From: "Bryan O'Sullivan" <bos@pathscale.com>
To: Adrian Bunk <bunk@stusta.de>
Cc: rjwalsh@pathscale.com, rolandd@cisco.com, gregkh@suse.de,
akpm@osdl.org, davem@davemloft.net, linux-kernel@vger.kernel.org,
openib-general@openib.org
Subject: Re: [PATCH 18 of 20] ipath - kbuild infrastructure
Date: Mon, 13 Mar 2006 11:24:28 -0800 [thread overview]
Message-ID: <1142277868.9032.14.camel@serpentine.pathscale.com> (raw)
In-Reply-To: <20060313181025.GA13973@stusta.de>
On Mon, 2006-03-13 at 19:10 +0100, Adrian Bunk wrote:
> I'm still a bit surprised, since in the rest of the kernel we are even
> going from -O2 to -Os for getting better performance.
>
> Robert said he wanted to post some numbers showing that -O3 is
> measurably better for you [1], but I haven't seen them.
I just ran some numbers. At large packet sizes, it doesn't matter what
options we use, because we spend all of our time in __iowrite_copy32,
which uses the string copy instructions.
For small packets, my quick tests indicate that -Os gives about 5%
*better* performance than -O3 (using gcc 4 on FC4). This is in line
with what people have been finding in the kernel in general recently.
So if I change that CFLAGS line from -O3 to -Os, are we in OK
shape? :-)
> > +_ipath_idstr:="PathScale $(shell date +%F)"
> > +EXTRA_CFLAGS += -DIPATH_IDSTR='$(_ipath_idstr)' -DIPATH_KERN_TYPE=0
> >...
>
> UTS_VERSION is already available and printed at the top of dmesg.
> I don't see the point in printing it a second time.
Good point. The idstr stuff is for our out-of-tree drivers.
Thanks,
<b
next prev parent reply other threads:[~2006-03-13 19:24 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-03-10 0:35 [PATCH 0 of 20] [RFC] ipath driver - another round for review Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 1 of 20] ipath - core driver header files Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 2 of 20] ipath - core device driver Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 3 of 20] ipath - copy and send routines for sending an skb Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 4 of 20] ipath - support for HyperTransport devices Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 5 of 20] ipath - support for PCI Express devices Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 6 of 20] ipath - chip initialisation code Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 7 of 20] ipath - misc driver support code Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 8 of 20] ipath - sysfs support for core driver Bryan O'Sullivan
2006-03-10 1:11 ` Greg KH
2006-03-10 5:09 ` Bryan O'Sullivan
2006-03-05 3:08 ` Pavel Machek
2006-03-10 6:37 ` Greg KH
2006-03-10 14:59 ` Roland Dreier
2006-03-10 15:08 ` [openib-general] " Hal Rosenstock
2006-03-10 16:54 ` Greg KH
2006-03-10 17:05 ` Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 9 of 20] ipath - char devices for diagnostics and lightweight subnet management Bryan O'Sullivan
2006-03-10 0:45 ` Roland Dreier
2006-03-10 0:47 ` Bryan O'Sullivan
2006-03-10 0:52 ` Roland Dreier
2006-03-10 0:35 ` [PATCH 10 of 20] ipath - support for userspace apps using core driver Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 11 of 20] ipath - layering interfaces used by higher-level driver code Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 12 of 20] ipath - infiniband header files Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 13 of 20] ipath - infiniband UC and UD protocol support Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 14 of 20] ipath - infiniband RC " Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 15 of 20] ipath - misc infiniband code, part 1 Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 16 of 20] ipath - misc infiniband code, part 2 Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 17 of 20] ipath - infiniband verbs support Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 18 of 20] ipath - kbuild infrastructure Bryan O'Sullivan
2006-03-13 18:10 ` Adrian Bunk
2006-03-13 18:38 ` Robert Walsh
2006-03-13 19:24 ` Bryan O'Sullivan [this message]
2006-03-13 19:36 ` Sam Ravnborg
2006-03-13 19:39 ` Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 19 of 20] ipath - integrate driver into infiniband " Bryan O'Sullivan
2006-03-10 0:35 ` [PATCH 20 of 20] ipath - ethernet emulation driver Bryan O'Sullivan
2006-03-10 15:35 ` [openib-general] [PATCH 0 of 20] [RFC] ipath driver - another round for review Michael S. Tsirkin
2006-03-10 16:02 ` Bryan O'Sullivan
2006-03-10 17:48 ` Grant Grundler
2006-03-10 17:54 ` Bryan O'Sullivan
2006-03-10 22:30 ` Grant Grundler
2006-03-11 4:20 ` Bryan O'Sullivan
[not found] <ac5354bb50d515de2a5c.1141922831@localhost.localdomain>
2006-03-09 17:53 ` [PATCH 18 of 20] ipath - kbuild infrastructure Roland Dreier
2006-03-09 18:56 ` Sam Ravnborg
2006-03-09 19:00 ` Roland Dreier
2006-03-09 21:15 ` Sam Ravnborg
2006-03-09 23:24 ` Sam Ravnborg
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=1142277868.9032.14.camel@serpentine.pathscale.com \
--to=bos@pathscale.com \
--cc=akpm@osdl.org \
--cc=bunk@stusta.de \
--cc=davem@davemloft.net \
--cc=gregkh@suse.de \
--cc=linux-kernel@vger.kernel.org \
--cc=openib-general@openib.org \
--cc=rjwalsh@pathscale.com \
--cc=rolandd@cisco.com \
/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
Powered by JetHome