Bei-z commented on a change in pull request #8542:
URL: https://github.com/apache/arrow/pull/8542#discussion_r517611814
##########
File path: cpp/src/arrow/util/decimal_benchmark.cc
##########
@@ -158,6 +158,7 @@ static void BinaryMathOp256(benchmark::State& state) { //
NOLINT non-const refe
for (auto _ : state) {
for (int x = 0; x < kValueSize; x += 5) {
benchmark::DoNotOptimize(v1[x + 2] * v2[x + 2]);
+ benchmark::DoNotOptimize(v1[x + 3] / v2[x + 3]);
Review comment:
As discussed offline, moved back the if statements in FillInArray for
BasicDecimal128, the that the current benchmark result is:
BinaryMathOp128 152 ns 152 ns 4284240
items_per_second=65.6334M/s
BinaryMathOp256 163 ns 163 ns 4186744
items_per_second=61.4706M/s
Compared with the benchmark result without this pr:
BinaryMathOp128 154 ns 154 ns 4423113
items_per_second=64.9309M/s
BinaryMathOp256 35.4 ns 35.4 ns 18341932
items_per_second=282.487M/s
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
