I have been working on some performance improvements to a model. By accident, I started two runs of the same model, and noticed hugely different performance. One finished in 7 hours, the other took 19 hours. As far as I can tell these should have been identical, and am looking for other sources of …

2.5x speed difference in runs of the identical model

wds15 May 4, 2018, 11:11am 6

Is your cluster configured right? Were other processes competing with Stan?

Topic		Replies	Views
Inconsistency when measuring computational efficiency Modeling performance	5	574	January 22, 2020
Runtime difference between cmdstan v2.28.2 and v2.27.0 CmdStan	7	958	December 10, 2021
Huge contrast in sampling time among chains for the same model? RStan	4	806	November 5, 2018
Why don't the execution time add up to real elapsed time General stanc	5	1112	August 22, 2017
Reproducibility across machines - hardware/C++ complier dependent? Modeling	2	470	June 29, 2019