On Tue, 25 Nov 2025 12:22:29 GMT, Per Minborg <[email protected]> wrote:
>> This PR proposes to consolidate and standardize the benchmarks related to >> the bulk operations for memory segments. >> >> All the benchmarks will now run with >> (heap|native)x(aligned|unaligned)x(intint|intlong|longlong) combinations. >> >> It should be noted that with this change, the running time for the >> benchmarks would increase significantly. > > Per Minborg has updated the pull request with a new target base due to a > merge or a rebase. The incremental webrev excludes the unrelated changes > brought in by the merge/rebase. The pull request contains nine additional > commits since the last revision: > > - Update after comments > - Merge branch 'master' into improved-bulk-benchmarks > - Add loop variant for mismatch > - Use other allocateFrom overload > - Remove lingering comment > - Refactor and use long arrays > - Remove BulkOps > - Fix copyright years > - Improve BulkOp benchmarks Benchmark (size) Mode Cnt Score Error Units SegmentBulkCopy.Array.bufferCopy 2 avgt 30 3.028 ± 0.094 ns/op SegmentBulkCopy.Array.bufferCopy 4 avgt 30 2.825 ± 0.047 ns/op SegmentBulkCopy.Array.bufferCopy 8 avgt 30 2.694 ± 0.048 ns/op SegmentBulkCopy.Array.bufferCopy 12 avgt 30 2.736 ± 0.019 ns/op SegmentBulkCopy.Array.bufferCopy 16 avgt 30 2.705 ± 0.024 ns/op SegmentBulkCopy.Array.bufferCopy 64 avgt 30 2.851 ± 0.017 ns/op SegmentBulkCopy.Array.bufferCopy 512 avgt 30 7.329 ± 0.116 ns/op SegmentBulkCopy.Array.bufferCopy 4096 avgt 30 51.591 ± 0.667 ns/op SegmentBulkCopy.Array.bufferCopy 32768 avgt 30 626.952 ± 43.598 ns/op SegmentBulkCopy.Array.bufferCopy 262144 avgt 30 3339.947 ± 16.218 ns/op SegmentBulkCopy.Array.bufferCopy 2097152 avgt 30 26569.111 ± 1974.834 ns/op SegmentBulkCopy.Array.bufferCopy 16777216 avgt 30 258056.303 ± 9005.471 ns/op SegmentBulkCopy.Array.bufferCopy 134217728 avgt 30 2335005.220 ± 43871.900 ns/op ------------- PR Comment: https://git.openjdk.org/jdk/pull/28260#issuecomment-3575400339
