Re:Are functional languages inherently slow?
by wonderzombie on Wednesday August 01, 2012 @05:34PM
(#4325879)
Not sure if anybody reads comments five days later, but thanks for the suggestion and the offer! I appreciate it.
Why do you suppose it's so hard to publish about them? A general lack of interest?
Having spent a large portion of the last five years working on a parallel concurrent collector that scales to 48-AMD and 64-Intel cores, I can say the game's totally different on NUMA architectures. Unfortunately, not much has been written yet on it (I have an MSPC paper; the Intel folks have a paper on Pillar that was at ISMM; that's about it). Feel free to ping me directly if you get to advanced material and need more pointers. Many of the systems have not yet been published, as it's shockingly difficult to publish on modern GC implementations these days.