Quote:
Originally posted by drdingo21
p4s suck because they hardly run at their rated speeds. for example my 1.2 tbird is faster than a p4 1.6 gig. My other cpu xp1600 is running at 1900 mhz and its stomping p4 2gig. If youd like i could throw up some benchamarks to further illustrate. The only reason pentium 4s are so popular is because they advertise and every big company picked them as their main cpus. You can buy a Xp2100, motherboard, a 32x cdrw, case and a stick of 256 mb pc133 ram for under the price of what you can get a p4 processer itself. And the amd machine will be every bit as fast as the pentium.
|
No, they do run at their rated speeds. In fact, AMD systems do not run at their rated speeds. A 2 GHz P4 will run at 2 GHz. An xp1600 doesn't actually run at 1.6 GHz, it runs at a slower clock. It is called an xp1600 because even at its slower clock it will keep up with a comparable Intel running at 1.6 GHz....now whether they compare to P3 or P4, I don't know.
P4s don't exactly suck. They have a very deep pipeline, thus meaning they were designed to run at high frequencies. This is partially due to Intel's whole marketing strategy of "more hertz equals better." The P4 has fallen victim to their very own ad campaign, because it was designed to run at very high clocks but doesn't deliver the clock-for-clock performance as some other chips. This is a side-effect of deepening the pipeline where you reduce cycle-efficiency but increase clock frequency. Increased clock frequency can translate to higher throughput. Of course there are drawbacks to higher frequency, such as the penalty paid with a cache miss, branch misprediction, or any other reason you need to flush the pipeline.
Of course another factor to consider with a P4 is that it has a slightly different microarchitecture than a P3. Not everything has been compiled and optimized for a P4, whereas much has already been optimized for P3s. If the code isn't compiled and optimized for the P4 then it won't unleash the full potential of the processor. Your Windows kernal, for example, is not optimized for the P4. I'm not sure how much is optimized for AMD's xp line of processors, but my guess is that it's somewhat comparable to the P3 optimization proliferation.
Benchmarks are, all things considered, generally meaningless. They are the tools that companies use to advertise their products, but they are sometimes easily cheated. Companies will do very much just to pump up their benchmark numbers. Benchmark performance doesn't translate well to real-world performance. For example, a benchmark could be optimized for a particular processor, giving weight to one over another. Or the benchmark could be optimized for each different processor, but real-world apps aren't always optimized, so the benchmark could easily report higher scores than you'd really experience when using the system.
I wouldn't say the P4 is a crap processor. Just don't fall victim to Intel's ad campaign and believe that processor performance is tied directly to frequency. There are many factors that contribute to processor and system performance. Just because a processor runs at a higher frequency but doesn't have the same throughput as another processor running at a lower frequency doesn't mean it's crap.
b