Bitcoin Forum

Anyone had a look at this

http://www.kickstarter.com/projects/adapteva/parallella-a-supercomputer-for-everyone

Is it suited for bitcoin mining ?

for mining not,
but for rendering video/images sure mega!

Mhm, it has opencl support for the 64 risc cores, should be good for something ;)

great for ltc and sc mining ;)

Seems interesting. They are essentially cutting an AVNet ZedBoard down from $395 to $99 with their own daughter boards.

The chip on the daughter board has some nice specs:
Source: http://www.adapteva.com/products/silicon-devices/e64g401/


It doesn't look good for the kickstarter (they need $750 000 and only have $200 000 so far) - I hope it goes through.

I agree, it doesn't look so good right now. Do you think they'll get a boost from releasing their documents (which occurred today, I believe)?

Does anyone still trust kickstarter?

Since they've clarified,

"Kickstarter does not guarantee projects or investigate a creator's ability to complete their project. It is the responsibility of the project creator to complete their project as promised, and the claims of this project are theirs alone."
http://www.kickstarter.com/help/faq/kickstarter%20basics#Acco

So, there is little reason to trust anyone. Although later on the same page, they do state that the creators agree to some type of obligation to deliver rewards or offer refunds to backers.

I'm surprised by how quickly it has evolved into a purchasing-oriented rather than donation-oriented system, but people tend to prefer tangible rewards if offered, I suppose.

Would you want something back if youinvested in a company? I would, why invest if tthat doesnt happen?
If there is a possiblblty woudt happen I would be better off keepin my 100K a year income in alow intrest
savings.

They are up to $300k now.

Interestingly check out their survey responses - some people are wanting it for Bitcoin:

http://www.kickstarter.com/projects/adapteva/parallella-a-supercomputer-for-everyone/posts

800 MHz times 64 cores is 51,200,000,000 instructions per second.

SHA-256 takes about 3000 instructions per round, and has 64 rounds, so 192,000 instructions per hash.

Gives about 266 kilohashes per second, per chip.  Not great for mining.

Once multi-threaded signature verification becomes standard in the Satoshi client, one of these $99 boards (given they reach their target) might be able to verify enough transactions per second to maintain the block chain if/when transaction volumes start reaching high levels. That's pretty cool for a $99 computer.

they are near $500k and 2 more days to go!
i don't see how the $750k will be reached unless a "big fish"
awakes the last moment :D

I think if they showed more practical uses for it they would make the funding

Damn, half a million is already a lot, it would be nice if someone with a quarter of a million appeared now.  :)

Sounds like a fun supercomputer project. How well would this perform compared to say a quadcore ivy bridge processor on perfectly threadable workloads?

I was just thinking the same thing.  Kickstarter has about a year left before it starts to implode due to people losing all trust with them.  Unless they do a better job of vetting what gets up there, noone will want to use it anymore.

What did they do that would cause people not to trust them? I haven't heard about kickstarter until now, so I don't have a clue, but on the outside it sounds like a very cool way to get funding for projects like these.

I think the point is that it's mostly a research project. Ie. the device is intended to discover what people would actually use 64 cores for. Currently there aren't many processes that would be sped up with this kind of configuration. It's a chicken-and-egg problem.

now @$674k and 27 hours to go...

These are floating point instructions per second, since Bitcoin uses integers there could be some wiggle room? ???

I'm very curious to see the Cypto-Related uses of the computer.

It does not have enough RAM to use it for LTC mining (fully).

Looks like they are gonna make it.

they DO make it!!!!
i am suprized, the last day surpass the target by far  :o

LOL, WTF? Awesome they made it! I wonder who decided to put forward like $300,000 in the last, like, 48 hours.

I thought about backing it for $200 to get the 64-core one, just for giggles, but there was no way they were going to hit the $3M stretch goal for that.

@$831k and 5 hours to go  :)

I was kinda hoping this project would succeed, but I didn't have any cash to send their way.

Side note: I find it funny that no one gives two shits about sending almost a million dollars to fund a startup company for computer hardware that could not be delivered for quite some time.  ;)

+1  ^^^ This.  :D

Can't wait for my 99$ board. Beside from learning I'll use it for data mining and 24/7 stochastics  :D, might be way cheaper than a quad core vps.

Same, should be an easy co-processor to use since it has existing libraries in the SDK (and opencl support). Can't wait to see what sort of streaming encryption performance i can get out of my 99$ unit.

All things considered, coming with packages like that will be very useful, even if the board's specs aren't great.

The company's projected improvements through 2014 do begin look competitive with at least GPU performance and presumably resell is not an issue.

http://www.kickstarter.com/projects/adapteva/parallella-a-supercomputer-for-everyone

Kickstarter has some interesting policies.  You must remember that until the funding goal is reached, all those numbers are just pledges.  The total does not represent money in the bank.  If the project does not meet the funding target, no one loses any money, since they haven't spent a penny (they've only pledged funds to the organizers). 

If the project meets its funding goal, then the money is made available to the developers (less a percentage kept by Kickstarter).  Should the product fail to ship, then the investors lose their funds.  If the product ships, however late, the investors still get something for their trouble.

Same here, they could have had a lot more supporters if they could have done 64 core within a $1,500,000 budget. I wanted two at least.

I wonder how it will go cracking WPA  :P

It might actually be very good at that, even more so than other cryptographical uses.