klion26 commented on PR #9689:
URL: https://github.com/apache/arrow-rs/pull/9689#issuecomment-4303511941
Seems there are still some regressions in the benchmark results. but I can't
reproduce in my laptop, do you have some insight here? Thanks @alamb @scovich
The things I tried:
- run benchmark on my laptop, didn't reproduce the result
- munually check the `cast decimal to integer` did not find difference
- write a demo to verify that the result of `cast_decimal_to_integer` did
not change (this wants to verify whether there are
- remove all the irrelavent benchmark cases, only keep the benchmarks cases
`cast decimal64 to uint8/uint16/int8/int16`, and make the array size to
`80M`(try to extend the run time and find the difference), did not find the
reason currently.
- the commands used to generate the pic are: 1) run `cargo bench
--features=arrow,async,test_common,experimental,object_store --bench
cast_kernels --no-run`; 2) run `samply record
./target/release/deps/cast_kernels-ac4bbb5659a24b1b` to generate the pic
- the first pic was generated on main branch, and the second was
generated on the current branch
- _one thing wired is that the call number of benchmark function is
different, there are more samples of `cast_kernels::main` but less
`cast_kernels::add_benchmark` on the main branch, not sure there is related to
my env, or the bench group setting or something else._
<img width="1207" height="676" alt="image"
src="https://github.com/user-attachments/assets/48eb6903-9425-42d7-baf7-4c6cf953b539";
/>
<img width="1237" height="626" alt="image"
src="https://github.com/user-attachments/assets/57ea0a15-8006-4eed-b70a-9a812ed138ec";
/>
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
