ABOUT BENCH FORUMS PODCAST LOGIN REGISTER
PC Components▼
Smartphones & tablets▼
Systems▼
ENTERPRISE & IT
GUIDES▼
DEALS

nForce4 Ultra Roundup: Charting the Mainstream

 by Wesley Fink on July 5, 2005 10:28 PM EST
75 Comments | Add A Comment
75 Comments

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 includes 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 Ultra 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:
Nttcps - m 4,0, -a 4 - l 256000 - n 30000
On the client side (the motherboard under test), we used the following Command Line:
Nttcpr - m 4,0, -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 can reduce CPU utilization below 10%, which is very respectable performance for on-chip gigabit LAN.

Firewire and USB Performance Audio Performance
PRINT THIS ARTICLE
Comments Locked

75 Comments

View All Comments

  • Zebo - Wednesday, July 6, 2005 - link

    Does ECS build EPOX's boards? just curious because they look pretty cheap like ECS IMO..

  • Heidfirst - Wednesday, July 6, 2005 - link

    "[b]#19 and Others - I'm sure you must have noticed that some web sites have never posted a negative review of an Abit board. Also water cooling and asynchronous ram is hardly comparable to our air-cooled tests.

    The first thing I did was check other reviewers and users of the Abit board. The great majority are running into problems at about 250 FSB - although a few are getting better performance. Abit has had so many complaints about the OC performance of this board that I would fully expect a hardware revision on the horizon.[/b]"
    Well the Fatality AN8 SLi, AN8 SLi, AN8 Ultra, AN8 V2.0 & AN8-V are effectively the new revision as I pointed out. Why buy a Fatality AN8 when the AN8 Ultra has better Vcore, better sound & is cheaper?
    & people have had HTT395 & DDR660 out of them on air ...

  • Heidfirst - Wednesday, July 6, 2005 - link

  • Wesley Fink - Wednesday, July 6, 2005 - link

    #19 and Others - I'm sure you must have noticed that some web sites have never posted a negative review of an Abit board. Also water cooling and asynchronous ram is hardly comparable to our air-cooled tests.

    The first thing I did was check other reviewers and users of the Abit board. The great majority are running into problems at about 250 FSB - although a few are getting better performance. Abit has had so many complaints about the OC performance of this board that I would fully expect a hardware revision on the horizon.

    Abit set the expectation that the AN8 Fatal1ty was the best of the best with a price tag to match. It's an interesting board with many interesting features, but it's performance as it now stands is nowhere near the best.

  • TheGlassman - Wednesday, July 6, 2005 - link

    I think that a round up such as this would serve the readers better if three types of memory were used and various dividers used.
    My Chaintech VnF4 is running at 256x11 quite happilly, so I know it will run well over 245 with a divider with my memory, and I'm sure most of the other boards as well. And yes, many people run it over 300 HTT with lower multi cpu's.
    This is not to say that 1:1 testing is not important, but since this is a round up, the various needs and budgets of your readers should be taken into account.
    Seeing bios's used that are dated during the testing, with a known single memory may if repeated cause readers to think that Anandtech doesn't deserve it's well earned reputation as a fair and complete tester of all things important to PC ethusiasts.
    Using memories with 3 different types of chips and using relevant dividers to find maximum HTT's and cpu speeds with each, while being more work, I think will be worth while to your readers, especially in a round up where boards are compared directly to each other.
    This current round up implies that most nF4 boards are not capable of high HTT's, but the truth is you have only shown that most do not run one type of memory at very high speeds. You have not exposed the limits of the boards, nor do we know if the situation is the same with any of the other commonly used memories.

  • bldckstark - Wednesday, July 6, 2005 - link

    Xpose-
    Not too early. I don't have my board yet. I have been waiting on X2. Using your same logic that means that nobody has a board yet right? I mean, since I am the only person I know that is going to build a A64 system soon then I should assume that nobody has one.
    Geez

  • xpose - Wednesday, July 6, 2005 - link

    This Editor's Recommendation for best motherboard is at least 3 months too late. We already have had all of our boards.

    Also, to say that the VN4F Ultra is a bad overclocking board is completely wrong. I have a 3000+ CPU running at 2.67gz now. That is about 49% OC and damn good reguardless of the MB.

  • Son of a N00b - Wednesday, July 6, 2005 - link

    great article! you can clearly see the hours of hard work you put into it. Great Job, I enjoyed it and it was jam packed with info.

    one quuestion though where was the gugabyte board ultra board?? sure you may have reviewed it in the SLI roundup, but then did you not do the same with DFI? Plus you had great results with the reference gigabyte board, but not the revision 1 board...i'd like to see how ir fairs now...maybe i missed something why you reviewed the DFI board again becuase I am not familiar witha ll their variations and naming scheme, but to me it looked the same...why review that one and not the others? sure its great to rehash what a great board the DFi one is but....

    just wondering as I have always had great success with gigabyte boards...but i probably missed something even thought i read it back to front, sorry if i did as i know that you would never do something without a good reason behind it...

    anyway thanks, keep the awesome articles rolling...

  • smn198 - Wednesday, July 6, 2005 - link

    Hi Wesley, thanks for the clarification on the HTT. do you know if it would have any more of an impact when dual core is brought into the equation?

    Thanks again. Good article BTW!

  • BigandSlimey - Wednesday, July 6, 2005 - link

    #18 I really like that idea, would probably be a headache to make it and keep it updated though.
Copyright © 2024. All rights reserved.
BENCH
TOPICS
FOLLOW
ABOUT

Log in

Don't have an account? Sign up now