From: Dave Jiang <dave.jiang@intel.com>
To: Ed Swierk <eswierk@skyportsystems.com>, linux-ntb@googlegroups.com
Subject: Re: Cabling for NTB between Intel Xeon E5 v4 servers
Date: Tue, 30 May 2017 09:34:02 -0700 [thread overview]
Message-ID: <657df2b1-18f5-9435-6477-98701a1aaa92@intel.com> (raw)
In-Reply-To: <CAO_EM_n81HfXZY-X+OuQWX0T=0E5BYpfJJEf+WQVC1DzyyWkLA@mail.gmail.com>
On 05/26/2017 07:00 PM, 'Ed Swierk' via linux-ntb wrote:
> I'm trying to set up a PCIe link between Intel Xeon E5 v4 (S2600WT)
> servers running Linux 4.10.17, as a testbed for performance
> measurements. My goal is to evaluate feasibility NTB on a custom Xeon
> D-based plug-in card, but as a first step I'm sticking with systems on
> which NTB is already known to (or supposed to) work.
So you'll need PCIe retimer cards. Not redrivers. The hw people
recommended using iDT based card pair. Others may work but have not
tried. I don't know what to recommend because we have an in-house
cabling set that's not commercially available for testing. Also, make
sure your BIOS support NTB and BAR size change and also it turns off SSC
when NTB is on.
>
> I read https://github.com/davejiang/linux/wiki/Intel-NTB-Startup-Guide
> and every other document I could find on Intel NTB, but I'm stuck on a
> basic question: what kind of cable should I use to connect the PCIe slots?
>
> I found some card-and-cable kits
> at http://www.onestopsystems.com/expansion-kits, but these are quite
> expensive ($thousands).
>
> Another idea is to use so-called mini-SAS HD aka SFF-8643, which I
> understand can carry PCIe x4, along with a pair of PCIe-to-SF-8643
> adapters
> (e.g. https://www.startech.com/HDD/Adapters/pcie-nvme-u2-ssd-adapter~PEX4SFF8643).
> The TX-to-RX wiring seems like it should work, but I have no idea about
> signal integrity through such a cable.
>
> There's also the SuperMicro AOC-SLG3-2E4R, which is similar to the
> StarTech card but claims to have redriver circuitry.
>
> What cabling have folks successfully used for NTB between Intel
> Xeon-based servers?
>
> --Ed
>
> --
> You received this message because you are subscribed to the Google
> Groups "linux-ntb" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to linux-ntb+unsubscribe@googlegroups.com
> <mailto:linux-ntb+unsubscribe@googlegroups.com>.
> To post to this group, send email to linux-ntb@googlegroups.com
> <mailto:linux-ntb@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/linux-ntb/CAO_EM_n81HfXZY-X%2BOuQWX0T%3D0E5BYpfJJEf%2BWQVC1DzyyWkLA%40mail.gmail.com
> <https://groups.google.com/d/msgid/linux-ntb/CAO_EM_n81HfXZY-X%2BOuQWX0T%3D0E5BYpfJJEf%2BWQVC1DzyyWkLA%40mail.gmail.com?utm_medium=email&utm_source=footer>.
> For more options, visit https://groups.google.com/d/optout.
next prev parent reply other threads:[~2017-05-30 16:34 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-05-27 2:00 Cabling for NTB between Intel Xeon E5 v4 servers Ed Swierk
2017-05-30 16:34 ` Dave Jiang [this message]
2017-06-01 4:47 ` Ed Swierk
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=657df2b1-18f5-9435-6477-98701a1aaa92@intel.com \
--to=dave.jiang@intel.com \
--cc=eswierk@skyportsystems.com \
--cc=linux-ntb@googlegroups.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