On Tue, 2018-03-06 at 16:23 -0800, Andrew Morton wrote:
> (switched to email.A A Please respond via emailed reply-to-all, not via
> the
> bugzilla web interface).
> 
> On Sat, 24 Feb 2018 03:56:38 +0000 bugzilla-daemon@bugzilla.kernel.or
> g wrote:
> 
> > 
> > https://bugzilla.kernel.org/show_bug.cgi?id=198913
> > 
> > A A A A A A A A A A A A Bug ID: 198913
> > A A A A A A A A A A A Summary: Swapping slows to about 20% of expected speed
> > when
> > A A A A A A A A A A A A A A A A A A A A using multiple swap partitions on separate
> > drives
> > A A A A A A A A A A A A A A A A A A A A (striping).
> > 
> > ...
> > 
> > Problems
> > 1) Linux's swap striping on nine drives performs at only 19.5% of
> > the speed of
> > a memory-mapped file on a RAID 0 partition on the same nine drives.
> > 2) Linux's swap striping performs best with exactly four swap
> > partitions and
> > performs almost as well with just three partitions.
> > 
> Well that's interesting, thanks.A A Let's get this onto the mm
> developers' mailing list and perhaps someone will have a theory.A A And
> perhaps Hugh will have a think about it when he returns to the fold.
Thanks, if let me know if you need me to run some test code or
something.