Our today's article will be devoted to Low-End processors. Not the cheapest ones as Celerons are still waiting for their turn. (Though even they can provide a sufficient performance level for many users, necessarily being very cheap -- top Core 2 Duo processors cost just as much or even more than the cheapest Core 2 Quad products.) Anyway, two processor cores are the reasonable minimum for most desktops today. Most processors we are going to review today easily fall into the magical price range of "up to $200" (according to extensive research, processors from the higher price range have the target audience of only 3% of users). So they can be used in $500 computers (this price including a monitor and a graphics card). Performance of such a PC will still be much higher than that of a relatively inexpensive notebook (or a four-year-old top computer), so this choice is not only justified when you have a strictly limited budget, but also when you just don't want to spend too much money.

Testbed configurations

Intel once claimed that its label "Intel Pentium Inside" guaranteed that you bought a powerful computer based on the most powerful processor. However, just like money, trademarks are prone to inflation. So Pentium has stopped being an elite accessory long ago. It only confirms that you haven't bought Celeron. To put it plainly, it's a Low-End but sterling family of processors from Intel, nothing more. But if you take a look at technical characteristics of modern dice in this family of processors, you will see that they haven't been cut down that much, it's just that the other processors have advanced far forward. Series E5000 got a 2M cache and caught up with Core 2 Duo E4000, and the E6000 that appeared in late spring (it presently includes two models -- E6300 and E6500) added FSB 1066 to the list of its features, that is it ascended to the level of lower models Core 2 Duo E6300/6400. But clock rates of modern Pentiums are significantly higher than those of old Core 2 Duo processors.

And what can modern Core 2 Duo models boast of? E7000 models have almost nothing to show off. They only have 3M cache. In other respects the C2D E7400 and Pentium E6300 look very much alike. All the more interesting to compare them in this article, even if in slightly different conditions (we'll describe them a tad later)! And maximum clock rates are different as well -- Core 2 Duo E7600 has already stepped beyond 3 GHz, which was impossible even for Core 2 Extreme several years ago. On the other hand, this processor costs just like Core 2 Quad Q8200. So we have a natural question: what's preferable these days (not in the past) -- frequency or number of cores. Top models of Core 2 Duo E8000 have clock rates of up to 3.33 GHz (just like Core i7 Extreme 975. and even 6M L2 Cache. But they are more expensive -- even the cheapest E8200 (discontinued, but still available in retail stores) falls just a little short of the C2Q Q8200, and the top E8600 (with its record-breaking clock rate) compares with the top Q9000. As it turns out, it's impossible to buy a E8600 in Moscow now. So we decided to do without the top model. And even without the E8500 -- it exceeds the psychological border of $200, unlike the cheapest quad-core processor for LGA775 and the models after it: Q8300, Q8400, Q9300, and Q9400. So for the reasons of pricing and availability in stores, dual-core processors with the largest L2 Cache (and their total cache size is exceeded only in Phenom II X2) will be represented in our tests by the E8200 only. We'll find out whether such cache gives any advantages or it affects the price only -- we'll compare it with processors from other families. At least the idea of buying a dual-core processor for the price of an average (!) quad-core processor does not look sane (besides, a top triple-core processor from AMD is much cheaper): soft makers are making progress, and there already appear applications that can use more than two processor cores. As a rule, such programs belong to tasks, where performance is of utmost importance. What's the point in accelerating office applications, if their slowest link is a user? What's the point in accelerating games with old engines, if even a Pentium provides hundred FPS there, and performance in heavy modes is limited by a graphics card anyway? And so forth. On the other hand, the process of transcoding video may take up several hours, and it will be useful to accelerate it even by one hour.

It should be also noted that popularity of E8000 processors is dented not only by their high prices, but also their FSB 1333, unlike lower frequencies in lower CPU series. It does only good in the nominal mode. But it's harmful, if you buy a processor for overclocking (which happens a lot with users pressed for money). The fact is, there exist practically no motherboards supporting Wolfdale processors that could not work with at least FSB 1333 -- fortunately, it was supported even by the relatively old family of Intel 3x chipsets (except for the budget G31). So we can stay in the nominal mode and overclock the processor -- for example, Pentium E5300 with the 1333 MHz FSB (instead of the nominal 800 MHz) will have the clock rate of 4 GHz (if it can do it, of course). Not bad, eh? It can be done practically with any motherboard (as it's a nominal mode for it), with absolutely normal memory modules, etc. But if you want to reach the same level with the E8200, the base frequency must be raised to 500 MHz (FSB 2000), which is much harder to do (and it cannot be done with any motherboard, if you are interested in practical usage, not in a screenshot to shock your friends). However, overclocking is a complex and ambiguous issue that depends on many factors (including plain luck), so we shall not dwell on it. We'll just note that such an outlook on technical characteristics of processors exists as well.

We reserved the most interesting fact for the end: we managed to get our hands on a couple of old processors in our lab -- Core 2 Duo E6600 and E6750. The former was one of the fastest processors from Intel, when desktop Core 2 processors had just arrived. But now it's no worse than Pentium (not all of them at that), that's all. On the other hand, we are interested in its results to facilitate comparison of new processors with the old ones using our new test procedure. Lots of users, who bought their computers three years ago, focused on the E6000 family. But now the old PC may be outdated enough to entice users for an upgrade. Should you pay attention to the modern line of Core 2 Duo, or it will be bargaining one trouble for another? The same applies to the E6750. It's a newer model that added FSB 1333 support to Core 2 Duo, while only the E8000 features its support among modern CPU families. This processor has the same clock rate as the E8200 (also present in this article), but they have different cache sizes.

We've taken four processors for comparison. Results of the Core 2 Quad Q9300 are implicitly present in all tests conducted by the new method -- we've just converted them into the explicit form. And we are interested in the performance ratio between dual-core processors and the cheapest (even cheaper than some of our today's contenders) Intel Core 2 Quad Q8200. By the way, it's actually two Pentiums glued together just as representatives of the Q9x00 series are two dice (similar to Core 2 Duo E7000) under a single cap. There are also two guests from AMD. For one, it's an Athlon II X2 250, which will compete with the Pentium and a little with the Core 2 Duo E7000. For two, it's a Phenom II X3 720, which must lure users away from the Core 2 Duo E8000. We also wanted to add results of the Phenom II X2 to the diagrams. However, we are not sure that shipments of such processors will be large enough. After all, this article is devoted to Intel, not to AMD. Results of this processor are published in the summary diagram, so you can compare the results by yourself.

Write a comment below. No registration needed!