>Dual-core processors widely available since 2005
>Twelve years later, consumer chips are pushing double digit core counts
>Parallel computing is still somehow in its infancy
How does that happen? You'd think that a decade and some change would see new developments there, but it seems that there's more focus on simple multitasking.
Secondary question: with ARM and big.LITTLE, can this change?
99% of what consumers do would show no appreciable benefit from parallelization
anyone who can actually benefit is already relegated to a handful of govt funded HPC centers
>have two or more applications open at the same time
You are now parallel computing sir.
>>62485106
What of enthusiasts?
>>62485131
The word you're looking for is multitasking.
>>62485153
the applications of parallel computing are overwhelmingly academic so I'd love to see these enthusiasts in 2017
>>62485063
>anime
hide
>>62485063
try to program an efficient way to use multiple cores
>>62485826
Isn't that what Rust is for, more or less?
>>62485063
>Parallel computing is still somehow in it's infancy
This is complete nonsense, just look at the scaling nwchem has at 60000+ cores and tell me that line again with a straight face.
You seem to be stuck in the mindset that just because it's incredibly difficult to program games that have decent scaling with core count that it can't and hasn't been done elsewhere.
>>62485826
It's actually quite easy. Mind you, I work with software that simulates chemical systems, so i wouldn't know the challenges of making a video run across multiple nodes.
>>62487763
no, forth and erlang, but not rust especially. It's good, but algorithms still need work. Parallelization is a fundamentally hard problem to untangle.
>>62485063
Parallel computing is as good as it gets now. Majority of stuff is simply inherently sequential.
>>62485826
async.