Julia call from Python3 running in single core

ihnorton · November 26, 2016, 11:19pm

Unfortunately that doesn’t seem to work via pyjulia. OP tried 0.3, and I tested several versions:

`addprocs` leads to segfault

opened 04:39AM - 24 Nov 16 UTC

closed 07:51PM - 06 Dec 16 UTC

``` import julia j = julia.Julia(jl_runtime_path="/Users/inorton/git/julia/usr…/bin/julia", jl_init_path="/Users/inorton/git/julia/usr/bin") j.addprocs(1) #segfault ``` (segfault stack [here](https://gist.github.com/ihnorton/63dabfa2316d2e29eb0941eaa70ab7ce)) Based on the stack trace, I thought there might be some issue with task switching because of ctypes (libffi) stack. However, running the following (equivalent?) code does not segfault: ``` JLPATH="/Users/inorton/git/julia" import ctypes jl = ctypes.PyDLL(JLPATH+"/usr/lib/libjulia.dylib", ctypes.RTLD_GLOBAL) jl.jl_init(JLPATH+"/usr/bin/") jl.jl_eval_string(""" addprocs(1) """) jl.jl_eval_string(""" println(nprocs()) """) # 2 ``` Tested with 0.4.6, 0.5, and a several-day-old trunk build. (cross-ref: https://discourse.julialang.org/t/julia-call-from-python3-running-in-single-core/508)

Topic		Replies	Views
Pyjulia in desperate need of attention form someone who knows what they're doing General Usage	22	5817	November 24, 2017
Construct and intermittently call a single Julia instance with JuliaCall Tooling python , juliacall , pyjulia	4	614	April 14, 2023
Questions about pyjulia usage General Usage pyjulia	6	1927	November 11, 2021
Baffling addprocs() with @everywhere Julia at Scale	17	3492	April 24, 2018
PyTorch and Julia Machine Learning	12	15407	March 27, 2019

Julia call from Python3 running in single core

Related topics