Hans Moritz Günther
Hans Moritz Günther
@DougBurke I didn't 't want to push it into 4.17, but that was several months ago. Maybe you can find time to have a look?
CI Failure is unrelated to this PR, please don't let that stop you from reviewing.
Currently, two things have to be true to use the cache: - `cache > 0` - `_use_caching = True` That's the current design - we could change it and (as...
> Have you seen an improvement in run-times when using the cache with more-than-one element? Or even with the old only-one-element in the cache? Yes and yes, but it really...
I pushed a commit that simplifies the caching even further as suggested in the review by removing the private `_use_cache` attributed altogether. Before that commit, caching happened if two things...
@DougBurke This changes how some XSPEC model Python wrappers are generated - those that do not allow cashing at all. I'm not sure if we have any of those at...
I run some benchmarking and the results are inconclusive. That doesn't mean that this fix is not worthwhile, just that, as a I said above, it depends a lot on...
I mean these two lines: https://github.com/sherpa/sherpa/pull/2166/files#diff-bbd8d73e0808ca92887b8ead47a0037440164909d021f66baf0ea7cfce9fe08eL879-R879
I'll rebase later to address conflicts, because if I rebase and force-push now then the link I just posted in https://github.com/sherpa/sherpa/pull/2166#issuecomment-2657817131 would break.
Address the most stupid merge conflicts I've ever seen (white space around the numbers in a copyright line)