On Tue, 14 Jul 2026 10:56:10 GMT, Andrew Haley <[email protected]> wrote:

>> Please use [this 
>> link](https://github.com/openjdk/jdk/pull/28541/changes?w=1) to view the 
>> files changed.
>> 
>> Profile counters scale very badly.
>> 
>> The overhead for profiled code isn't too bad with one thread, but as the 
>> thread count increases, things go wrong very quickly.
>> 
>> For example, here's a benchmark from the OpenJDK test suite, run at 
>> TieredLevel 3 with one thread, then three threads:
>> 
>> 
>> Benchmark (randomized) Mode Cnt Score Error Units
>> InterfaceCalls.test2ndInt5Types false avgt 4 27.468 ± 2.631 ns/op
>> InterfaceCalls.test2ndInt5Types false avgt 4 240.010 ± 6.329 ns/op
>> 
>> 
>> This slowdown is caused by high memory contention on the profile counters. 
>> Not only is this slow, but it can also lose profile counts.
>> 
>> This patch is for C1 only. It'd be easy to randomize C1 counters as well in 
>> another PR, if anyone thinks it's worth doing.
>> 
>> One other thing to note is that randomized profile counters degrade very 
>> badly with small decimation ratios. For example, using a ratio of 2 with 
>> `-XX:ProfileCaptureRatio=2` with a single thread results in
>> 
>> 
>> Benchmark                        (randomized)  Mode  Cnt   Score   Error  
>> Units
>> InterfaceCalls.test2ndInt5Types         false  avgt    4  80.147 ± 9.991  
>> ns/op
>> 
>> 
>> The problem is that the branch prediction rate drops away very badly, 
>> leading to many mispredictions. It only really makes sense to use higher 
>> decimation ratios, e.g. 64.
>> 
>> ---------
>> - [x] I confirm that I make this contribution in accordance with the 
>> [OpenJDK Interim AI Policy](https://openjdk.org/legal/ai).
>
> Andrew Haley has updated the pull request incrementally with two additional 
> commits since the last revision:
> 
>  - Merge branch 'JDK-8134940' of https://github.com/theRealAph/jdk into 
> JDK-8134940
>  - Simplify overflow handling

Looks ok, modulo a few queries of my own and a small number of hanging comments 
from @shipilev.

src/hotspot/cpu/aarch64/c1_LIRAssembler_aarch64.cpp line 2670:

> 2668: 
> 2669:   if (counter_stub != nullptr) {
> 2670:     __ ubfx(rscratch1, r_profile_rng, 28 - ratio_shift, ratio_shift);

Why is this 28 - ratio_shift?

src/hotspot/cpu/x86/c1_MacroAssembler_x86.cpp line 278:

> 276:     // In effect, the CRC instruction is being used here for its
> 277:     // linear feedback shift register.
> 278:     movl(temp, 0);

On aarch64 we have asserts to detect a zero state when using crc32. Why do we 
not also need them here on x86?
Likewise in the save and restore routines?

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

PR Review: https://git.openjdk.org/jdk/pull/28541#pullrequestreview-4693412181
PR Review Comment: https://git.openjdk.org/jdk/pull/28541#discussion_r3579498463
PR Review Comment: https://git.openjdk.org/jdk/pull/28541#discussion_r3579993366

Reply via email to