rixoff
Re:OpenBenchmarking.org – A Platform for Benchmarking and Performance ...
on Saturday February 20, 2016 @03:22PM (#11141775)
This site is a pile of confusion. Of all the benchmarks I'm interested, the HPC Challenge, the results are buried. No overview, and when you finally get into the results, the performances are all terrible. (Should be closer to 500GFLOPS per node, not 100!)
This also appears geared towards consumer use, which is akin to adding throttle body spacers and tall spoilers to your car... not practical applications.
Re:Teaching Deep Convolutional Neural Networks to Play Go
on Monday December 15, 2014 @07:49PM (#8755372)
Happy to see that this paper is freely available for all to download and read!
Re:Stanford Javascript Crypto Library
on Thursday August 29, 2013 @12:54PM (#6296928)
Sure! Check out crypton.io. SpiderOak is going to be shifting towards this method. Although we already use client-side cryptography, it's in fat clients written in Python. We're going to reimplement the client using Crypton, which uses SJCL.