RFR: 8304824: NMT should not use ThreadCritical [v8]

[Robert Toyonaga]

On Sat, 12 Oct 2024 18:42:34 GMT, Afshin Zafari <azafari at openjdk.org> wrote:

>> Robert Toyonaga has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Update src/hotspot/share/utilities/vmError.cpp
>>   
>>   Co-authored-by: David Holmes <62092539+dholmes-ora at users.noreply.github.com>
>
> Just an important test or comparison is needed here. Do you have any comparison of  the performance before and after this PR change set? Even if no degrading is expected, better to test it.

Hi @afshin-zafari, I made a [simple test](https://gist.github.com/roberttoyonaga/b2315de83f62ef0f4cef8e4d08ba18d3) and could not find any significant difference in performance. The test basically spawns 16 threads (on my 16 core amd linux machine) and each thread reserves and commits memory thousands of times before finishing. I record how long it takes for all the threads to finish.

 The after 10 trials, the average difference between with/without the mutex change is ~3%, which is less than the stdev (5% of the average values).  It's a pretty rough test, but at least it supports our expectation and shows that there's no obvious degradation.

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

PR Comment: https://git.openjdk.org/jdk/pull/20852#issuecomment-2427685056