When we launched the “$99 Parallella project” on Kickstarter in 2012, the goal was to “democratize access to supercomputing”. With a $99 price point, we knew that we would not make a profit, but we figured the value of reaching the whole world with our technology was enough of a reward for Adapteva as a company. With over 8,500 boards shipped to 75 countries (and over 150 Universities), we are certainly off to a great start!
Despite the enormous interest, it’s going take years until our Epiphany chip business can stand on its own.This is just the reality of the semiconductor design-in cycle. This means that despite the generous investment from Ericsson and Carmel Ventures, we can no longer afford to subsidize the cost of the Parallella boards and will have to raise the price slightly so that everyone involved can stay motivated (including Adapteva, our US manufacture, and our distributors). As we the sales ramp up, we will work hard to reduce prices and bring to market Parallella versions at $99 and below!
The great news is that these new products will all ship with a larger heatsink and an improved power supply included so the ramp-up process is MUCH simpler. This new bundling is in response to loud and clear feedback from our Kickstater backers. “Get rid of the fan!”
The “Desktop” Parallella version is currently in stock at shop.adapteva.com and should ship via Amazon within 24-48 hours. (US customers only, international distribution starting soon). The “Microserver” and “Embedded” Parallella boards will be available in stock within a few weeks. Still not convinced that the Parallella is something special? Check out some of the open source projects below. Look forward to hearing from all of you and seeing the projects you create! Sincerely, Andreas Olofsson Founder/CEO at Adapteva @adapteva
Check out some of the demo videos/examples from early users below! Slowly but surely, the evidence is accumulating that the Parallella platform delivers!
- “r-opencl”: R Running on Parallella
- “kinect_test”: Kinect demo running with only 1 Epiphany core utilized (lots of headroom!)
- “aobench”: Raytracing demo
- “john”: John the ripper demo (25x more efficient than GPUs on Epiphany-IV)
- “Para-para”: Hello world examples for MPI, OpenMP, and OpenCL runs out of the box with Ubuntu 14.04 image
- “Lena”: 2D FFT based filtering example
All of the examples are open source and available at: http://github.com/parallella/parallella-examples