I’m trying to fit a simple MvNormal model. Initially, I’m using a little synthetic dataset to make sure I have things correct. Here’s the test model, where I have a two-dimensional dataset with 1,000 rows: import numpy as np import scipy as sp import pymc as pm import arviz as az mvn = sp.stats.m…

Yes, thanks for pointing me there. I didn’t turn it up as I was initially searching around. I’m guessing it’s conflicting processes/threads.

Slow MvNormal sampling with multiple cores

ricardoV94 July 21, 2023, 2:59pm 2

Topic		Replies	Views
Sampling hangs with multiple cores Questions	5	4252	May 21, 2020
Sampling gets stuck with more than one core Questions	5	1122	May 30, 2020
Increasing sampling speed using multiple cores Questions	0	392	March 2, 2021
Pm.sample gets stuck after init with cores > 1 Questions	17	4113	January 4, 2021
Is multivariate inference always so slow? version agnostic	9	1159	July 7, 2023