Closed r30shah closed 4 years ago
FYI @fjeremic
@fjeremic Made changes as recommended in https://github.com/AdoptOpenJDK/bumblebench/pull/11/commits/167f8a2a7723d38696de4ab3ce082f3db0679e35
@fjeremic I separated out the benchmark in net.adoptopenjdk.bumblebench.arraycopy
package.
I kept all sized elements yet as opposed to just Byte and Object. @0dvictor mentioned he would like experiment how much difference an element have on the arraycopy speed on the platform he is working on.
Thank you @r30shah. On X86, the generated code varies on different element sizes, and the best choice may be different. Therefore, keeping all the tests for all element sizes would be desirable.
@piyush286 / @smlambert do you might know how typically gets assigned to be a committer for these PRs? It would still be valuable to have them merged.
Will review/run shortly, thanks for the reminder.
Add benchmark to test array copy speed in forward or backward direction.
Signed-off-by: Rahil Shah rahil@ca.ibm.com