Yuhao Su
Yuhao Su
> Do you have time to look into why HashJoin is causing rustc to stuck? I'll take a look.
Preliminary benching (tpch q12) shows both policy produced similar result. However we are currently facing a compactor OOM issue that causes performance deterioration within 5 min. I will bench it...
> Will this PR be part of #5395 or #5395 will make this PR obsolete? No. They are not related.
Indeed hash join might be using const generic to generate duplicate code for non-critical sections. I'll take a look to see if this can be improved.
I'm curious if a struct has a generic `K`, will those functions of the struct that does not use `K` be expanded and compiled multiple time?
After a closer look, I think it is hard to make improvements by adjusting generic use. generic `K` (join key type) is already used across the hash join executor. We...
Hmmm, it might be possible. We can make the fetching from remote part return `Box`. But this will make our code looks less readable and harder to refactor those code...
> Operator cache doesn't have TTL so it is inconsistent with storage and may cause wrong results. It can be hard to define a _correct_ result when something is TTLed?...
> To generalize a bit, we should do things as if vectorization provides 0 benefits. But since vectorization actually makes things faster, we should have a tradeoff/balance here. Vectorization can...