Jaxley runtime compared to NEURON

[VENOM314]

In the original Jaxley paper (https://www.biorxiv.org/content/10.1101/2024.08.21.608979v2.abstract) it was mentioned that Jaxley was similar to NEURON in speed on CPU. I was wondering if it is true that after Jaxley V0.9.0 and V0.11.2 Jaxley is faster than NEURON on both CPU (since V0.9.0 has a 20% runtime speedup and 50x compile time speedup) and GPU (V0.11.2 fixed a bug in GPU-based network solver following V0.9.0). Is the reasoning correct?

[michaeldeistler]

Hi there,

yes, I would expect that v0.11.2 is a bit (~20%) faster than NEURON on CPU. The paper did not compare compile time, so the 50x speed-up there does not have an effect on the numbers shown in the paper.

GPU should also be a bit faster now as compared to what is reported in the paper.

Michael