Does not GPU compute

Re:Does not GPU compute by vessenes on Thursday June 02, 2016 @07:18PM (#11826982)

Even if you're doing fintech simulations, FP16 could well be plenty of precision for a first pass, and then you'd get all those extra cores and ops/watt.

FP64 seems like a very small use case for most of the parallelized workflows I can imagine.

[ Parent | on HN ]

Re:Does not GPU compute by Robi395 on Friday June 03, 2016 @04:14AM (#11828999)

For scientific workloads, double precision is a must have. The 7 digits of FP32 is not enough. In my lab, we haven't updated our Kepler based GPU since 2013 for this reason.

[ Parent | on HN ]

Re:Does not GPU compute by bnegreve on Friday June 03, 2016 @05:00AM (#11829136)

I agree but the article has an answer to this.

Since I don’t plan to run any DNA analysis or fintech simulations on my smartphone anytime soon, I am very satisfied having FP32/FP16 precision in mobile right now. And so should you.

[ Parent | on HN ]

Re:Does not GPU compute by Robi395 on Friday June 03, 2016 @05:08AM (#11829159)

You are right, I thought vessenes meant: "FP64 seems like a very small use case for most of the parallelized workflows I can imagine" for any platform.

[ Parent | on HN ]

Re:Does not GPU compute by vessenes on Sunday June 05, 2016 @03:38AM (#11840023)

That makes sense. I'm assuming that the majority of GPUs are not running scientific workloads, instead deep learning and matrix ops for finance.

[ Parent | on HN ]