Memory usage in Julia v1.9.0-beta2

General Usage
1

This caught my attention, as I had problems with the 1.8 series. However, I’ve found that 1.9 consumes about 5 x the memory for my use case, which went up from 1.7G (1.8.3) to 10.3G in 1.9.0-beta2 . In turn v1.8.3 is worse than v1.7.3, but not nearly as much as v1.8.0 was. (see [Resource usage in versions 1.7.2 and 1.8.0](https:// Resource usage in versions 1.7.2 and 1.8.0 ))
The tested code is FMA/fullstate-amp-1.jl at master · rs1909/FMA · GitHub , which is part of GitHub - rs1909/FMA: Foliations, Manifolds and Autoencoders

1 Like
2

How do you measure this? “Consumes memory” is a bit imprecise.

1 Like
3

I use the ‘top’ command on Linux and read off the RES value, that is “anything occupying physical memory” as per the man page of ‘top’.

4

Is it possible that the issue is related to the number OpenBLAS threads?
At some times the default changed.

5

Don’t know if related but if one looks at the size of the new DLL some of them are BIG. Biggest one I saw had ~140 MB.

6

v1.9.0-beta2 has 1 thread, v1.8.4 has 8 threads. There is certainly a difference, but the effect should be the opposite? It is on a ryzen 3700X processor.

~/julia-1.9.0-beta2/bin/julia 
               _
   _       _ _(_)_     |  Documentation: https://docs.julialang.org
  (_)     | (_) (_)    |
   _ _   _| |_  __ _   |  Type "?" for help, "]?" for Pkg help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 1.9.0-beta2 (2022-12-29)
 _/ |\__'_|_|_|\__'_|  |  Official https://julialang.org/ release
|__/                   |

julia> using LinearAlgebra

julia> BLAS.get_num_threads()
1

~/julia-1.8.4/bin/julia 
               _
   _       _ _(_)_     |  Documentation: https://docs.julialang.org
  (_)     | (_) (_)    |
   _ _   _| |_  __ _   |  Type "?" for help, "]?" for Pkg help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 1.8.4 (2022-12-23)
 _/ |\__'_|_|_|\__'_|  |  Official https://julialang.org/ release
|__/                   |

julia> using LinearAlgebra

julia> BLAS.get_num_threads()
8
1 Like