Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 30 Jun 2001 12:02:36 -0700
From:      "Crist J. Clark" <cristjc@earthlink.net>
To:        David Leimbach <dleimbac@earthlink.net>
Cc:        questions@freebsd.org
Subject:   Re: Kernel options for faster loopback?
Message-ID:  <20010630120236.E348@blossom.cjclark.org>
In-Reply-To: <20010630103533.A548@mutt.home.net>; from dleimbac@earthlink.net on Sat, Jun 30, 2001 at 10:35:33AM -0500
References:  <20010629114724.A1074@mpi-plusplus.mpi-softtech.com> <20010629235736.C348@blossom.cjclark.org> <20010630103533.A548@mutt.home.net>

next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Jun 30, 2001 at 10:35:33AM -0500, David Leimbach wrote:
> 
> The reason for me asking this is that FreeBSD has really shitty performance
> when it comes to using loopback for any kind of communications.
> 
> Its about 8x faster on Linux to use the loopback device for communication 
> between processes of parallel jobs using an MPI library.

I'd be curious what you are using to benchmark this. I in your
previous mail, you said your ping(1)s on the loopback were
slow. That's what was so funny. OTOH, if you have some meaningful
measurements showing the bandwidth on the loopback is slow, that would
be interesting.

Also, there are better mechanisms for interprocess
communications. Ways to avoid the overhead of the network stack
completely.
-- 
Crist J. Clark                           cjclark@alum.mit.edu

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010630120236.E348>