There is nothing surprising (by the way, and nothing bad) in the fact that the manufacturer of any product considers his product to be the best product of “all time.” Nevertheless, it is useful to present the useful qualities of the product in relation to other similar products. In the case we are considering, this is actually quite difficult to do. The fact is that the lines of AMD and Intel have long diverged. They produce completely different processors. Apparently, different processors and should be used for different applications. Nevertheless, common features are known: Intel produces processors with a higher clock speed, but at equal frequencies the performance of AMD processors is still higher. It is already considered virtually the norm to compare AMD processors with a frequency of 1.3 … 1.4 GHz and Intel processors with a frequency of 1.7 GHz. In most cases, these are really the same processors in terms of performance. In some applications, however, at approximately equal frequencies Pentium 4 “outperforms” Athlon, but in some other applications the Athlon processor with a frequency of 1.4 GHz “outperforms” Pentium 4 with a frequency of 2 GHz.
Nevertheless, in spite of any tests, the owner of the 1.4 GHz Athlon processor in the shower probably feels some discomfort when he finds out that his friend has a Pentium 4 processor running at 1.7 GHz.
To reduce this discomfort, the Athlon XP processor designation indicates not the processor clock speed, but the so-called model number – the specific model number in the general Athlon processor family. For example, Athlon XP 1500+, Athlon XP 1800+.
AMD was also forced to take this step by the fact that the Athlon XP processor is undoubtedly more productive than previous Athlon models. However, how to properly compare a processor with a Thunderbird core operating at 1.4 GHz and a processor with a more advanced Palomino core but operating at 1.33 GHz?
Of course, we have already encountered this designation technique before. This is the notorious PR – Pentium Rating. At one time, AMD released processors that at the same clock speed were more efficient than Pentium processors. True, only on fixed-point operations. Then this PR was introduced. The lag in floating point operations was carefully hushed up. In general, there is some logic in this, but real applications rarely do not use floating point operations, and everyone quickly figured out that PR does not correspond to the real processor performance ratio.
For Athlon XP processors, the choice of model number values has been approached much more carefully. Undoubtedly, the value of model number is tied to a certain value of the clock frequency of Pentium 4 processors. However, in this case, the value of model number, for example, 1500+ means that in the vast majority of applications this processor will not yield to Pentium 4 with a clock frequency of 1.5 GHz. But they do not hide from us that the processors are so different that you can find such an application in which the Pentium 4 turns out to be faster.
Therefore, the value of model number must be treated exactly as it is interpreted: as the model number in the series of Athlon processors. By the way, at the same time, it is completely optional that an Athlon XP processor with a model number equal to 1500+ will be faster than a regular 1.4 GHz Athlon processor.
To find real differences, the Athlon XP 1500+ processor was tested in the EPOS research department in comparison with the capabilities of the Pentium 4 (1.5 GHz) and Athlon (1.4 GHz) processors.
The Athlon XP 1500+ processor is clocked at 1333 MHz. However, he inherited from his predecessor all the best qualities. Improvements made in the Palomino core are mainly aimed at accelerating the streaming processing of large data arrays – in this area the Pentium 4 positions are quite strong. Therefore, it is clear that the Athlon XP 1500+ in the vast majority of cases should be faster than the Pentium 4 1.5 GHz processor. This was confirmed, even in the most difficult tasks, the Pentium 4 lag turned out to be very significant. So, in particular, in fig. Figure 4 shows the values of the rendering time in the 3DMax 3.1 environment of one frame from a real task.
Of course, such a result could be foreseen in advance. To overtake even the regular Athlon, the Pentium 4 must have a clock frequency of at least 1.7 … 1.8 GHz. So, indeed, AMD when choosing the model number value does not deceive us, and does not deceive us with a large margin.
Of much greater interest is the question: “But how did the performance increase compared to previous Athlon models?”
Of course, one should not expect a big superiority of the Athlon XP 1500+ processor over the 1.4 GHz Athlon processor. After all, the basis of high performance is still the computing capabilities of the processor and coprocessor, and they are the same for both models. Improved only the ability to stream large data arrays.