nushell: Rendering with `tabled` seems about 10x slower than before
Describe the bug
After merging https://github.com/nushell/nushell/pull/5969 I’m seeing about 10x performance drop.
How to reproduce
> 1..100000
...
99997 99998
99998 99999
99999 100000
> $env.CMD_DURATION_MS
3795
On my machine, before the PR, it takes about 350 ms. After the PR it takes 3.7 s.
Expected behavior
Ideally, the performance difference wouldn’t be so big. Rendering this big data is not useful but it can happen by accident and then you need to wait.
Screenshots
No response
Configuration
| key | value |
|---|---|
| version | 0.65.1 |
| branch | |
| commit_hash | d1687df0678ec55241031860e6f7132dec3f2503 |
| build_os | linux-x86_64 |
| build_target | x86_64-unknown-linux-gnu |
| rust_version | rustc 1.62.0 (a8314ef7d 2022-06-27) |
| rust_channel | stable-x86_64-unknown-linux-gnu |
| cargo_version | cargo 1.62.0 (a748cf5a3 2022-06-08) |
| pkg_version | 0.65.1 |
| build_time | 2022-07-06 23:11:48 +03:00 |
| build_rust_channel | release |
| features | database, dataframe, default, trash, which, zip |
| installed_plugins | gstat |
Additional context
I’m using the light table mode and Alacritty terminal.
About this issue
- Original URL
- State: closed
- Created 2 years ago
- Comments: 29 (17 by maintainers)
@zhiburt benchmarks update with nu-table above.
@zhiburt I tried https://github.com/nushell/nushell/pull/6049 and it takes like 430 ms! That’s definitely more than good enough.
overall it’s AWESOME! whatever was going on with table_mode light in WSL is fixed apparently.
I was thinking the same thing. I just need to figure out what commit to go back to. Any ideas?
Ok, i’ll just make a new table under the one above.
yes
I’ve already told you once. it’s not me all the credit to criterion https://github.com/bheisler/criterion.rs
The benches you can find here https://github.com/zhiburt/tabled/blob/comparision-libs-bench/benches/benchmark.rs