What Will It Get Me, And What Will It Cost Me?

There’s a review over at Tom’s Hardware in which a 3.06GHz was coaxed into 4GHz.

Per the rather heavily self-congratulatory tone, I’ll just say that it’s a lot nicer when other people say such things.

P.S. Expert overclockers use liquid nitrogen, not liquid hydrogen. If you have an accident with liquid nitrogen, you’ll just lose chunks of your body. Have an accident with liquid hydrogen, you lose chunks of your neighborhood.

I know, they sound so alike, but this is like a doctor confusing iodine and cyanide. 🙂

Now, seriously . . . .

The results achieved probably indicates about the best you can expect from any reasonably-priced Intel-based system towards the end of 2003, and probably an overoptimistic one at that. (Prescotts towards the end will probably do a little better, but probably won’t be affordable until 2004).

Actual MHz will probably be a bit lower, and memory speeds will likely be a bit higher.

The effort and expense involve will change dramatically as the year progresses. Today, you’re looking at spending well close to two thousand dollars on CPU/dual DDR motherboard/dual DDR sticks and cooling system. By next December, the cost to get close will be more in the range of $500+.

What Will It Get Me, And What Will It Cost Me?

When you look at your favorite benchmarks, don’t just look at the top one. Look at a few others.

For those of you AMD-inclined, look at the Athlon 2700+ and 2800+ figures. This will give you a good idea of what to expect from a cheap TBredB shortly, for about $70. Is the difference between it and the top worth the additional expense, effort and/or wait?

For those of you who are Intel inclined, look at the P4 2.8-3.06 figures without hyperthreading. This is roughly what you can expect out of P4 system now and in the near future. Is the difference between it and the top worth the additional expense, effort and/or wait?

Finally, look at the chart to see where you are right now. Is the jump from where you are to where either of these will get you worth whatever the expense or effort will be to you? Keep in mind that in most situations, you might notice a 15% difference, and it takes a whole lot more than that for it to become immediately obvious.

This is a good way to use any multiprocessor comparison to buy wisely.

This is not a good way to glory, but if that’s what you think you’ll get, let me ask you before you buy: Who had the fastest computer wherever you electronically hang out six months ago? Three months ago? A month ago? A week ago? Now?

Don’t remember? Then why expect anyone else to? Nothing wrong with pursuing your hobby, but don’t think you buy long-time fame. All that does is make you a legend in your own mind.

The Next Jump

By 2004, machines are going to be more than a bit different than they are now. We’ll be looking at .09 micron chips, dual DDR-II memory, SATA hard drives, gigabit Ethernet connections, faster PCI busses, maybe even Wi-Fi built in. In other words, a different platform.

In 2003, we’ll be on the back end of .13 micron CPUs and transitioning from what essentially is still recognizably the descendant of the Pentium II-class mobo to a new generation.

Any purchases you make should keep this in mind. That sort of wholesale upgrade is going to cost you, so better to get the old stuff up to speed as cheaply as you can, and save the major spending for 2004.


Be the first to comment

Leave a Reply