What If......?

[bgracia]

First off, I would like to say that I do not work for Adapteva, nor any other company and did not buy a Parallella board. I found out about it too late.

I have replied to a few threads concerning Epiphany chips, performance, availability, etc. Mr. Aolofsson, in another thread, talked about R&D costs, etc. He mentions $3.5M for R&D. With all external factors being left out and the price to buy 1 chip was $3.5M then if say 500K chips were made, the price would be say $7 each.

With this in mind, I was wondering if Adapteva had it's 1K core chips available today,

1. How many of us would but them?
2. How many would you be willing to buy?
3. What would you pay?

Personally, I would buy at least 20-25 maybe more.
I would be willing to pay in the neighborhood of 7-20 for each chip.

Brian

[bgracia]

Oh, 20-25 was just a start. If my project works out, then I would be willing to purchase several hundred at those prices. I would only be able to purchase in small quantities though.

Brian

[LamsonNguyen]

I don't really think it's a question of demand. Imho, it's rather pre-mature to start talking about 1k chips, though I can understand how it would be slightly more interesting than twiddling our thumbs while we wait.

[piotr5]

once upon a time there was such a strange economic system where you went to the bank with your fresh ideas and you did get 4m$ in a nick of time. either you would pay it back or you would go broke, the bank didn't care. they just sold your debts in a way that they would get tripple A and lived happily everafter.

truth is, economy is a showstopper for every technology, that's why we don't have dozens of batmans and gi-joes running around the world like in the comics. first you need to have the money, then you can do r&d. no matter how high the demand, the only possibility to change that is to increse the risk of many people suffering for your ruthlessly fast development plans. so in effect you need a generous donation up front or you must earn the money. if instead of donations you choose kickstarter, then you must already have something to trade for those donations, some part of r&d must already be finished.

however, if we are lucky, the previous kickstarter-boons will be distributed and enough income will be generated, just in time for a software-side support of a theoretical 1k-core parallella. then a new kickstarter would make sense. and by that time, maybe some new technology would be available, to make those better than if things would have been rushed now...

I was told that the time needed for setting up mass-production of some chip and the surrounding board is definitely over one year (even when research is finished), so I do not even expect any parallella-sale going up earlier than 1 year after the kickstarter finished. the 1k-core epiphany I expect not earlier than 3 years from now. by then we'll probably see a 16-core desktop pc too, and people will realize that lots of cores don't make for a fast computer, so there really are much more urgent problems to solve till that time...