Benchmarks

[ukd1]

Hey @bensheldon - just found this, nice that you forked it and maintain!

So, I'm curious why Que performs so much better than QC on - I thought, based on this release it was comparable or slightly better, but the docs here suggest otherwise: but, my comment was on a laptop - these benchmarks on big machine. Could that be it? If you have any thoughts, great.

Only I think I noticed inside the benchmarks themselves was that Que seems to enqueue, and destroy in the same transaction; I'd assume this results in zero writes to tables or indexes; however QueueClassic has a comment about not knowing how to delete it - i.e. it's actually going to be writing it to disk. I'm curious how much difference that would make, and will make a PR.

[bensheldon]

Yay 🙌

fyi, I haven't updated any of the backends to their latest versions, so the numbers are likely not representative of their contemporary performance.

I'm now using Github Actions as the "benchmark" system: https://github.com/bensheldon/queue-shootout/actions/workflows/shootout.yml

My own goals here are not to make any assertions about which library is better so much as have a test-bench: less shootout, more firing range.

[ukd1]

@bensheldon nice, saw the github actions update - looks like a great idea!