RFR: 8318986: Improve GenericWaitBarrier performance [v5]

[Claes Redestad]

On Tue, 7 Nov 2023 20:22:13 GMT, Aleksey Shipilev <shade at openjdk.org> wrote:

>> See the symptoms, reproducer and analysis in the bug.
>> 
>> Current code waits on `disarm()`, which effectively stalls leaving the safepoint if some threads lag behind. Having more runnable threads than CPUs nearly guarantees that we would wait for quite some time, but it also reproduces well if you have enough threads near the CPU count. Just waiting at `arm()` is insufficient, but we can have several `Semaphores` to do what we want. 
>> 
>> This PR implements a more efficient `GenericWaitBarrier` to recover the performance. Most of the implementation discussion is in the code comments. The key observation that drives this work is that we want to reuse `Semaphore` and related counters without being stuck waiting for threads to leave.
>> 
>> (AFAICS, futex-based `LinuxWaitBarrier` does roughly the same, but handles this reuse on futex side, by assigning the "address" per futex.)
>> 
>> This issue affects everything except Linux. I initially found this on my M1 Mac, but pretty sure it is easy to reproduce on Windows as well. The safepoints from the reproducer in the bug improved dramatically on a Mac. Note not only the orders of magnitude better safepoint times, but also the several times more GC safepoints in the time-bound allocation test, which means the attainable GC throughput is similarly better, since we don't waste time at this wait barrier.
>> 
>> ![plot-generic-wait-barrier-macos](https://github.com/openjdk/jdk/assets/1858943/3440f200-2360-482a-b8b2-71385cca35b7)
>> 
>> Additional testing:
>>   - [x] MacOS AArch64 server fastdebug, `tier1`
>>   - [x] Linux x86_64 server fastdebug, `tier1 tier2 tier3` (generic wait barrier enabled explicitly)
>>   - [x] Linux AArch64 server fastdebug, `tier1 tier2 tier3` (generic wait barrier enabled explicitly)
>>   - [x] MacOS AArch64 server fastdebug, `tier2 tier3`
>>   - [x] Linux x86_64 server fastdebug, `tier4` (generic wait barrier enabled explicitly)
>>   - [x] Linux AArch64 server fastdebug, `tier4` (generic wait barrier enabled explicitly)
>
> Aleksey Shipilev has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Rework paddings

Most quick benchmarks are done, but the bulk will take the night to complete. 

Most results so far are in the noise, though I see a few minor Windows-only regressions that raise suspicion in aggregate: Most of the `org.openjdk.bench.javax.crypto.small.SignatureBench.RSA` and `org.openjdk.bench.javax.crypto.small.RSABench` microbenchmarks regress (0.3-0.8%), then a few J2DBench sub-benchmarks that regress 0.8-1.5% across, similarly across all tested GC combinations (parallel, G1, ZGC). None of these look like integration blockers on their own, but if they are real it'd be nice to get on top of them early. I'll have to re-run and diagnose a bit further to say for sure that this isn't a red herring (or caused by something sneaking into my baseline build). 

Perhaps you can check if the RSA micro regressions are reproducible on your end?

-------------

PR Comment: https://git.openjdk.org/jdk/pull/16404#issuecomment-1809187098