By Steven Walton on November 19, 2007
Editor: Julio Franco
Post/read user's feedback
Today we have in our hands the latest Intel Core 2 Extreme quad-core processor, and I know what you are thinking but no, this is not yet another extremely expensive processor featuring the same four cores with only a speed bump over the last QX6850 we looked at last time. Hmm, ok so I told a bit of a white lie there, as this is still an extremely expensive processor, but it’s not just another in a chain of speed bumps. The new Core 2 Extreme QX9650 represents Intel’s new flagship processor technology which until now was mainly known by its code-name “Penryn”, manufactured on a 45nm design process, adding fifty new SSE4 instructions, among other things.
Now, the actual QX9650 processor and other Penryn quad-cores have received its own codename which is “Yorkfield” and because this particular processor belongs to the Extreme line up, its complete name is Yorkfield XE. Similarly, future dual-core CPUs based on Penryn will go by the code-name “Wolfdale” once they are released. To complete the equation you should know that Yorkfields are really two Wolfdale chips stuck together. This is the same as with previous Kentsfield quad-core processors which in reality were just two Conroe chips stuck together in one package. This may be a bit confusing but when time to buy comes, it's important being able to identify each different core version and know what advantages one generation of chips has over another, ultimately making the better buying decision.
Intel "Penryn" processors incorporate 410 million transistors for each dual core chip,and 820 million for each quad core chip. The original Intel Pentium Processor only has 3.1 million transistors.
The Core 2 Extreme QX9650 is the first of many products to be released using the Penryn architecture. Like the previous Core 2 Extreme QX6850, the QX9650 also works at 3.0GHz but adds a number of enhancements. Perhaps the most prominent, Intel's 45nm High-k metal gate silicon technology, which is claimed to be an industry's first, featuring transistors with reduced current leakage designed to decrease power consumption while also accommodating for increased clock speeds. This is a big deal considering Intel has used the conventional silicon-oxide technique since 1960.
Besides the apparent die shrink, the Penryn architecture features a faster Radix-16 divider, faster OS primitive support, improved store forwarding, split load cache enhancements, enhanced virtualization technology and enhanced dynamic acceleration technology. While all of these may sound nice in paper, in real-life terms what we do know is that this newer Core 2 architecture is faster and more efficient. We were able to observe power consumption levels that are significantly lower as well as operating temperatures. These are all aspects that we will uncover in this review, but for now let me just say that when heavily overclocked the QX9650, it produced stress temperatures that were comparable to a dual-core Conroe processor operating at its default frequency.
When we move on to benchmarking the QX9650, it will extremely easy to spot the differences and advantages Penryn offers today because the QX9650 and the QX6850 (65nm Kentsfield) both operate at 3.0GHz and use the same front side bus speeds, so we will be putting them head to head for comparison. Also to note, the QX9650 does receive a L2 cache boost at 12MB versus only 8MB found in the QX6850.