Closed GoogleCodeExporter closed 9 years ago
I had the same thought.
I imagine that when we hit memory limit PageHeap::IncrementalScavenge should be
pushed forward (and if there is nothing more to Scavenge then call new_handler).
My current problem with IncrementalScavenge is that it's capable to return
unused memory to the system, but it's too difficult to control.
Original comment by pafi...@gmail.com
on 23 Aug 2012 at 9:40
Here you go. Not very stress tested yet, but unittest passes.
You can control it by TCMALLOC_HEAP_LIMIT_MB env variable (see description in
the patch).
Original comment by pafi...@gmail.com
on 1 Sep 2012 at 1:15
Attachments:
[deleted comment]
Updated patch with unittests fixes + comments and indentation changes.
TCMALLOC_DISABLE_MEMORY_RELEASE switch added to emulate platform with no
MADV_DONTNEED support (for testing purposes).
Original comment by pafi...@gmail.com
on 2 Sep 2012 at 9:31
Attachments:
Original comment by chapp...@gmail.com
on 3 Nov 2012 at 4:56
r178 | chappedm@gmail.com | 2012-11-04 14:02:15 -0500 (Sun, 04 Nov 2012) | 5
lines
issue-448: New environment setting that allows you to set an explicit heap limit
TCMALLOC_HEAP_LIMIT_MB - The maximum amount of heap memory that tcmalloc can
use.
TCMALLOC_DISABLE_MEMORY_RELEASE - emulate platform with no MADV_DONTNEED
support (generally for testing purposes).
Original comment by chapp...@gmail.com
on 4 Nov 2012 at 7:02
Original issue reported on code.google.com by
ice...@google.com
on 12 Jul 2012 at 7:36