Re: hardware question
Hi David,
I am not the best person advising you but i am into a similar situation. I am
using ..
Dual 733MHz Pentium III processors
1 GB memory
5 PCI slots (three 64 bit and two 32 bit),
2 Intel 1000BT dual Ethernet cards ( giving 4 ethernet ports on each machine )
on
the 64 bit PCI bus
running on FreeBSD 4.9
the max i can send out on the gigabit link is 750Mbps. Infact even when i try
to use all the four interfaces , the aggregate does not go above 750Mbps. Now
is this a limitation of the 64 bit PCI bus ( which i seriously doubt ) or
something else , i am not able to figure out.
This should give you some insight on the kind of hardware ( atleast the PCI
bus ) that u should be using if you want to saturate that link/use two cards.
- Nirav.
----- Original Message -----
From: David <dpnsw --at-- tpg.com.au>
Date: Sunday, February 15, 2004 11:21 pm
Subject: hardware question
> My appologies for what may seem like a pretty dumb question but it doesn't
> seem to be covered anywhere I have looked.
>
> While I understand iperf can help me to optimise an existing link and do
> some
> measurement on it I am comming at this from the other angle.
>
> I have an optical Gig E transport that I need to fully load and get some
> sort
> of measuement back on delay, jitter and packet loss. I gather therefore I
> need simply to run iperf sending UDP packets and set the bandwidth to 1
> gig.
> The test will be run between 2 points which preferably would be the same
> machine.
>
> My question is about what sort of hardware do I need to be able to do this.
> What sort of CPU do you need to be able to saturate an optical GigE link
> and
> if I wanted to put 2 cards into one computer would the PCI bus be able to
> handle this?
>
> While I suspect that GNU/Linux may be a better choice to run ipref on in
> such
> a case is it likely that win32 would be able to do the same thing?
>
> I ask this question here because I figure that if anybody would know how
> to do
> this it would be the people developing and using iperf.
>
> Many thanks.
> --
> Best Regards,
>
> David Price
>
>
>