DragonFly BSD
DragonFly kernel List (threaded) for 2012-08
[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]

Re: GSoC: Add SMT/HT awareness to DragonFlyBSD scheduler


From: Mihai Carabas <mihai.carabas@xxxxxxxxx>
Date: Sun, 5 Aug 2012 19:45:00 +0300

--14dae93407a56ae87804c6877b8e
Content-Type: text/plain; charset=ISO-8859-1

Hello,

At [1] you have a chart with a pgbench running on monster for a range of 5
to100 threads/clients (for 5, there are 5 threads and 5 clients, so a total
of 10 new processes). As you can see, when the number of running processes
is greater than the number of cores, the cache-coherent heuristic has much
better results (in some cases more than ~15%).
I modified my cache-coherent heuristic, for the monster, so that it tries
to stick a process to a socket, no matter what core. The L3 cache comes to
work, at that's why we have these results.
Now I'm working to make a tunable option for selecting the level you want
to stick a process (thread level, core level, socket/package level). By
tomorrow, I will commit it.
Also, this week I cleaned up the code and introduce KTR debug options
instead of my old kprintfs. The next week, I will come with this kind of
charts for core i3, core i7 (alexh) and dual-xeon (ftigeot).

[1] http://leaf.dragonflybsd.org/~mihaic/pgbench_monster.pdf

Mihai

--14dae93407a56ae87804c6877b8e
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Hello,<div><br></div><div>At [1] you have a chart with a pgbench running on=
 monster for a range of 5 to100 threads/clients (for 5, there are 5 threads=
 and 5 clients, so a total of 10 new processes). As you can see, when the n=
umber of running processes is greater than the number of cores, the cache-c=
oherent heuristic has much better results (in some cases more than ~15%).</=
div>
<div>I modified my cache-coherent heuristic, for the monster, so that it tr=
ies to stick a process to a socket, no matter what core. The L3 cache comes=
 to work, at that&#39;s why we have these results.</div><div>Now I&#39;m wo=
rking to make a tunable option for selecting the level you want to stick a =
process (thread level, core level, socket/package level). By tomorrow, I wi=
ll commit it.</div>
<div>Also, this week I cleaned up the code and introduce KTR debug options =
instead of my old kprintfs. The next week, I will come with this kind of ch=
arts for core i3, core i7 (alexh) and dual-xeon (ftigeot).</div><div><br>
</div><div>[1] <a href=3D"http://leaf.dragonflybsd.org/~mihaic/pgbench_mons=
ter.pdf">http://leaf.dragonflybsd.org/~mihaic/pgbench_monster.pdf</a></div>=
<div><br></div><div>Mihai</div>

--14dae93407a56ae87804c6877b8e--



[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]