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

[Joshua Cao]

On Wed, 3 Jan 2024 14:34:50 GMT, Aleksey Shipilev <shade at openjdk.org> wrote:

>> ConcurrentHashMap's copy constructor calls `putAll()` -> `tryPresize()` -> `transfer()`. When coming from the copy constructor, the Map is empty, so there is nothing to transfer. But `transfer()` will still copy all the empty nodes from the old table to the new table.
>> 
>> This patch avoids this work by only calling `tryPresize()` if the table is already initialized. If `table` is null, the initialization is deferred to `putVal()`, which calls `initTable()`.
>> 
>> ---
>> 
>> ### JMH results for testCopyConstructor
>> 
>> before patch:
>> 
>> 
>> Result "org.openjdk.bench.java.util.concurrent.Maps.testCopyConstructor":
>>   937395.686 ±(99.9%) 99074.324 ns/op [Average]
>>   (min, avg, max) = (825732.550, 937395.686, 1072024.041), stdev = 92674.184
>>   CI (99.9%): [838321.362, 1036470.010] (assumes normal distribution)
>> 
>> 
>> after patch:
>> 
>> 
>> Result "org.openjdk.bench.java.util.concurrent.Maps.testCopyConstructor":
>>   620871.469 ±(99.9%) 59195.406 ns/op [Average]
>>   (min, avg, max) = (545304.633, 620871.469, 689013.573), stdev = 55371.419
>>   CI (99.9%): [561676.063, 680066.875] (assumes normal distribution)
>> 
>> 
>> Average time is decreased by about 33%.
>
> src/java.base/share/classes/java/util/concurrent/ConcurrentHashMap.java line 851:
> 
>> 849:      */
>> 850:     public ConcurrentHashMap(Map<? extends K, ? extends V> m) {
>> 851:         this(m.size(), LOAD_FACTOR, 1);
> 
> This looks like just `this(m.size())`, right? Not sure if we want to save an additional chained constructor call, though.

Yep, same thing. I'll change it.

> 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

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

PR Review Comment: https://git.openjdk.org/jdk/pull/17116#discussion_r1440798729
PR Review Comment: https://git.openjdk.org/jdk/pull/17116#discussion_r1440875149