RFR: 8322149: ConcurrentHashMap copy constructor should not transfer from old table on presizing [v2]

[Joshua Cao]

On Wed, 3 Jan 2024 19:52:34 GMT, Joshua Cao <duke at openjdk.org> wrote:

>> test/micro/org/openjdk/bench/java/util/concurrent/Maps.java line 122:
>> 
>>> 120:     public void testCopyConstructor() {
>>> 121:         ConcurrentHashMap<Integer, Integer> clone = new ConcurrentHashMap<>(staticMap);
>>> 122:         dummy = clone;
>> 
>> Is this for preventing dead-code elimination? If so, just do:
>> 
>> 
>>   return new ConcurrentHashMap<>(staticMap);
>
> I tried this. I don't think its working. When I just return the map, we lose the improvements in performance, and the benchmark overall just runs much faster. I'm guessing the map got DCE'ed

Scratch that. Was my mistake. Was running the wrong test.

I'm renaming `testCopyConstructor` to `testConcurrentHashMapCopyConstructor`, since the file name + test don't actually explicitly say its for ConcurrentHashMap.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/17116#discussion_r1440968465