RFR (S/M): 8142484: Let IHOP follow the current heap size, not the maximum heap size (original) (raw)
Bengt Rutisson bengt.rutisson at oracle.com
Thu Mar 3 14🔞17 UTC 2016
- Previous message (by thread): RFR (S/M): 8142484: Let IHOP follow the current heap size, not the maximum heap size
- Next message (by thread): RFR (S/M): 8142484: Let IHOP follow the current heap size, not the maximum heap size
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Hi Thomas,
On 2016-03-03 15:07, Thomas Schatzl wrote:
Hi all,
can I have reviews for this change that updates what maximum threshold (capacity) the IHOP mechanisms follow? Instead of following the maximum capacity, IHOP follows the current capacity now. Changes are relatively small, instead of passing a fixed threshold capacity at initialization, add and use a method in the G1IHOPControl class to update to the current threshold. CR: https://bugs.openjdk.java.net/browse/JDK-8142484 Webrev: cr.openjdk.java.net/~tschatzl/8142484/webrev
I think this looks good.
One very minor nit for line 569 in g1CollectorPolicy.cpp:
569 _ihop_control->update_target_occupancy(HeapRegion::GrainBytes * new_number_of_regions);
This would read more naturally for me if the multiplication was swapped around, so:
569 _ihop_control->update_target_occupancy(new_number_of_regions * HeapRegion::GrainBytes);
I don't need an updated review for that.
Thanks, Bengt
Testing: jprt, local testing that it works, vm.gc run
Thanks, Thomas
- Previous message (by thread): RFR (S/M): 8142484: Let IHOP follow the current heap size, not the maximum heap size
- Next message (by thread): RFR (S/M): 8142484: Let IHOP follow the current heap size, not the maximum heap size
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]