Alden Keefe Sampson
Alden Keefe Sampson
I'm curious what folks think of this and looking for input! One particular design question to note: this adds a nested concurrent_map (there's already one in the codec pipeline's read/read_batch)...
I'm working on a couple performance tests as Davis suggested above. Trying [airspeed velocity](https://asv.readthedocs.io/en/latest/index.html) which is new to me but looks useful.
Got pulled away after pushing those changes but before describing them -- I'll update this tomorrow with profiling results.
alright! - Tests added that check for coalesced (reduced count) of calls to store - Profiled execution time and store request count I don't necessarily need a detailed review at...
Thanks @JackKelly! I'd be happy for more benchmarking of this. I took a look at LDeakin's zarr_benchmarks and I would expect the changes in the PR to _not_ have any...
Any maintainers, I don't necessarily need a detailed review at this point -- if yall let me know any open questions you have or give a thumbs up ill work...
I'm approaching having time to wrap up the TODOs in the description and still eager to get this merged
**Ready for review!** at long last The PR optimizes reading part of a shard in two ways: 1. Requests for nearby chunks in the same shard are combined into a...
@d-v-b who should I request a review from on this?
Thank you for the first review @d-v-b! I didn't get to this before going out of office but ill address your comments when im back in ~1 week.