The New Core 2 Duo: 45-nm Wolfdale Replaces 65-nm Conroe

Ad

Intel’s 65-nm generation isn’t dead yet, as the firm will still be offering multiple models in the server, desktop and notebook spaces for a while. The processors used in Apple’s ultra portable Mac Book Air and Lenovo’s upcoming X300 series are good examples. Both devices are based on power-optimized sub-20-Watt versions of the 65-nm Merom core, which means that the first generation of Core 2 will still be around for a while. However, Intel has recently introduced its latest and greatest product generation: the Core 2 Duo E8000 family, based on the 45-nm Wolfdale dual core. The new architecture offers new features, more clock speed and additional cache speed—all the while lowering power consumption. We looked at the new architecture to find out how much of an improvement the new processor generation really offers.

Since Intel started to relax its information policy a few years ago, details about future products have been made available earlier than in the past. Pre-launch data about the new 45-nm Core 2 Duo was no exception. While the top model, the Core 2 Extreme QX9650, which is based on four processing cores, was already launched in late 2006, the dual-core processors for the mainstream have been available for only a few weeks. The key improvements hence didn’t come as a surprise. The new 45-nm core, code named Wolfdale, is still based on the second generation Core micro architecture, and it succeeds the 65-nm Conroe core. You may have heard about Penryn as well; this is Intel’s code name for the entire 45-nm product family as well as the mobile version. Merom refers to the 65-nm desktop and mobile CPU generation.

The 45-nm Wolfdale processors offer a 6-MB L2 cache size, which is 50% larger than the 65-nm Merom’s 4-MB cache. Intel has shrunk the die size from 143 to 107 mm². Cut-down versions with only 3-MB L2 cache are expected to follow shortly. All models are powered by the hafnium-based high-K metal gate silicon manufacturing process, which helps to reduce leakage currents and leads to power management improvements. This has translated into a reduction in both maximum and idle power requirements. As we saw in our review of the 45-nm Core 2 Extreme QX9650, the improvements are welcome features.

So far, Intel has launched three mainstream models: the Core 2 Duo E8200, E8400 and E8500, running at 2.66, 3.0 and 3.16 GHz, and a 333-MHz system clock speed (FSB1333). Since a dual-core Xeon X5272 running at 3.4 GHz (80 W TDP) has already been launched, we’re sure that there are clock speed margins that Intel should be able to access almost at any time. Although we’d love to see a high-speed dual-core processor, it will likely not arrive anytime soon, as AMD hasn’t been able to compete with Intel in the high end. But you may get the results you want yourself; in fact, we have been successful in overclocking the Core 2 Duo Wolfdale by a significant margin.

We will look at the new processors’ technical details and their performance in this first part of our Wolfdale review, and in the second part, we will analyze their overclocking capabilities and the power requirements when they’re used for everyday workloads over a defined period of time.


Talkback
human_error 14/02/2008 01:57
Hide
-0+
human_error
"While the top model, the Core 2 Extreme QX9650, which is based on four processing cores, was already launched in late 2006" isn't that meant to be 2007?
Wild9 14/02/2008 08:36
Hide
-0+
Wild9
Quote :While the die size decreased, transistor size increased by 40% from 291 million to 410 million.


I'd say transistor count rather than size :)
Wild9 16/02/2008 01:17
Hide
-0+
Wild9
I still maintain that AMD is not so bad - in this article you are effectively comparing a Ferrari and a Porsche; both are fast products.

Does it matter whether you drive a Ferrari to go to the corner store? Most people would say no, because all that most people want to do is get from A-to-B as cheaply as possible. And that is where AMD fits in.

People are too quick to slate AMD..without the A64 there would be no C2D/C2Q. And still, the A64 which is a generation behind C2D, can still compete. The 6000+ puts in a stirling effort.

These 45nm parts are good devices, but they lack a dedicated bus interconnect and of course, they lack a memory controller. Even with these things I would not buy Intel because historically they have always lagged behind in terms of innovation (and some would say, fair play). Phenom does need work, and hopefully AMD can address this, but I think the only reason you're suddenly seeing half-decent products from Intel is because it has had to. Right up until recently AMD has been a fierce competitor. In certain, bandwidth-intensive applications AMD still has the lead.
MrRimmer 18/02/2008 01:23
Hide
-0+
MrRimmer
Deep Power Down might not always make sense in a desktop environment, where more applications and services are active than on the desktop.

Note You are going to post a comment as anonymous.



Google Ads