Bouncner

Results 182 comments of Bouncner

Improve output writing of hash-based aggregate

@dey4ss : the PR is still blocked by others (benchmarking script). But in case you're bored: is there anything that would block this PR here?

Improve output writing of hash-based aggregate

### ~Benchmarks~ Will redo once #2635 is in.

Improve output writing of hash-based aggregate

Re-running benchmarks. The following on AMD show some weird results (SSB gets slower in queries where the aggregate takes

Improve output writing of hash-based aggregate

~Something is odd. I made it kaputt.~

Improve output writing of hash-based aggregate

~I am waiting for the DELab which might still be down for a couple of days.~ Nobody knows when it's coming back.

Improve output writing of hash-based aggregate

@dey4ss : there have been tiny changes since your last review.

Improve output writing of hash-based aggregate

### Intel - Nemea **System** nemea - click to expand | property | value | | -- | -- | | Hostname | nemea | | CPU | Intel(R) Xeon(R)...

Improve output writing of hash-based aggregate

### AMD - Koroniea **System** koroneia - click to expand | property | value | | -- | -- | | Hostname | koroneia | | CPU | AMD EPYC...

Re-Evaluate Hash Maps

#2613 changes all third party maps to `boost::flat_unorderered_map`. To reduce the memory consumption, we still use the sparse map though (actual impact has not been evaluated).

Re-Evaluate Hash Maps

Given that there is currently no relevant alternative to `tsl::sparse_map` (apart from Google's abseil, that that's basically the same idea) and we voted for the memory saving in this data...