RFR: JDK-8319314: NMT detail report slow or hangs for large number of mappings

[Johan Sjölen]

On Thu, 2 Nov 2023 16:42:58 GMT, Thomas Stuefe <stuefe at openjdk.org> wrote:

> If a lot of mappings have been registered with NMT, NMT detail report may get so slow as to hang seemingly. 
> 
> 100 k mappings will take > 1 minute
> 1 million mappings never returns (I canceled the test after 90 minutes)
> 
> We burn all the time doing a sorted insert into the target array of the MemBaseline object. That sorted insert is pointless since the source array is already sorted by region base address.
> 
> Omitting the pointless sort drops the runtime for 1 million mappings from > 90 minutes to 48 seconds.
> 
> Tests: Tested locally. GHAs (Windows error unrelated).

Hi,

Thanks for this, one small thing to be fixed.

src/hotspot/share/nmt/memBaseline.cpp line 130:

> 128: 
> 129:   bool do_allocation_site(const ReservedMemoryRegion* rgn)  {
> 130:     assert(rgn->base() >= _last_base, "region unordered?");

It seems like we don't assign to `_last_base`?

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

Marked as reviewed by jsjolen (Reviewer).

PR Review: https://git.openjdk.org/jdk/pull/16480#pullrequestreview-1712181521
PR Review Comment: https://git.openjdk.org/jdk/pull/16480#discussion_r1381383524