3 ms·
You can get that without a programming language revolution by reformulating the problem a bit. For instance 3D-API shading languages do just that under the hoo
by flohofwoe 6y ago
You can get that without a programming language revolution by reformulating the problem a bit.
For instance 3D-API shading languages do just that under the hood, just without the "parallel-for". Instead you tell a 3D-API what code should run per-vertex or per-pixel in (usually) a traditional sequential C-style language, and the GPU driver and hardware care about the parallelization and scheduling.