westonpace commented on code in PR #12755:
URL: https://github.com/apache/arrow/pull/12755#discussion_r849857816
##########
cpp/src/arrow/compute/exec/expression_benchmark.cc:
##########
@@ -120,5 +148,31 @@ BENCHMARK_CAPTURE(BindAndEvaluate, nested_array,
BENCHMARK_CAPTURE(BindAndEvaluate, nested_scalar,
field_ref(FieldRef("struct_scalar", "float")));
+BENCHMARK_CAPTURE(ExecuteScalarExpressionOverhead, complex_expression,
complex_expression)
+ ->ArgNames({"rows_per_batch"})
+ ->RangeMultiplier(10)
Review Comment:
I think we can safely drop the 10M down to 1M. I originally asked for 10M
to see the effects when the batch size was greater than L3 but I don't think we
are seriously considering doing things with that size of batch. I think we can
also adjust the lower limit to 1000.
--
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]
