Page 1 of 2

Parallella Epiphany IV

PostPosted: Thu Nov 20, 2014 4:54 pm
by cardosojc
When will it be available to buy?
Regards

Re: Parallella Epiphany IV

PostPosted: Fri Nov 21, 2014 10:16 am
by 9600
No date has been announced.

Regards,

Andrew

Re: Parallella Epiphany IV

PostPosted: Mon Dec 08, 2014 11:16 am
by gbonin
Hello,
I working for the French Aerospace Lab (ONERA) and I would be interested by your board with the Epiphany IV (64 cores) for multiple projects (Camera, Autopilot, Robotic, 2 axis gimbal controller, etc...)
When do you think it will be available in 2015 ?
And is it fully compatible with the 16 cores board version ?
Could you tell more about your Parallella version 2.0 ?
Thanks a lot
Best Regards.

Re: Parallella Epiphany IV

PostPosted: Mon Dec 29, 2014 1:37 pm
by jiapei100
Same question...
When and where can I buy Epiphany IV?
The news came out 2 years ago....

Cheers
Pei

Re: Parallella Epiphany IV

PostPosted: Mon Feb 09, 2015 1:58 pm
by jiapei100
Thank you ...
We are buying one....


Cheers
Pei

prosperkatula wrote:this is january, someone tell me the Parallella Epiphany IV is now available

Re: Parallella Epiphany IV

PostPosted: Tue Feb 10, 2015 2:27 am
by jimmystone
I am sorry, I do not find the Epiphany IV version on the RS website
Can anyone tell me where can I get a Epiphany IV?
Thanks.

Re: Parallella Epiphany IV

PostPosted: Tue Feb 10, 2015 10:27 am
by pmnow
Hello,

Any updates about the 64Core version of the Board?- The response is too slow & disappointed...

BR/Nowshad

Re: Parallella Epiphany IV

PostPosted: Tue Feb 10, 2015 7:01 pm
by aolofsson
No firm timetable for production release of 64-core board. The Parallella-16 is already the leading SBC in terms of cost/power/performance efficiency in some application areas. Need ~$3M to make it happen and it's unclear that it would be a good ROI for us without more software in place.

Currently our focus is on building out the software needed to make the Parallella-16 much easier to use.
If you can't wait, then I suggest you pick one of the many fine SBCs on the market today. (jetson, odroid, etc, etc)

Andreas

Re: Parallella Epiphany IV

PostPosted: Wed Feb 11, 2015 5:15 pm
by Sundance
When we (Sundance Multiprocessor Technology - our company) became an "Early Access" Partner many years ago, I had hopes it would create a revolution and a shift towards "Low-Power, Scalable and Distributed Processing", like many other old 'Transputer Nutters' hopes for.

The limited amount of local memory for each DSP makes it very hard to use for many application. I would even say MOST "Real-life" application and I think that is why Epiphany has not landed on all door-steps, but found their way into the heart of 'dreamers', like us.

As it happens, then I have 10x Epiphany III on my desk that I smile to every day and I just do not want to let the 'dream' go away.

So, I buy two lottery tickets every week; one for my (currently empty) Pension-fund and the other I write "Epiphany V" on. I honestly think that "Epiphany IV" is just more of not quite enough as having 64x DSP Cores with not-enough memory is not worth $3millions of cash.

When I win, I will send you a cheque in the post, as the Epiphany DSP is still close to the perfect Multiprocessor concept and if you can make a Epiphany DSP with 16x DSP Cores, each having, say 256kb, of local memory, then software application would be much more likely to happen that can find 'Real-Life' solutions.

Keep drumming, plesae

Flemming

Re: Parallella Epiphany IV

PostPosted: Thu Feb 12, 2015 12:45 pm
by piotr5
Sundance wrote:The limited amount of local memory for each DSP makes it very hard to use for many application. I would even say MOST "Real-life" application and I think that is why Epiphany has not landed on all door-steps, but found their way into the heart of 'dreamers', like us.

tell me, what application did you have in mind which would require more than 32k, for example requiring 256k?
if 256k is enough, did you try executing code located on a neighbour core? is it really that slow?

if memory really is preventing epiphany landing at every door-step, then why did such computers like c64 zxspectrum and so on land on so many door-steps back in the eighties? did back then programmers exist which do not exist anymore? for the c64 literally thousands of programs were created, it was a gaming platform. if you are a dreamer and you want your dream come true, why not turn parallella into a new gaming platform with optional programming&electronics-training? if memory really is such a big constraint, did you actually read the article Efficient code placement management for Epiphany? are your programs full of functions/procedures which take up more than say 16kb? do you really create objects which monolithically store data of size>16k, instead of being full of pointers? it's not better hardware parallella needs, it's better programs, better compilers!

btw, imho the real reason why epiphany isn't useful yet is because of the speed constraint. only 2x4GB/s throughput? that's somehow insufficient for fibreoptical applications. add to that the constraints of fpga memory access on parallella. now imagine cern replacing their servers for parallella...

as for EpiphanyIV, maybe it would help to put a countdown on the site displaying how much money is needed for that? I think it would be a good feeling for customers when they get the impression their purchase influenced that contdown. do it like a kickstarter site, estimate generously some amount, and once that amount is reached put up an actual timer counting down for the release...