Preamble

Today Intel Russia shared some information about the forthcoming changes in the systems of names of its processors. The information provided is official but preliminary, and the final conception can be changed. 

The key change is the lack of clock speeds in names of Intel's future CPUs. Actually, it will be replaced with a so called processor number.

The official version is that the extension of lines, especially in the desktop sector, and availability of processors based on different architectures and even cores make the situation pretty confusing. One of the examples is 4 modifications of the Pentium 4 CPU clocked at 2.8 GHz:

1. Pentium 4 2.8 GHz (Northwood, 512 KB L2, 533 MHz FSB)
2. Pentium 4 2.8A GHz (Prescott, 1 MB L2, 533 MHz FSB)
3. Pentium 4 2.8C (Northwood, 512 KB L2, 800 MHz FSB, Hyper-Threading)
4. Pentium 4 2.8E (Prescott, 1 MB L2, 800 MHz FSB, Hyper-Threading)

The difference in their parameters is great, therefore, they might have different performance, price and an area of application. Sometimes it's also used by hardware sellers to play tricks with buyers. That is why from Intel's point of view the current situation makes it necessary to redefine the conception of naming its processors in order to prevent possible mistakes and makes processor names simple and logical from an end-user's standpoint. 

By introducing the processor number Intel is bringing its system of names to the standard practice successfully used by other manufacturers of consumer devices. Digital cameras, cars, cellphones, audio and video equipment do not contain any signs of the number of megapixels, engine power or acoustic system power. These parameters are not hidden but the names of the products might indirectly indicate them. This appeal sounds very interesting: by comparing processors and digital cameras and even cellphones Intel admits that these products belong to the same market category. Cameras, audio centers... and processors! This is the consumer electronic equipment of XXI century...

Details and explanation 

But it doesn't mean that Intel is going to cancel its system of popular brands like Pentium 4, Pentium M, Celeron and others. All their (tm) remain as is, and the trade names even get some extra weight: by replacing the clock speed with an abstract identifier Intel focuses our attention on the shared name for a given CPU class.

As a result, Intel's new system of CPU naming follows this scheme:

1. Main brand (Intel)
2. Line brand (for example, Pentium M Processor)
3. Series (the first figure in the processor number, for example, 7xx)
4. Two more figures which together with the first one form a unique identifier of a certain model

CPU names will look like Intel Pentium M Processor 735. They underline that all data on the key features and architectural peculiarities will be provided both at Intel's site, in documentation and on the CPU box. The fact that the clock speed won't be indicated in CPU names doesn't means that it won't be indicated at all. And it also refers to the bus clock, cache size and so on. The clock speed will be indicated in small font for those who are interested in it.

The processor number implies a certain combination of parameters, and currently, the list consists of the following ones:

1. Architecture
2. Core clock
3. FSB clock
4. CPU cache sizes
5. Future Intel Technologies

The items 1 to 4 are obvious, and the last one is of the most interest. I'm nearly sure that this auxiliary item is left for processors with Vanderpool and LaGrande support. I.e. some hypothetic Pentium 4 556 can have only one difference from Pentium 4 555 - and this is the Vanderpool support.

However, the main point is very clear: within one CPU brand a processor number will be referred to a single model with a fixed set of features. The idea is actually to prevent situations like that with 4 modifications of the Pentium 4 2.8 GHz. By the way, it defines one more significant difference between the name systems of Intel and AMD: AMD can give the same name to several chips with different features. For example, Athlon XP 2600+ covers three models: 2133 MHz / 266 FSB / 256 L2, 2083 MHz / 333 FSB / 256 L2 and 1917 MHz / 333 FSB / 512 L2. 

At the moment Intel's conception of naming its future CPUs looks the following way:

Note that the "series" for the desktop and mobile platforms repeat: the figure 7 at the beginning is used for the processor numbers of Pentium M and desktop Pentium 4 eXtreme Edition, 5 is used for mobile and desktop Pentium 4, and 3 for mobile and desktop Celeron. As someone noticed at the briefing, the guys at Intel must like BMW very much because of the figures chosen and because the respective series on the market look very similar to those of BMW :). But the new system of names doesn't touch upon the server sector - the Xeon and Itanium lines will still have clock speeds in their names.

The company says that it doesn't intend to rename the old processors, and the processor number will be used only for new models. Supposedly, first processors with the processor number in their names will be released in May, and first desktop models will be available in June. It looks like that all processors based on the mobile Dothan core and desktop CPUs for Socket T (Socket 775) will get the new-type names. It's also possible that some of the currently produced Banias based Pentium M CPUs will also get the processor number, but no desktop CPUs will be renamed. 

The most important is to avoid misunderstanding!

Looking at the above diagram and taking all the above facts into account remember what a processor number is and what it is not:

1. Processor number doesn't reflect a performance level. A CPU with a greater processor number must not necessarily run faster. On the other hand, it can be faster - the fact is that the processor number and performance are not directly connected. Some Pentium 4 780 can have a higher clock rate than Pentium 4 770 or it can have the same clock rate and a greater cache size, or it can have all the same physical features but a different instruction set.
2. Processor number can't be used for direct comparison of processors of different product lines. Even if the processor number is the same it's useless to draw any conclusion regarding similar features of, for example, Pentium M 755 and Pentium 4 755 eXtreme Edition. Moreover, they underline that the increased processor number of the Pentium M can indicate a completely different change compared to the Pentium 4 XE, even if the new models get the same identifier again.
3. Although an increase in the processor number generally means an improvement, this figure shouldn't be used as a determining factor when making a choice. A greater processor number doesn't mean that such model is preferable for a given end user.

Remember that the processor number is not its technical characteristic. This is just its unique identifier within its product line. Intel gave us an example of SKU (Stock Keeping Unit) as a synonym of processor number usually used to identify a product in stock. So, if a user really knows what he needs he is just considered to know the unique identifier of the product he needs. 

Conclusion

Well, something like a processor number was long awaited for from Intel. This solution looks logical and doesn't make things complicated. Finally we don't have to find out what all those letters coming after the clock rate in the name mean. As to removed clock speeds from CPU names, this is not a vital loss - this parameter lost its attractiveness as a key performance index long time ago. Frankly speaking we don't see a great difference between AMD's recent approach which voiced its opinion about CPU speeds in their names (they already refused TPI!) and Intel's future approach. Anyway, we will have to clarify this issue using the tests. 

Discuss in our forum