> -----Original Message----- > From: Nitin Gupta <nigupta@xxxxxxxxxx> > Sent: Monday, March 2, 2020 1:34 PM > To: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx>; Michal Hocko > <mhocko@xxxxxxxx>; Vlastimil Babka <vbabka@xxxxxxx> > Cc: Nitin Gupta <nigupta@xxxxxxxxxx>; Matthew Wilcox > <willy@xxxxxxxxxxxxx>; Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>; > Mike Kravetz <mike.kravetz@xxxxxxxxxx>; linux-kernel <linux- > kernel@xxxxxxxxxxxxxxx>; linux-mm <linux-mm@xxxxxxxxx> > Subject: [PATCH v2] mm: Proactive compaction > > - With 5.6.0-rc3 + this patch, with proactiveness=20 > > echo 20 | sudo tee /sys/kernel/mm/compaction/node-*/proactiveness > > percentile latency > –––––––––– ––––––– > 5 2 > 10 2 > 25 3 > 30 3 > 40 3 > 50 4 > 60 4 > 75 4 > 80 4 > 90 5 > 95 429 > The percentile latency data above is correct but the summary line is Incorrect: > Total 2M hugepages allocated = 11120 (21.7G worth of hugepages out of 25G > total free => 98% of free memory could be allocated as hugepages) > Corrected: Total 2M hugepages allocated = 384105 (750G worth of hugepages out of 762G total free => 98% of free memory could be allocated as hugepages) Thanks, Nitin
