[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: GigE recommendations
- To: misc_(_at_)_openbsd_(_dot_)_org
- Subject: Re: GigE recommendations
- From: Michael Shalayeff <mickey_(_at_)_lucifier_(_dot_)_net>
- Date: Mon, 8 Nov 2004 09:50:40 -0500 (EST)
- Cc: misc_(_at_)_openbsd_(_dot_)_org
Making, drinking tea and reading an opus magnum from Henning Brauer:
> * Breen Ouellette <the_(_dot_)_man_(_at_)_breeno_(_dot_)_net> [2004-11-06 21:50]:
> > Theo de Raadt wrote:
> > >The OpenBSD development networks (ie. my basement) are are mostly
> > >switched gigabit. i386, amd64, alpha, macppc, sparc64, and perhaps
> > >even something else, are gigabit. Using sk, em, bge, gem, and I dunno
> > >what else. Other developers in the group are using the other gigabit
> > >cards in their networks.
> > >Of course it works. It's not non-free-firmware locked Windows-only
> > >wireless crap.
> > I realize that they don't have any firmware problems. Let me clarify my
> > question.
> > With the recent posts about NFS problems being related to particular
> > NICs, I want to know if anyone has experience with the Intel GigE NICs
> > behaving badly under any particular applications. I am particularily
> > interested in the Intel PRO/1000MT Desktop Adapter (PWLA8390MT) and the
> > Intel PRO/1000MT Dual Port Server Adapter (PWLA8492MT). These use (em),
> > btw. They seem like good cards, I just want to be sure before I buy 20
> > and 2 of them, respectively.
> I haven't seen the specific em(4) model making a difference - I didn't
> really bench, tho.
82542/543/544 chips lack some interrupts coalescing features
although on defaults it shall not differ much from aother
chips unless some alyte tweaking is done in the sources (:
paranoic mickey (my employers have changed but, the name has remained)
Visit your host, monkey.org