Ethernet Performance

We have often talked about the throughput advantage of PCI Express compared to PCI and its impact on Gigabit LAN performance. To show you this, the new motherboard test suite needed to include LAN performance measurements.

The Windows 2000 Driver Development Kit (DDK) includes a useful LAN testing utility called NTttcp. We used the NTttcp tool to test Ethernet throughput and the CPU utilization of the various Ethernet Controllers used on the nForce4 SLI motherboards.

We set up one machine as the server; in this case, an Intel box with an Intel CSA Gigabit LAN connection. Intel CSA has a reputation for providing fast throughput and this seemed a reasonable choice to serve our Gigabit LAN clients. At the server side, we used the following Command Line as suggested by the VIA whitepaper on LAN testing:
Ntttcps -m 4,0,‹client IP› -a 4 -l 256000 -n 30000
On the client side (the motherboard under test), we used the following Command Line:
Ntttcpr -m 4,0,‹server IP› -a 4 -l 256000 -n 30000
At the conclusion of the test, we captured the throughput and CPU utilization figures from the client screen.

Ethernet Throughput

Ethernet Overhead

While Marvel PCI has a bit lower overhead, the throughput for Gigabit LAN on PCI is almost 30% lower than Gigabit on PCI Express. The nVidia on-chip PCIe LAN also exhibits lower CPU utilization than the full Gigabit LAN on PCIe. It has also been reported that enabling nVidia Active Armor lowers CPU overhead even further.

All Ethernet tests were performed with standard frames, but Gigabit Ethernet supports Jumbo frames as well. Jumbo frames will theoretically provide a further reduction in CPU overhead. We have seen test results that show the combination of Active Armor and Jumbo Frames, which can reduce CPU utilization below 10%, a very respectable performance for on-chip gigabit LAN.

Firewire and USB Performance Audio Performance
Comments Locked

108 Comments

View All Comments

  • eva2000 - Monday, February 28, 2005 - link

    whoops no mention of psu was used in system config listing but didn't read till page 20 of the review it mentions OCZ 520W PS psu heh
  • Slaimus - Monday, February 28, 2005 - link

    They used a OCZ 520W.
  • neologan - Monday, February 28, 2005 - link

    I think the test results for 3dmark2003 single and SLI are the wrong way around?

    http://www.anandtech.com/mb/showdoc.aspx?i=2358&am...
  • F4810 - Monday, February 28, 2005 - link

    Why did they show no benchmarks with the mobo's overclocked? It doesnt make sense to say these boards are better becuase you can clock the memory higher if the overall CPU clock is roughly the same. The reason they dont show you is that due to the onchip memory controller on the AMD chips, the high memory frequency doesnt make much of a difference at all in real world terms. As long as you can clock the CPU high that is all that really matters. Also they dont take into account cost as some boards cost 50% more that the others.
  • dornick - Monday, February 28, 2005 - link

    I was considering jumping on the SLI bandwagon until I had some sense knocked into me.

    I'd like to see a comparison of the Ultra chipset MBs, including the Chaintech, Epox, etc... since that's where I think the nForce 4 market will go.
  • eva2000 - Monday, February 28, 2005 - link

    missing one vital piece of info, what PSU you used hehe
  • Spacecomber - Monday, February 28, 2005 - link

    Thanks for doing the indepth analysis of this new chipset and how it is being implemented by some major motherboard manufacturers.

    There was one detail that I was hoping to see some reference to. I understand, from a friend, who has the MSI SLI motherboard, that the Creative Live sound chip only works if your power supply has a -5 volt connector on it. It looks to me like the OCZ power supply that you used has this, but many of the new power supplies, such as the Enermax v2.0 power supplies, no longer have a -5 volt connection. He was using a Enermax 535 watt Whisper II (SLI ready) when he ran into this issue. This kind of compatability problem slipping through QA seems like another indication that the everyone was in a big hurry to get these motherboards to market, maybe before they were thoroughly tested.

    Space
  • Regs - Monday, February 28, 2005 - link

    I liked the subtle hints Wes.

    In your final words you stated, "If you want the best performance possible then the answer would likely be yes". Then how would this apply for users getting two 6600GTs?
  • xsilver - Monday, February 28, 2005 - link

    great article
    minor gripe -- the overclocking "graphs" are useless -- what would be better is the resulting fps of overclocking to show people if its worth it to get that extra xxx fps
  • arfan - Monday, February 28, 2005 - link

    now i am waiting ultra mobo benchmark. What about the price of all this mobo ? i fell disappointed with msi doesn't have PCI 1x. (Sorry my english language is very bad)

Log in

Don't have an account? Sign up now