Review: Intel Core 2 Duo/Extreme processors

The new dual-core processors have arrived and they don’t disappoint

Intel’s next generation of dual-core desktop processors has finally arrived and we've managed to get hold of two before their official release date on the 27 July.

The samples we've been testing are the Core 2 Duo E6700 with a clock speed of 2.67GHz and a Core 2 Extreme X6800 at 2.93GHz.

The Core 2 Duo and Core 2 Extreme are the first desktop processors to incorporate Intel’s Core Microarchitecture enhancements, which we'll talk about in more detail later.

They are fabricated using a 65nanometre manufacturing process, which Intel initially used to construct high-end Pentium D chips.

Both processors are well equipped with 4MB of Level 2 (L2) cache and an effective front-side bus speed of 1,066MHz (actually 266MHz but data is transferred four times per clock cycle).

Such a large, unified cache and speedy channel to access Ram, among the other architectural improvements, more than make up for the clock speed deficiency of these chips when compared with the 3.6GHz that later revisions of the Pentium D were pushed to.

The Duo performed impressively in our benchmark tests with a PCmark05 score of 6,678. In Cinebench it clocked up 448 (single threaded) and 818 (multithreaded). If you take a look at our previous processor benchmark scores, you'll see just how impressive the Core 2 Duo E6700 is.

These scores are broadly similar to the ones we reported on when we performed Intel-supervised controlled tests in Munich in May.

The E6700 processor is only surpassed by the blazingly fast X6800, which managed an unrivalled PCmark05 score of 7,404 and Cinebench scores of 484 (single threaded) and 899 (multithreaded).

Put simply, the Core 2 Extreme X6800 is the fastest desktop processor we’ve ever seen.

AMD recently released improved processors running on its new AM2 socket, but even when overclocked they fall well short of Intel's from a raw performance perspective.

The E6700 consumes an average of 65w of power. This is quite low, so power-conscious users should be more concerned about how much power is being drawn by their graphics card; up to 240w with a two card SLI/Crossfire setup.

The X6800 Extreme chip, however, sacrifices the energy efficiency of the Duo in favour of pushing the Core Microarchitecture to its performance limits.

All the advancements outlined above contribute to a smarter, faster processor that doesn’t disappoint when it comes to multitasking of processor-intesive applications.

Although much of the motivation for Core Microarchitecture came from the success of the Pentium M line, it’s not directly derived from the mobile platform.

In August, Intel will release its Core 2 Duo mobile chips, codenamed Merom. These will be very similar to the desktop Core 2 Duo, except for a slower front-side bus and more battery conservation features.

As for the Core 2 Duo and Core 2 Extreme processors, we're seriously impressed with their performance.

See our review of Hi-Grade's Ultis Tachyon, the first PC to reach us with one of Intel's new processors. Alternatively, click on to page 2 to read more about the technical aspects of the Core 2 Duo and Core 2 Extreme.

Related reviews:
Hi-Grade Ultis Tachyon Core 2 Extreme PC
A super-fast PC with decent components supporting the Core 2 Extreme.

Gigabyte GA-965P-DQ6 Core 2 Duo/Extreme
A new motherboard that supports Intel's latest processors

Page 2: Core Microarchitecture explained...

Essentially there are five strands of Core Microarchitecture development in the Core 2 Duo range.

The first of is Wide Dynamic Execution. This encompasses the ability of each core to process four instructions at a time and the support for macrofusion, which involves combining commonly coupled processor instructions to save execution time. Previously, processor cores could only handle three instructions simultaneously.

Intelligent Power Capability is the management of power the processor draws when running. A sophisticated control in the cores only powers up individual logic systems when they’re needed. As a result power consumption is reduced.

Advanced Smart Cache is the mechanism by which L2 cache is shared between the two cores. The cache is shared and uncontrolled, meaning one core can be dynamically allocated up to 100% of it.

An advantage over independent cache for each core is that data need only be stored in one place if both cores are working on it.

The ability to unify a 4MB cache is great during tasks such as the transcoding of movies; each frame can be completely loaded into cache rather than Ram.

Smart Memory Access ensures that the location of data that the cores are working on is optimally placed in memory, thus reducing the effect of memory latency.

Further to this, improved efficiency of out-of-order instruction processing is achieved by an improvement in the algorithm which deals with the speculative loading of data into the processor’s memory banks.

It sounds complicated (indeed, it is), but it basically means the processor makes use of each clock cycle rather than letting some wastefully slip by.

Intel has also worked on the advanced pre-fetching of instructions in an attempt to decrease the amount of Ram accesses made during an operation. As a result, it's able to better use the cache.

Finally, Advanced Digital Media Boost sees a great improvement in the handling of 128-bit SSE (Streaming SIMD Extensions) instructions. These are now completed in a one clock cycle rather than split into two 64-bit calculations as was the case with prior architectures.

The end result is an increase in speed during intensive applications such as video editing.

If that's left you feeling more than a little confused, it's worth taking a look at Intel's promotional video on Core Microarchitecture .