Tuning Gigabit

[Nick Evans]

On Sun, 29 Jun 2003 07:58:27 -0400
Christian Brueffer <chris at unixpages.org> wrote:

> On Sat, Jun 28, 2003 at 10:05:56PM -0400, David Gilbert wrote:
> > >>>>> "Gregory" == Gregory Sutter <gsutter at zer0.org> writes:
> > 
> > Gregory> Will you please summarize the motherboard performance data so
> > Gregory> we know which boards to buy and which to skip?  Thanks.
> > 
> > I've been working on such a summary.  So far, the 'nvidia' chipset
> > boards have all tested badly.  They couldn't be coaxed to pass more
> > than 100meg of traffic by any means we could discern.
> > 
> > The K7S5A has been our mainstay.  Many of them are DOA, but the ones
> > that pass a couple weeks of cpuburn (see port) ... both on cpu and
> > memory tests ... work amazingly well.  These boards are limited to 300
> > megabit total thruput by being a 33Mhz 32bit PCI bus.
> > 
> > We've been testing mainly Athlon boards ... we havn't seen good P4
> > boards ... but most of the boards we've had through for the P4 have
> > been workstation and not server boards.
> > 
> > The tiger tyan MPX is a dual board with 64 bit slots.  I havn't had
> > time to fully benchmark it becuase we use it as a fairly primary
> > database server ... but it has generally been able to perform at or
> > near the top of the class.
> > 
> > There is an ASUS dusl board with 32-bit only slots and the AMD 76x
> > chipset (unfortunately it's far away and I can't look at it).  it's
> > 32-bit slots run at 66Mhz and have extrodinarily good thruput.
> > AFAICT, it's currently out of production ... but the dual board on the
> > ASUS site looks very good.
> > 
> 
> If you're talking about the ASUS A7M266-D, it actually has two 64bit
> PCI slots (powering a SCSI controller and a NIC here).
> It runs extremely well so far.
> 
> - Christian
> 
> -- 
> Christian Brueffer	chris at unixpages.org	brueffer at FreeBSD.org
> GPG Key:	 http://people.freebsd.org/~brueffer/brueffer.key.asc
> GPG Fingerprint: A5C8 2099 19FF AACA F41B  B29B 6C76 178C A0ED 982D

I recently tested a Compaq DL380, 733Mhz processor, 512meg ram with two fiber 64-bit Intel Pro 1000's. The system (4.8-R) forwarded 500 megabits at about 20% idle with IPFilter loaded but with no rules. The system uses a ServerWorks chipset and has several 64-bit pci slots. I used netperf on several systems on either side of the firewall. In the next week or so I'll be testing two Pentium 4 2.4 gig systems with ServerWorks chipsets using a much larger cluster of load applying systems. I'll post those results.

When I tested the same configuration on a Compaq 1850R (450Mhz, i440BX, 32-bit PCI) I could only muster about 250-270 megabits of traffic. I figured the PCI bus was the limiting factor and stepped up to the DL380.

Previously I've forwarded 50,000 packets per second on a Celeron 900Mhz, Intel 810 chipset, 32-bit PCI (4.4-R) system with IPFilter loaded and IIRC 150 rules in effect. The system used Intel PRO/100S dual port server adapters. I used ping -f for those tests.

I have access to about 25-30 systems to apply load with so if anyone wants something specific tested I might be able to work it into my schedule.

Nick

-------------------------------
nick.evans
network.engineering
talkpoint communications, inc.
land 212.909.2967
cell