News for hackers. Stuff that matters.

nelsonje

Re:Go hits the concurrency nail on the head on Thursday October 04, 2018 @08:35PM (#18144800)
This is the canonical Cilk paper: http://supertech.csail.mit.edu/papers/cilk5.pdf
I asked some of the authors about this---it sounded like the primary reason was that, in the FPGAs they were considering, the Altera PCIe endpoint was much better behaved than the Xilinx one (due to excessive link training time, IIRC).