The following code comes from a real-world example with a beginner Julia-programmer calculating the position vector for a scalar field. The second version with the array replaced by a tuple is about 10x faster. This is because in each loop iteration the array causes a memory allocation, which is avo…

The TLDR is that it should, and that there is current work in the compiler on escape analysis that facilitates this type of operation.

Should Julia be able to optimize away small temporary arrays?

Oscar_Smith July 23, 2022, 12:44am 8

We’re getting there. 1.8 added the effect system, and 1.9 will likely have some form of escape analysis.

5 Likes

Topic		Replies	Views
Strange performance of a loop Performance	37	3327	July 21, 2018
Memory allocations when returning vectors General Usage array , memory-allocation	15	1546	June 6, 2018
Why is Julia's generated assembly more complicated than Swift's for a simple function? Performance	16	2579	March 6, 2020
Vectors vs. Tuples and multiple dispatch New to Julia performance	3	961	April 14, 2021
Fortran vs Julia stack allocated arrays Performance array , memory-allocation	18	6599	November 29, 2018