Francesco Nigro

QPIDJMS-552 JMS 2 Completion threads shouldn't scale with the number of sessions

28

This implements an elastic per-connection thread pool that allows session completions to be consumed in order (as now) while saving create a dedicated thread per session. In a shared connection...

QPIDJMS-417 Reduce GC pressure while using BytesMessage

3

Using directly ByteBuf-based streams allows to avoid unnecessary creations of intermediate instances to operate on the underline ByteBuf content

Reusing existing Netty types to speedup http encoding and save scalability issues

2

This can be completed only with: - proper benchmarking - https://github.com/netty/netty/pull/12796 is completed

Using https://github.com/franz1981/type-pollution-agent in a JMH benchmark, this is the profiling data:  I'm new of java assist; but is it expected?

Improve array ring's size performance

1

see https://github.com/belaban/jgroups-raft/commit/a749eb55b597e177a7f3653cd9ce31314de3daae#r74658924

InMemoryLog should use a growing mechanism similar to FilePositionCache

1

`FilePositionCache` allow accumulating entries while appending fixed sized chunks/pages saving expensive memory copies while enlarging capacity: `InMemoryLog` could make use of a similar mechanism for its entries. Given that `InMemoryLog`...

InMemoryLog entries on repeated counter perf runs keep on growing

5

While running single node repeated tests with `InMemoryLog` performance degrades over runs because LogEntry on `InMemoryLog::entries` keep on growing.

Support target throughput benchmarks and microseconds precision

1

In order to allow fair comparisons of resource utilization and/or latencies between runs that push different load against a system, would be great to have a target throughput benchmark mode....

Inconsistent composite score best score computation

2

Looking at `plaintext`, for `officefloor`: https://www.techempower.com/benchmarks/#hw=ph&test=plaintext&section=data-r22&c=e&o=e&l=xan3i7-cn3 it scores its best at `33,107`, but while moving to the `Composite Score` tab (at https://www.techempower.com/benchmarks/#hw=ph&test=composite&section=data-r22&c=e&o=e&l=xan3i7-cn3), it get `5,648,781` resulting far higher in the...

Update Citrine to Ubuntu 22.04 performance issues

21

As commented by @nbrady-techempower in https://github.com/TechEmpower/FrameworkBenchmarks/issues/7321#issuecomment-1466365129 here we are! I don't have a clue why/what's going on, but let's take a look at both vertx and redkale before/after the upgrade,...