vertexclique commented on a change in pull request #8664:
URL: https://github.com/apache/arrow/pull/8664#discussion_r530327927
##########
File path: rust/arrow/src/compute/kernels/aggregate.rs
##########
@@ -141,32 +146,41 @@ where
match array.data().null_buffer() {
None => {
- let sum = data.iter().fold(T::default_value(), |accumulator,
value| {
- accumulator + *value
- });
+ let total = data
+ .par_iter()
Review comment:
This is what MKL suggests, Array level non-primitive parallelism has
shared pointer overhead, that's the reason. I had so much bad time before with
arc overhead in my projects.
----------------------------------------------------------------
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:
us...@infra.apache.org
