280066.4341.421.6
21000129.6380.91.6
3990.130.043.1
32101.210.383.23
33998.242.453.36
360028.038.043.49
379865.919.63.36
3990125.7335.433.55
41000.140.026.45
42001.040.176.06
44008.311.386.04
460028.034.546.18
480066.3511.016.03
41000129.7321.56.03
louloizides
 
Posts: 26
Joined: Fri Mar 20, 2015 11:53 am

Re: Epiphany Questions

Postby sebraa » Mon May 04, 2015 11:17 am

louloizides wrote: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?
It is not the Epiphany which gets too hot, it is the Zynq chip.

louloizides wrote:2) Why does an MIMD processor use less power than an SIMD one?
Because the cores are different. Multiple simpler processors may use less power than a single complicated one...

louloizides wrote: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?
This is the theoretical maximum obtainable by the Epiphany chip. As far as I understand, the Epiphany is clocked slower than it could be (600 MHz) and the FPGA logic is limiting the bandwidth between ARM and Epiphany as well.
sebraa
 
Posts: 495
Joined: Mon Jul 21, 2014 7:54 pm

Re: Epiphany Questions

Postby 9600 » Mon May 04, 2015 11:35 am

louloizides wrote: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?


As sebraa says, it's the Zynq that generates most of the heat and on top of which it's a tiny, densely populated board. And if you look at the Epiphany-IV that is scaled up to 64-cores, this still only consumes 2 watts.

Regards,

Andrew
Andrew Back (a.k.a. 9600 / carrierdetect)
User avatar
9600
 
Posts: 997
Joined: Mon Dec 17, 2012 3:25 am

Re: Epiphany Questions

Postby louloizides » Mon May 04, 2015 5:00 pm

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?
louloizides
 
Posts: 26
Joined: Fri Mar 20, 2015 11:53 am

Re: Epiphany Questions

Postby 9600 » Mon May 04, 2015 5:50 pm

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
Andrew Back (a.k.a. 9600 / carrierdetect)
User avatar
9600
 
Posts: 997
Joined: Mon Dec 17, 2012 3:25 am

Re: Epiphany Questions

Postby aolofsson » Tue May 05, 2015 1:46 pm

louloizides wrote: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.


..great questions.

1.) This is physics and has to do with the size of the board. For a board the size of a credit card, the limit is about 3W. This is also why your phone and tablet gets hot. Devices with boxes have a larger surface area and if designed correctly draw the heat out of the board/chip into the box. The bare board with a tiny heatsink is a bad scenario.

2.) The SIMD processor has a lot of stuff on it (FPGA, ethernet, buses, etc). SIMD will generall be more efficient than MIMD given the same ISA. However, note that the Epiphany ISA/arch is a lot smaller than the ARM A9 core..thus more efficient, even though it's not SIMD.

3.) The Epiphany chip has 4 2GB/s links (north east west south). Only one of them is connected to the Zynq chip so the peak theoretical is 2GB/s. With the chip only working at 600MHz, this brings it down to 1.6GB/s. With the current sub optimal implementation, it's much less than that. This is being remedied as we speak.. http://github.com/parallella/oh

Andreas
User avatar
aolofsson
 
Posts: 1005
Joined: Tue Dec 11, 2012 6:59 pm
Location: Lexington, Massachusetts,USA

Re: Epiphany Questions

Postby louloizides » Tue May 05, 2015 8:56 pm

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
louloizides
 
Posts: 26
Joined: Fri Mar 20, 2015 11:53 am

Re: Epiphany Questions

Postby aolofsson » Thu May 07, 2015 7:37 pm

Yes, this is an FPGA update (ie, new bit stream for SD card image)...May 30th release date.
(believe it when you see it. :D )
Andreas
User avatar
aolofsson
 
Posts: 1005
Joined: Tue Dec 11, 2012 6:59 pm
Location: Lexington, Massachusetts,USA

Re: Epiphany Questions

Postby piotr5 » Mon May 11, 2015 10:47 am

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...
piotr5
 
Posts: 230
Joined: Sun Dec 23, 2012 2:48 pm

Re: Epiphany Questions

Postby louloizides » Mon May 11, 2015 11:18 am

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.
louloizides
 
Posts: 26
Joined: Fri Mar 20, 2015 11:53 am

Next

Return to General Discussion

Who is online

Users browsing this forum: No registered users and 2 guests