I?ve learned a lot from this thread, and I think I?m just about ready to purchase my new system. I really like the idea of overclocking the memory and the FSB up to 333. Thanks for pointing that out, Insane 3D. I have a CPU question. A few posts ago, Insane 3D mentioned that a 1600+ could be overclocked to the performance of a 2100+ or 2200+. That sounds good (and inexpensive!). However, I was planning (and had budgeted for) a 2100+ T-bred all along. My question is, as a first time OCer (if I could afford it) should I just get the 2100+? Because I?m concerned about stretching my system too far as a first-timer. Also, Insane3D, may I e-mail you privately when I get all this in check out your mobo BIOS settings vs. mine?
Ok, I'll try to answer this question for you. First off, you need to understand with overclocking, there are no guarantees. There is always the possibility that your new CPU won't o/c as well as you would like. Now, that being said, most of these 1600+ XP's are basically 2100+. Most of them will easily do 1.6ghz - 1.7ghz without any additional voltage. With a kick up of the vcore a bit, it's very common to hit the 1.8ghz level, or basically a 2200+. Let me explain a bit about CPU's first to help you understand.
AMD makes a batch of XP's. They test each one and find out the max speed it will run stable at the default vcore. Some will be 1600+, some will be 2100+, and some won?t work at all and aren't used. This is commonly referred to as yields. This is over-simplied of course, but you get the idea. It's sort of like baking in a way. Say you make a batch of cookie dough. That amount of cookie dough will yield so many cookies, depending on the amount of dough you use in each one. Smaller cookies you will get more, larger cookies you will get less. As you make more and more batches, you use what you learned baking the previous batches to maximize the yields. Anyway, back to the subject at hand.
As they make each batch of processors, they improve the process slightly and improve the yields. As each batch is completed, the amount of CPU's that will run the top speeds is increased, and there are less and less low end or non - functioning ones. What starts to happen is that almost all the CPU's are basically the high end 2100+, but AMD still needs to sell the lower speeds, so they set them to run lower than they will run to fill that market demand. So, basically, almost all of the 1600+ are no different than the high-end 2100+.
Now, another advantage to the 1600+ XP is the low 10.5 multiplier. If you get a 2100+ with a 13 multiplier you have a good deal less headroom without unlocking the CPU and changing the multiplier. If you wanted to run a 166mhz FSB, on a 2100+ that would make the CPU try to run @ 2.16ghz (13 x 166 = 2158mhz), much higher than it would ever run. With a 1600+ XP and a 10.5 multiplier, if you cranked the FSB to 166mhz you would only be running 1.74ghz (10.5 x 166 = 1743mhz), just a tad over a 2100+, but with the extra bandwidth of a 333mhz FSB. So not only would you be saving your money on the much cheaper 1600+, you would be faster than a stock 2100+ due to the faster FSB (266mhz vs. 333mhz).
The Epox board is basically designed to be overclocked, so you are starting out with some very good components. This makes it even more likely that you will hit those speeds. However, like I said before, there are no guarantees. As to your other question, I would be glad to help you set up your new system. Feel free to send me a private message when you get your stuff and I will be glad to help you with any setup or tweaking questions you have.
🙂