pvlib / pvlib-benchmarks Goto Github PK

The benchmark results history is getting a bit confusing. The plots are supposed to quickly draw the eye to performance jumps, but the majority of jumps are just due to environment changes. Example: SolarPositionNumba.time_sun_rise_set_transit_spa

Although changes in performance due to environment is surely interesting, I think the main goal for our benchmarks should be for evaluating performance changes in pvlib itself (holding the environment ~constant). Should we wipe out and rebuild the history using the current set of environments?

We may also want to change our environment management strategy to reduce future fragmentation.

After following the steps given in the Readme.md, I got an error in step 10 i.e. Validate and build environments using asv check.

Versions:
asv = 0.4.2
python = 3.7.1
conda = 4.10.3
virtualenv 20.14.0

@AdamRJensen suggests, and I think I agree, that it would make sense to migrate this benchmarks repository from pvlib-benchmarker/pvlib-benchmarks to pvlib/pvlib-benchmarks. I solicit thumbs oriented up or down on this idea.

It seems I can't tag the collective handle from here, so: @wholmgren @cwhanse @adriesse @mikofski @AdamRJensen