Testing for Part II done

by Anand Lal Shimpi on 4/6/2005 3:27 AM EST
Comments Locked

16 Comments

Back to Article

  • Rand - Wednesday, April 6, 2005 - link

    #12- Thanks, but I've seen the article. It's better then some and has a decent enough description of the implementation but the testing is still largely relegated to the stereotypical PCMark/SiSoft bandwidth tests.
    Nothing on latency, seperating read/write bandwdith, impacts of prefatching or anything else.

    Anand- No need to thank me, I'm merely stating my opinion. In the end it's you that's made me come to said impression, so the thanks should be directed towards you. :)

    I don't envy you in dealing with the numerous complaints/requests and questions of bias you must be constantly inundated with.
    AnandTech is a large and influential site thats bound to draw considerable attention.

    I've only wrote for smaller sites, and even then primarily editorials, and architectural analysis. Not nearly so controversial as benchmarking, but even so I've come to expect a certain number of accusations of bias after any given piece.
  • Anand Lal Shimpi - Wednesday, April 6, 2005 - link

    Rand

    Thanks for that, I appreciate your honesty and your confidence.

    Take care,
    Anand
  • Rand - Wednesday, April 6, 2005 - link

    "but with all of the conspiracy theories bubbling out there I'm sure someone would interpret it as an Intel/NVIDIA forced move :)"

    Frankly I suspect that's bound to happen even if you do include them, any unusual or unexpected results or surprising gains that cause the NF4 Intel Edition to approach AMD where it normally wouldn't will be looked upon with suspicion by some.

    I'd simply try to ignore it and continue on as you do, people will have to decide for themselves whether their willing to put their faith in your integrity.

    Personally I long since decided you were reasonably knowledgeable and legitimately concerned with presenting the most accurate and realistic results you could for readers.

    Ganted, there have been articles I found very questionable and others I simply felt poorly written but the evidence has always led me to feel those rare incidents are more likely due to a rushed article or incompetence (No offence intended, even the best of writers make mistakes) rather then outright deception or bias.

    There are many sites out there and many more reviewers, given time people will come to their own conclusions on whom they feel they can/cannot put their faith in to present valid results.
  • Son of a N00b - Wednesday, April 6, 2005 - link

    w0w, the articles that you are going to try and tackle are exactly what I would like to see.

    I think a nice quick and to the point article of AMD nforce4 vs Intel nforce4 would be suffcient. Just some fps graphs of Halkf Life 2, Doom 3, UT...and your ever so treausred final words...no need to go to in depth right now, wait for some more Dual Core action so you could right an in depth article of AMD dual core SLI vs Intel dual core SLI, and how it compares in extreme computing...
  • biffzinker - Wednesday, April 6, 2005 - link

    In reponse to #2:

    HH had a more in depth preview discussing the memory controller.
    Here's the link if your interested: http://www.hothardware.com/viewarticle.cfm?article...

  • Aquila76 - Wednesday, April 6, 2005 - link

    If you're looking to see how AMD SLI fairs against Intel SLI, Tom's Hardware has compared them, so give Anand a break this review! And yeah, AMD still beats Intel in gaming while Intel wins decoding and stuff (like always).

    http://tinyurl.com/6t2oz
  • Illissius - Wednesday, April 6, 2005 - link

    Yes, I agree... judging by the other articles out there there the nForce4-IE is fast, but not fast enough to make any significant difference in the Intel vs AMD comparison (I'd say just paste in some numbers from older articles to save people the trouble of going back and checking manually, and leave it at that). The only new thing it brings to the table compared to the AMD version is RAID 5, so some numbers on that would be nice.
  • Anand Lal Shimpi - Wednesday, April 6, 2005 - link

    Thanks for the responses - not having to run the AMD numbers will save me a lot of time. That's normally the approach I'd take because it makes sense, but with all of the conspiracy theories bubbling out there I'm sure someone would interpret it as an Intel/NVIDIA forced move :)

    Take care,
    Anand
  • Houdani - Wednesday, April 6, 2005 - link

    Of the three upcoming projects you'd like to tackle, I'd rank the SC:CT guide the lowest and flip a coin between the Win64 and NF4 vs. 955x article.

    For the NF4/955x study:
    AMD numbers would be nice to have as a visual reference, but I wouldn't go too far out of my way to include them. I imagine you'll wind up repeating a bunch of these tests when the dual AMDs arrive, so it's really up to you whether you want to collect data on the single core AMDs now or later.
  • GTaudiophile - Wednesday, April 6, 2005 - link

    #6: Wow, I had forgotten all about VIA! They must be on their way down the drain at this point...in terms of motherboard chipsets. They are quickly becomming irrelevant, IMO.
  • Heron Kusanagi - Wednesday, April 6, 2005 - link

    We all know how AMD works in SLI setups. And it is a power hungry beast...so want to know how Intel SLI setups can hit PSUs...

    Also, OT a bit, but is it possible to speak abt AMD vs Intel vs VIA in terms of Dual Core?
  • GTaudiophile - Wednesday, April 6, 2005 - link

    These must be Dark Days for ATI. All this news about nVidia and Intel and SLI, and they have nothing to respond with at this time. But I will try to remain positive and dream of that not-too-distant day when we see two R520 cards in SLI on an ATI motherboard. ATI had better get that out the door as smooth as nVidia has!
  • HammerFan - Wednesday, April 6, 2005 - link

    IF someone wishes to see how the AMD SLI performs, just point them to the previous reviews. As for me, I agree with Rand's first post, with the ATi+NVidia and 955x+NF4 comparo.

    my $.02
    HF
  • Rand - Wednesday, April 6, 2005 - link

    Another thought just occured to me, I'd be curious to see what the power consumption figures look like for a high end Intel SLI configuration.
  • Rand - Wednesday, April 6, 2005 - link

    I'm not concerned at all with whether you bother throwing an AMD processors in the NF4 SLI Intel Edition review, we already know exactly how AMD's processors perform and how they compare relative to Intel's, where they lead and fall short isn't going to dramatically change due to nVidia's chipset.

    I'm more interested in an indepth comparison of the 955X vs. NF4, most particularly I'd really be interested in getting good impression of the memory controller, most reviews thus far on other sites have done little more examinatio then a simple test of memory bandwidth via SiSoft Sandra/PCMark and nothing else.
    Some look into IDE/SATA controller implementation would be nice as well.

    The only other bit I can add is I'd prefer to see the platforms compared using an ATi graphics card... it's already relatively well known that nVidia's chipsets tend to perform marginally better relative to the competition when using nVidia graphics cards, other reviews have shown that trend to have continued with the NF4 SLI Intel Edition, so I'm more interested in seeing how the NF4 compares without such benefits.
  • TX8 - Wednesday, April 6, 2005 - link

    IF your working on a Splinter Cell Chaos Theory guide your probably wondering why theres no SM 2.0 path, well below is the reason according to one of the mods over at Beyond3D who is contact with the devs of SC:CT

    The games target market is the current generation consoles which are SM 1.1 equivalent so thats the reason why that mode was chosen, the reason there is a SM 3.0 mode in the PC version is because it was a practice run for the next generation consoles which will be powered by SM 3.0 hardware from both ATI & NVIDIA.

    SC:CT wont appear on the next gen consoles but the engine that powers it will be used on those next gen consoles for the next SC game in the series this time using the SM 3.0 path/features etc.

Log in

Don't have an account? Sign up now