From: Michael Ellerman <michael@ellerman.id.au>
To: Dennis Schridde <devurandom@gmx.net>
Cc: cbe-oss-dev@lists.ozlabs.org, linuxppc-dev@lists.ozlabs.org,
arnd@arndb.de
Subject: Re: PROBLEM: Only 2 of 4 cores used on IBM Cell blades and no threads shown in spufs
Date: Tue, 23 Apr 2013 19:12:47 +1000 [thread overview]
Message-ID: <20130423091246.GA18195@concordia> (raw)
In-Reply-To: <1470334.YUWOQ37ijW@ernie>
On Mon, Apr 22, 2013 at 06:44:13PM +0200, Dennis Schridde wrote:
> Hello!
Hi Dennis,
> [1.] One line summary of the problem:
> Only 2 of 4 cores used on IBM Cell blades and no threads shown in spufs
>
>
> [2.] Full description of the problem/report:
> On my IBM Cell blades, only 2 out of the 4 CPU cores are being used, even when
> several threads are running.
Yes you're right, I see that too.
For me it is fixed by applying the following patch, it should be in v3.10:
http://patchwork.ozlabs.org/patch/230103/
Maybe you can try it.
> Also /spu is always empty, despite the SPUs being
Yes I see that too.
> For testing, I started an instance of cellminer with --ppe 2 --spe 16 (2
> threads on the PPEs, 16 on the SPEs) and htop reports two CPU cores being used
> at 100% and two others at 0%, while 4 threads are running (each at 50% CPU
> utilisation -> they share a core).
Does your cellminer actually work? ie. does it run OK?
I am using a fractal benchmark I had lying around, and it appears to
work, and runs fast enough that it must be running on the SPUs (I think).
So it sounds like we have some weirdness with stuff not appearing in
/spu, but spu programs are still able to run. Which is odd to say the
least.
cheers
next prev parent reply other threads:[~2013-04-23 9:12 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-04-22 16:44 PROBLEM: Only 2 of 4 cores used on IBM Cell blades and no threads shown in spufs Dennis Schridde
2013-04-23 9:12 ` Michael Ellerman [this message]
2013-04-23 11:51 ` Dennis Schridde
2013-04-23 13:16 ` Dennis Schridde
2013-04-23 13:19 ` Michael Ellerman
2013-05-17 15:46 ` Dennis Schridde
2013-05-22 3:37 ` Michael Ellerman
2013-04-23 12:01 ` Dennis Schridde
2013-04-23 12:16 ` Michael Ellerman
2013-04-23 12:41 ` Dennis Schridde
2013-04-23 12:53 ` Michael Ellerman
2013-04-23 13:24 ` Michael Ellerman
2013-04-23 18:13 ` Dennis Schridde
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20130423091246.GA18195@concordia \
--to=michael@ellerman.id.au \
--cc=arnd@arndb.de \
--cc=cbe-oss-dev@lists.ozlabs.org \
--cc=devurandom@gmx.net \
--cc=linuxppc-dev@lists.ozlabs.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).