[GitHub] [arrow] pitrou commented on a diff in pull request #12755: ARROW-16014: [C++] Create more benchmarks for measuring expression evaluation overhead

[GitBox <gi...@apache.org>]

pitrou commented on code in PR #12755:
URL: https://github.com/apache/arrow/pull/12755#discussion_r849591197


##########
cpp/src/arrow/compute/exec/expression_benchmark.cc:
##########
@@ -69,6 +70,26 @@ static void SimplifyFilterWithGuarantee(benchmark::State& state, Expression filt
   }
 }
 
+static void ExecuteScalarExpressionOverhead(benchmark::State& state, Expression expr) {
+  const auto rows_per_batch = static_cast<int32_t>(state.range(0));
+  const auto num_batches = 10000000 / rows_per_batch;
+
+  ExecContext ctx;
+  auto dataset_schema = schema({
+      field("x", int64()),
+  });
+  ExecBatch input({Datum(ConstantArrayGenerator::Int64(rows_per_batch, 5))},
+                  /*length=*/1);
+
+  ASSIGN_OR_ABORT(auto bound, expr.Bind(*dataset_schema));
+  for (auto _ : state) {
+    for (int it = 0; it < num_batches; ++it)
+      ABORT_NOT_OK(ExecuteScalarExpression(bound, input, &ctx).status());
+  }
+  state.counters["rows_per_second"] = benchmark::Counter(
+      state.iterations() * num_batches * rows_per_batch, benchmark::Counter::kIsRate);

Review Comment:
   Is this rows per second per thread?



-- 
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: github-unsubscribe@arrow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org