Intel's next-generation Core processor architecture

Because Core 2 Duo's performance gains are not simply due to ramped-up clock speeds, it's not a simple task to uncover the key developments -- and unless you're an engineer, chip architecture discussions are pretty tedious. With this in mind, Intel is trying to simplify the Core 2 Duo story by breaking out its advances into five major components. Below you'll find Intel's terms, followed by our translations:

Wide Dynamic Execution
Instead of sending data through the pipe faster, Core 2 Duo will bite off larger chunks of data to process at once. It can also combine the identical parts of different data requests. This should translate to faster overall performance.

Advanced Digital Media Boost
Every step in the multimedia processing chain is now 128-bit, which wasn't always the case. Since the CPU doesn't have to translate the non-128 bit parts, it has less work to do, and you get faster multimedia performance.

Smart Memory Access
A smarter traffic cop between the CPU and the memory.

Advanced Smart Cache
Cache is the place between the processor and main system memory where frequently used data lives. It's broken into different levels -- L1, L2 and so on -- with the amount of data stored traded off against the speed at which the chip can access that data. Core 2 Duo is a dual-core chip, and although each core has its own L1 cache, they share a dynamic L2 cache. So if core 1 needs only 1MB of L2, and core 2 could benefit from 3MB, Core 2 Duo chips can adjust that distribution on the fly. Intel's previous-generation Pentium Ds and AMD's dual-core Athlons have a fixed L2 cache per core, which means no load balancing, potentially wasting L2 cache space. Translation: better performance.

Intelligent Power Capability
Intel has incorporated the power management techniques from its notebook CPUs into the multiplatform Core architecture. This means a Core 2 Duo can shut off whatever parts it's not using, reducing power consumption. So, not only do you get smaller heat sinks and fans (which should make DIYers lives' easier and overclockers' potentially more exciting -- see the image above), but it also makes room for more micro-form-factor PCs such as Apple's Mac mini and the AOpen Mini PC.

Intel did tell us when Core 2 Duo is shipping, but it also bound us to secrecy by way of an NDA (Non Disclosure Agreement). At least the date is firm, and it's in Q3 2006. Intel also said that we'd have samples for testing soon. If you're interested in what's been published around the Web, your best bet is these non-Intel-supervised results (wherein the Core 2 Duo indeed lives up to the hype) from Anandtech and Firingsquad. Anandtech's report is especially interesting because it highlights the upcoming market realities facing Intel. With lots of older chips in the channel through the end of the year, Dell and others will probably cut prices drastically to unload old supplies. Expect tempting desktop offers this holiday shopping season, along with some difficult price/performance choices.

Intel actually had another big revelation for us. At the end of the Core presentation, one of the company's representatives called up Windows Task Manager's performance tab to reveal four processor gauges dancing before us. In fact, they were flatlined, since the alleged quad-core chip in the PC before us, code-named Kentsfield, was running only PowerPoint. We weren't able to verify it for ourselves, but given AMD's Socket 4x4 announcement last week, we're not surprised that Intel was eager to respond with its own multi-core technology.