[GitHub] [arrow] jianxind commented on a change in pull request #7607: ARROW-8996: [C++] Add AVX version for aggregate sum/mean with runtime dispatch

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

jianxind commented on a change in pull request #7607:
URL: https://github.com/apache/arrow/pull/7607#discussion_r451914666



##########
File path: cpp/src/arrow/compute/registry.cc
##########
@@ -115,6 +116,19 @@ static std::unique_ptr<FunctionRegistry> CreateBuiltInRegistry() {
   RegisterVectorNested(registry.get());
   RegisterVectorSort(registry.get());
 
+  // SIMD functions
+  auto cpu_info = arrow::internal::CpuInfo::GetInstance();
+#if defined(ARROW_HAVE_RUNTIME_AVX2)
+  if (cpu_info->IsSupported(arrow::internal::CpuInfo::AVX2)) {
+    RegisterScalarAggregateSumAvx2(registry.get());

Review comment:
       I don't note the simd_level parameter before. We can put it a further optimization then, I guess there's still some jobs to make simd_level work as expected.




----------------------------------------------------------------
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:
users@infra.apache.org