Various Items

nForce2 To Be Improved You can read the details here.

Does it strike anyone as being strange that nVidia and Via are planning on new revisions to their Athlon boards to come out just around when they are supposed to be obsoleted by Athlon64s?

I don’t think that means we’re not going to see Athlon64s and boards sometimes around September; I just don’t think we’re going to supplant the Athlons very soon.

The one tidbit of information we could use at this point would be the initial pricing of Athlon64s. My bet is that the initial 130nm 1Mb cache Athlon64 will carry a premium price far beyond what the average member of this audience will pay. The somewhat later 256Kb cache Mini64s may be somewhat reasonably priced.

I don’t think either is going to convince a sizable percentage of current socket A owners to shift.

Your Last AGP Card By this time next year, PCI-Express will begin to replace AGP. In all likelihood, not too much thereafter, the hottest video cards will be PCI-Express only.

While it is not a good time to buy a reasonably hot video card at a reasonable price, any one who thinks he or she will shell out big time for a card, perhaps to play Doom III, might want to factor this in. No point in buying a $400 one day, and three months later, your system won’t accommodate it.

2.4C or 2.6C?

There’s a more than a bit of forum talk around as to whether it is better to spend another $40 and buy a 2.6C rather than a 2.4C, given that people are having problems reaching FSBs of more than 275MHz.

This makes me bang my head against the wall, because for the life of me, many if not most people in forums these don’t seem to be the least concerned as to why they don’t get past 275MHz, and they’re either too ignorant or lazy to do a few simple things and report on them.

I don’t see how anybody can call themselves an overclocker when they can’t or won’t troubleshoot a situation and determine what specifically is a bottleneck. That’s not an overclocker. That’s an accident hoping to happen.

Anybody faced with this 275MHz problem should do the following:

  1. Lower your memory ratio to 3:2. Don’t argue, just do it. Tighten the timings if you like. See how much you can now increase the FSB. If you now can increase it by more than 6MHz or 7MHz, your memory is the problem. Given the cost of RAM and the relative lack of improvement you’ll get from somewhat faster RAM, it’s a problem many will not want to “fix.”
  2. If you can’t increase the FSB very much, get better cooling on the Northbridge (everyone should do this anyway). People are asking me, “How much improvement can I expect if I do this?” I tell you, “I don’t and can’t know.” Look, I’m not asking you to buy a Prometeia. I’m asking you at most to buy a $15 cooler, or do a little work with something you already have. That makes a lot more sense to do first than spending at least ten times more money on component replacements that won’t work if the Northbridge is too hot (and it does get hot).

  3. If lowering the memory ratio doesn’t work, and better cooling on the Northbridge doesn’t work, the prime culprit becomes the CPU. Before you throw that out, if you’re not using much of a PC cooler, maybe you ought to improve that. To hell with probes and sensors, just touch the damn heatsink while it’s in operation. If you get a near-burn from a cheap heatsink, you need something better. Pull the whole motherboard out (especially if it’s a small case and you don’t have much ventilation inside the case) and see if it runs faster in an open environment. If it does, you need better case cooling.
    If your problem is overheating, buying a new CPU isn’t going to fix that, and it’s probably not going to do any better if that’s the case.

  4. It never hurts to swap components with someone else. If your RAM will only do 215MHz in your system, but 240MHz in someone else’s, your problem isn’t the RAM.

    If you think your problem is, “My system doesn’t run fast enough,” and nothing more specific than that, that isn’t the real problem. The problem is you’re not an overclocker; you’re an accident hoping to happen. An overclocker would troubleshoot and figure out what in his system wasn’t running fast enough before doing something about it, and actually have some good reason to suspect that.

    Overclocking isn’t hard, but it takes some effort and understanding, and willingness to learn something new once in a while. That seems to be increasingly rare these days.

Ed

Be the first to comment

Leave a Reply