3 ms·
It really isn't that hard to pivot. It's worth saying that if you were already writing OpenMP and MPI code then learning CUDA wasn't particularly difficult to g
by physicsguy 1y ago
It really isn't that hard to pivot. It's worth saying that if you were already writing OpenMP and MPI code then learning CUDA wasn't particularly difficult to get started, and learning to write more performant CUDA code would also help you write faster CPU bound code. It's an evolution of existing models of compute, not a revolution.
- Q6T46nT668w6i3m 1y agoI agree that “learning CUDA wasn’t particularly difficult to get started,” there are Grand Canyon sized chasms between CUDA and its alternatives when attempting to crank performance.
- j45 1y agoIt will improve for sure but this shouldn’t be downplayed.
- physicsguy 1y agoWell, I think to a degree that depends what you're targeting. Single socket 8 core CPU? Yes. If you spent some time playing with trying to eke out performance on Xeon Phi and have done NUMA-aware code for multi socket boards and optimising for the memory hierarchy of L1/L2/L3 then it really isn't that different.