I’m trying to populate a 3D array (predetermined size of about a million) with values of a
Float64 function on a cartesian grid.
I think the term for this is “discretely sampled volumetric data” in scientific visualization.
Anyway, I wrote 3 different ways to do it in Julia and measured the performance with
@time but all of them were less than half the speed of an equivalent C code (
I fiddled with my Julia code just enough to learn that some allocations need to be pulled out of the loop, that I need to be careful how I select a range of elements while avoiding slowdowns, etc. I basically checked through all of the points in Docs » Performance Tips that apply to me, but the speed isn’t that satisfying.
Perhaps this is the limit of Julia as far as populating an array is concerned…
Could anyone let me know if there’s a faster way? Thanks.