Re: Abridged summary of mechanical-sympathy@googlegroups.com - 7 updates in 2 topics

105 views
Skip to first unread message

isaiah perumalla

unread,
Aug 8, 2017, 11:46:40 AM8/8/17
to mechanica...@googlegroups.com

On Tue, 8 Aug 2017 at 15:10, <mechanica...@googlegroups.com> wrote:
Oleg Mazurov <omaz...@splicemachine.com>: Aug 07 10:29PM -0700

I ran the benchmark with a profiler and was able to reproduce both modes,
fast and slow. The difference appears to be due to how HotSpot compiles the
DefaultMailbox.run() -> ...more
Peter Veentjer <alarm...@gmail.com>: Aug 07 08:42AM -0700

Hi Everyone,
 
I have a failing understanding the problem with transparent huge pages.
 
I 'understand' how normal pages work. A page is typically 4kb in a virtual
address space; each process has ...more
Marshall Pierce <mars...@mpierce.org>: Aug 07 10:54AM -0500

AFAIK, The issue is not huge pages that applications opt in to and
control; it's when they are "transparent", as in when the kernel
"helpfully" decides when to use them. This can lead to significant ...more
Peter Veentjer <alarm...@gmail.com>: Aug 07 09:02AM -0700

Your answer makes a lot of sense. It makes it the 'transparent' part clear.
 
What doesn't make sense is why the kernel would decide to switch between
large/small pages. So what is the benefit to ...more
Gil Tene <g...@azul.com>: Aug 07 10:25AM -0700

THP certainly sits in my "just don't do it" list of tuning things due to
it's fundamental dramatic latency disruption in current implementations,
seen as occasional 10s to 100s of msec (and ...more
Gil Tene <g...@azul.com>: Aug 07 11:14AM -0700

To highlight the problematic path in current THP kernel implementations,
here is an example call trace that can happen (pulled from the discussion
linked to below). It shows that a simple ...more
"Alen Vrečko" <alen....@gmail.com>: Aug 07 08:50PM +0200

Saw this a while back.
 
https://shipilev.net/jvm-anatomy-park/2-transparent-huge-pages/
 
Basically using THP/defrag with madvise and using
-XX:+UseTransparentHugePages -XX:+AlwaysPreTouch JVM ...more
You have received this digest because you're subscribed to updates for this group. You can change your settings on the group membership page.
To unsubscribe from this group and stop receiving emails from it send an email to mechanical-symp...@googlegroups.com.
--
Isaiah Perumalla
Reply all
Reply to author
Forward
0 new messages