RFR: 8322043: HeapDumper should use parallel dump by default [v3]

[David Holmes]

On Tue, 16 Apr 2024 19:45:12 GMT, Alex Menkov <amenkov at openjdk.org> wrote:

>> src/hotspot/share/services/heapDumper.hpp line 63:
>> 
>>> 61:   // additional info is written to out if not null.
>>> 62:   // compression >= 0 creates a gzipped file with the given compression level.
>>> 63:   // parallel_thread_num >= 0 indicates thread numbers of parallel object dump, -1 means "auto select".
>> 
>> I don't understand why you need to add `-1` to mean "auto-select" instead of just setting the default parameter to be `default_num_of_dump_threads()`?
>
> I think it makes the code more flexible - it allows to distinguish between "use default value" and "I don't care" cases.

I'm not sure it is a worthwhile distinction. Not passing an actual parameter means "I don't care - take the default".

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

PR Review Comment: https://git.openjdk.org/jdk/pull/18748#discussion_r1568350751