Hattonuri commented on PR #39054:
URL: https://github.com/apache/arrow/pull/39054#issuecomment-1843783701
> I agree this looks a bit fragile even if in the current context this
wouldn't be a problem.
>
> @Hattonuri Did you run any benchmarks on this? Are there any tangible
benefits?
I run Decimal128Conversions and got these results
`benchmark baseline contender change % counters
Decimal128Conversion 18.102M items/sec 18.877M items/sec 4.285
{'family_index': 0, 'per_family_instance_index': 0, 'run_name':
'Decimal128Conversion', 'repetitions': 1, 'repetition_index': 0, 'threads': 1,
'iterations': 1921, 'allocs_per_iter': 2.375, 'max_bytes_used': 209536,
'total_allocated_bytes': 1931840}`
But this 4.285% increase is not stable.
In my reading code i have lowered cpu usage from ~590% to ~530%.
I did not find any other benchmarks with decimals
--
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]
