From: Ingo Molnar <mingo@kernel.org>
To: Christoph Hellwig <hch@lst.de>
Cc: "Elliott, Robert (Server Storage)" <Elliott@hp.com>,
"linux-nvdimm@ml01.01.org" <linux-nvdimm@ml01.01.org>,
"linux-fsdevel@vger.kernel.org" <linux-fsdevel@vger.kernel.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"x86@kernel.org" <x86@kernel.org>,
"ross.zwisler@linux.intel.com" <ross.zwisler@linux.intel.com>,
"axboe@kernel.dk" <axboe@kernel.dk>,
"boaz@plexistor.com" <boaz@plexistor.com>,
"Kani, Toshimitsu" <toshi.kani@hp.com>
Subject: Re: another pmem variant V2
Date: Thu, 2 Apr 2015 20:03:30 +0200 [thread overview]
Message-ID: <20150402180329.GA8503@gmail.com> (raw)
In-Reply-To: <20150402164139.GA9774@lst.de>
* Christoph Hellwig <hch@lst.de> wrote:
> On Thu, Apr 02, 2015 at 03:11:36PM +0000, Elliott, Robert (Server Storage) wrote:
> > Attr Copy Read IOPS Write IOPS
> > ==== ==== ========= ==========
> > UC memcpy 36 K 22 K
> > UC NT rd,wr 513 K 326 K
> >
> > WB memcpy 3.4 M 2.5 M
> > WB NT rd,wr 3.3 M 3.5 M
> >
> > WC memcpy 776 K 3.5 M
> > WC NT rd,wr 3.0 M 3.9 M
> >
> > WT memcpy 2.1 M 22 K
> > WT NT rd,wr 3.3 M 2.1 M
> >
> > a few other variations yielded the peak numbers:
> > WC NT rd only 3.2 M 4.1 M
> > WC NT wr only 712 K 4.6 M
> > WT NT wr only 2.6 M 4.0 M
> >
> > There are lots of tuning considerations for those memcpy
> > functions - how far to unroll the loop, whether to
> > include PRFETCHNTA instructions, etc.
>
> Looks like we should aіm for WC + NT would be a good start.
>
> Can you prepare a patch to add your NT memcpy variants and
> a second one to use them in the pmem driver?
So we already have NT memcpy variants, see the copy_*_user_*_nocache()
primitives in arch/x86/. They could be used almost straight away for
kernel memory as well, as kernel buffers will not fault.
Thanks,
Ingo
next prev parent reply other threads:[~2015-04-02 18:03 UTC|newest]
Thread overview: 69+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-03-26 8:32 another pmem variant V2 Christoph Hellwig
2015-03-26 8:32 ` [PATCH 1/3] pmem: Initial version of persistent memory driver Christoph Hellwig
2015-03-26 14:12 ` [Linux-nvdimm] " Dan Williams
2015-03-26 14:35 ` Christoph Hellwig
2015-03-26 21:37 ` Ross Zwisler
2015-03-26 14:52 ` Boaz Harrosh
2015-03-26 15:59 ` Dan Williams
2015-03-26 8:32 ` [PATCH 2/3] x86: add a is_e820_ram() helper Christoph Hellwig
2015-03-26 9:02 ` Ingo Molnar
2015-03-26 9:34 ` Christoph Hellwig
2015-03-26 10:04 ` Ingo Molnar
2015-03-26 10:19 ` Christoph Hellwig
2015-03-26 10:28 ` Ingo Molnar
2015-03-26 10:29 ` Christoph Hellwig
2015-03-26 15:49 ` Boaz Harrosh
2015-03-26 16:02 ` [Linux-nvdimm] " Dan Williams
2015-03-26 16:07 ` Boaz Harrosh
2015-03-26 16:43 ` Christoph Hellwig
2015-03-26 18:46 ` Elliott, Robert (Server Storage)
2015-03-26 19:25 ` [Linux-nvdimm] " Dan Williams
2015-03-26 20:53 ` Ross Zwisler
2015-03-26 22:59 ` Yinghai Lu
2015-03-27 8:10 ` Christoph Hellwig
2015-03-26 8:32 ` [PATCH 3/3] x86: add support for the non-standard protected e820 type Christoph Hellwig
2015-03-26 16:57 ` another pmem variant V2 Boaz Harrosh
2015-03-26 17:02 ` [PATCH] SQUASHME: Streamline pmem.c Boaz Harrosh
2015-03-26 17:23 ` Christoph Hellwig
2015-03-26 22:17 ` Ross Zwisler
2015-03-26 22:22 ` Ross Zwisler
2015-03-26 23:31 ` [Linux-nvdimm] " Dan Williams
2015-03-31 13:44 ` Boaz Harrosh
2015-03-26 17:18 ` another pmem variant V2 Christoph Hellwig
2015-03-26 17:31 ` Boaz Harrosh
2015-03-26 18:38 ` Christoph Hellwig
2015-03-31 9:25 ` Christoph Hellwig
2015-03-31 10:25 ` Boaz Harrosh
2015-03-31 10:31 ` Boaz Harrosh
2015-03-31 14:21 ` [RFC] SQUASHME: pmem: Split up pmem_probe from pmem_alloc Boaz Harrosh
2015-03-31 16:10 ` Christoph Hellwig
2015-03-31 16:08 ` another pmem variant V2 Christoph Hellwig
2015-03-31 13:18 ` [SQUASHME 0/6] Streamline of Initial pmem submission Boaz Harrosh
2015-03-31 13:23 ` [PATCH 1/6] SQUASHME: Don't let e820_PMEM sections Boaz Harrosh
2015-03-31 17:16 ` [Linux-nvdimm] " Brooks, Adam J
2015-03-31 13:24 ` [PATCH 2/6] SQUASHME: pmem: Remove getgeo Boaz Harrosh
2015-03-31 13:25 ` [PATCH 3/6] SQUASHME: pmem: Streamline pmem driver Boaz Harrosh
2015-03-31 13:27 ` [PATCH 4/6] SQUSHME: pmem: Micro cleaning Boaz Harrosh
2015-03-31 15:17 ` [Linux-nvdimm] " Dan Williams
2015-03-31 15:24 ` Boaz Harrosh
2015-03-31 15:30 ` Dan Williams
2015-03-31 15:43 ` Boaz Harrosh
2015-03-31 19:40 ` Matthew Wilcox
2015-03-31 13:28 ` [PATCH 5/6] SQUASHME: pmem: Remove SECTOR_SHIFT Boaz Harrosh
2015-03-31 13:33 ` [PATCH 6/6] SQUASHME: pmem: Remove "... based on brd.c" + Copyright Boaz Harrosh
2015-03-31 15:14 ` another pmem variant V2 Boaz Harrosh
2015-03-31 16:16 ` Christoph Hellwig
2015-03-31 16:44 ` Ingo Molnar
2015-03-31 17:24 ` Christoph Hellwig
2015-03-31 17:33 ` [Linux-nvdimm] " Dan Williams
2015-04-01 7:50 ` Ingo Molnar
2015-04-01 8:06 ` Boaz Harrosh
2015-04-01 12:49 ` Boaz Harrosh
2015-03-31 22:11 ` Elliott, Robert (Server Storage)
2015-04-01 7:26 ` Christoph Hellwig
2015-04-02 15:11 ` Elliott, Robert (Server Storage)
2015-04-02 16:41 ` Christoph Hellwig
2015-04-02 18:03 ` Ingo Molnar [this message]
2015-04-01 19:33 ` Elliott, Robert (Server Storage)
2015-04-02 9:37 ` Christoph Hellwig
2015-03-26 18:38 Christoph Hellwig
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=20150402180329.GA8503@gmail.com \
--to=mingo@kernel.org \
--cc=Elliott@hp.com \
--cc=axboe@kernel.dk \
--cc=boaz@plexistor.com \
--cc=hch@lst.de \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-nvdimm@ml01.01.org \
--cc=ross.zwisler@linux.intel.com \
--cc=toshi.kani@hp.com \
--cc=x86@kernel.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).