Slashdot Mirror

← Back to Stories (view on slashdot.org)

Socket-A Chipset Roundup

Posted by CmdrTaco on from the little-bit-of-this-a-little-bit-of-that dept.

EconolineCrush writes: "The Tech Report has a review up of VIA's new KT333 Socket A chipset. Though it's really a review of the KT333, a total of seven different chipsets from VIA, SiS, AMD, and NVIDIA are compared to determine the uniprocessor Socket A performance king. This is definitely worth checking out if you're in the market for an AMD platform, or are curious to see how your current chipset stacks up against the latest and greatest."

4 of 144 comments (clear)

  1. Stability, or performance? by clockwork18 · · Score: 3, Insightful

    Perhaps they should be reviewing the total stability of any given solution. I for one know that I used to have an Iwill board based on the ALi MaGIK chipset, and it was a hell of a lot more stable than my current KT266A.

    1. Re:Stability, or performance? by syzxys · · Score: 3, Insightful

      I agree completely about VIA's stability problems. Out of all the computers I've had based on VIA chipsets (about 4 or so), the only one with a decent amount of stability was the ASUS A7M266. Except for that one oddball stable board, it's been a horror story of:

      Personally, I'm terrified of VIA chipsets at this point. I like the AMD 760MP much better. :-)


      ---
      Windows 2000/XP stable? safe? secure? 5 lines of simple C code say otherwise!
    2. Re:Stability, or performance? by beanball75 · · Score: 2, Insightful

      Well, considering the people currently writing reviews are using reference designs, I think full scale stability testing is premature.

  2. Re:Upgradeable Chipsets? by syzxys · · Score: 3, Insightful

    Why can't someone produce motherboards wherein components other than the CPU are quickly upgradeable?

    Wow, that's a good idea. I wonder how the cost of manufacturing for, e.g. a ZIF socket + chip compares to just surface-mounting or whatever the motherboard makers do now. (Although, actually I think at least one of my current motherboards has the southbridge in a socket).

    I see two problems though:

    1. How big of a target market would there be, initially? I mean, sure probably 50% of the people on /. would be in the market, but what about the big OEM's? I don't see "consumers" going to the local CompUSA to upgrade their northbridge, I mean hell, (a) most people don't even know what a "northbridge" is, (b) the CompUSA folks would be more likely to tell them they just needed a new motherboard or new computer. "Oh yeah, your power switch is broken, better upgrade your case, motherboard, and processor while you're at it." :-)
    2. The programming interface to each chipset is proprietary. Ever try to get information about the registers on a VIA chipset? They want you to sign a giant NDA just to look at the specs PDF, for crying out loud! (Although last time I looked, some people had slipped up and posted NDA'd VIA specsheets where google could find it. Shh, don't tell anyone. :-) And since it's proprietary, that means the developers are used to being able to change it whenever they want. IOW, there's no engineering pressure to make things backward-compatible, because the only software that is affected is the BIOS. So, you could say, just distribute BIOS images with the new chipsets. But how many tech support calls do you think they're going to get when people accidentally plug in the new chipset with ACPI power off registers are in the same place the DRAM timing registers used to be? (so the board won't power on anymore).

    So in other words, I think it's a great idea, but there's no way the chipset companies are going to have it while they're still acting like it's the 1950's and every single chipset is (a) proprietary, (b) guarded like it's the secret to eternal life or something. Oh well, we can always dream. :-)

    ---
    NEW! Crash Windows NT/2000/XP from any account using only printf!