Intel's Penryn core finally hits mass production under the codenames Yorkfield (quad core) and Wolfsdale (dual core), beating AMD to the punch with the release of the first lead-free CPU and bringing a massive 6 or 12MB L2 cache to the table. Built using Intel's P1266 process with Hafnium-based transistors in a 45 nm design, the new cores are geared towards exceptional energy efficiency with the added bonus of Nehalem New Instructions (NNI) also known as SSE4 instruction set.

We'll be reviewing the feature set of the new CPU, and look at performance, overclocking and energy efficiency of Yorkfield compared to the older versions of Intel processors and .. of course, the competition from Austin.

Intel’s Yorkfield Processor

The Core2 architecture has been a major success for Intel; after years of trying to compensate for Netburst architecture inadequacies with core frequency at the expense of power consumption and heat dissipation, the new design brings the word smart processing back to Intel. After a few minor adjustments with respect to bus interface clock, Intel is now releasing the first major revision of the successful design by bringing 45 nm process technology to mass production and adding 50 Nehalem New Instructions (NNI), more commonly referred to as SSE4. Most of the architectural and process intricacies have been covered in our earlier articles following the official unveiling of the Hafnium-gate transistor technology earlier this year, but there are other recent announcements that add to the low-power / environmental soundness of the new product line. A case in point is the completely lead-free manufacture of the processor with silver-tin-copper-based solder replacing the lead-tin bumps on the die.

A penny on a wafer of Penryn dies - not indicative of the actual manufacturing cost of the CPU! Photopgraph courtesy of Intel Corp.

In a nutshell, the technical specifications are:

• Available at 3.00 GHz
• 333 MHz (1333 MHz) PSB interface
• Supported Multipliers: 6, 7, 7.5, 8, 8.5, ... 13.5, 14, 15 for frequencies from 2-5 GHz
• 775-land Package
• 107 mm2 die size
• 820 million transistors
• 12 MB L2 Cache (2 x 6 MB L2 cache) 8-way set associative for improved cache hit rate on load/store operations
• Voltage ID Range: 0.850 – 1.3625V
• Max Core Voltage: 1.45V
• Max PSB Termination Voltage: 1.45V
• PSB Termination Voltage (DC + AC specifications): 1.045V – 1.155V
• VCC_BOOT Default VCC voltage for initial power up - 1.10 - V
• ICC Max: 125A (580 mA) per land*

Functional Specifications

• Enhanced Intel Speedstep® Technology
• Supports Intel® 64 architecture
• Supports Intel® Virtualization Technology
• Supports Execute Disable Bit capability
• PSB frequency at 1333 MHz
• Advance Dynamic Execution
• Very deep out-of-order execution
• Enhanced branch prediction
• Optimized for 32-bit applications running on advanced 32-bit operating systems
• Intel® Advanced Smart Cache
• Intel® Advanced Digital Media Boost
• Enhanced floating point and multimedia unit for enhanced video, audio, encryption, and 3D performance
• Binary compatible with applications running on previous members of the Intel microprocessor line
• Power Management capabilities
• System Management mode
• Multiple low-power states*

*The Core2* design features a variety of different power saving states from frequency / voltage throttling to selective clock gating.

As with all new CPU releases, the three questions are:

• How fast is it?
• Does it overclock?
• How energy-efficient is it?

In addition, we were looking at the effects of the new SSE4 instruction set, heralded to produce double digit percentile improvements in selective applications like for example VirtualDub 1.7.2 and higher with DivX 6.7.

We have the answers and some of them were predictable, some blew us away, and some were absolutely unexpected.