lkml.org 
[lkml]   [2020]   [Jun]   [18]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: [PATCH v16 03/14] mm/damon: Implement region based sampling
Date
On 2020-06-15T18:19:16+02:00 SeongJae Park <sjpark@amazon.com> wrote:

> From: SeongJae Park <sjpark@amazon.de>
>
> This commit implements DAMON's target address space independent high
> level logics for basic access check and region based sampling. The
> target address space specific logics for the monitoring target address
> regions construction and the access check are required, though. The
> following commits will provide reference implementations of those for
> the general virtual address spaces and the physical address space.
> Users can implement and use their own versions for their specific use
> cases, though.
>
> Basic Access Check
> ------------------
>
> DAMON basically reports what pages are how frequently accessed. The
> frequency is not an absolute number of accesses, but a ratio.
>
> For this, DAMON first calls target monitoring construction callback
> (``init_target_regions``), and then the access check callbacks, which is
> assumed to check the access to each page and aggregates the number of
> observed accesses of each page, for every ``sampling interval``.
> Finally, DAMON resets the aggregated count per ``aggregation interval``.
>
> This is thus similar to the common periodic access checks based
> monitoring mechanisms but provides the access frequency. The overhead
> will increase as the size of the target process grows.
>
> Region Based Sampling
> ---------------------
>
> To avoid the unbounded increase of the overhead, DAMON groups a number
> of adjacent pages that assumed to have same access frequencies into a
> region. As long as the assumption (pages in a region have same access
> frequencies) is kept, only one page in the region is required to be
> checked. Therefore, the monitoring overhead is controllable by setting
> the number of regions.
>
> Nonetheless, this scheme cannot preserve the quality of the output if
> the assumption is not kept. Following commit will introduce how we can
> make the guarantee with some sort of best effort.
>
> Signed-off-by: SeongJae Park <sjpark@amazon.de>
> ---

Reviewed-by: Leonard Foerster <foersleo@amazon.de>

\
 
 \ /
  Last update: 2020-06-18 11:24    [W:0.182 / U:0.168 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site