Original 45nm Intel® Core™ Microarchitecture

  Section 3 of 10  

Original 45nm Intel® Core™2 Processor Performance

MICROARCHITECTURE ENHANCEMENTS

The PenrynΔ family of processors is the next generation of Intel® processors based on the Intel® Core™2 microarchitecture, implemented on Intel's 45nm, Hi-k metal gate process technology. Frequency improvements, within existing power and thermal envelopes, over previous-generation processors, a larger L2 cache, microarchitectural enhancements, and improvements in power- and thermal-management schemes deliver improved performance per watt and energy efficiency for a broad range of client and server applications. The Penryn family of processors also added 47 new SSE4 instructions that can improve the performance of audio, video, image-editing applications, video encoders, 3-D applications, and games.

Microarchitecture enhancements that improve performance in the Penryn family of processors include the following:

• Larger Cache: Penryn processors include up to a 50 percent larger L2 cache with a higher degree of associativity that further improves the hit rate, maximizing its utilization. Dual-core Penryn processors feature up to a 6-MB L2 cache and quad-core processors up to a 12-MB L2 cache.
• Faster Divider: Penryn processors provide faster divider performance, roughly doubling the divider speed over previous generations through the inclusion of a new, faster divide technique called Radix 16.
• Super Shuffle Engine: Shuffles (the repositioning of bits) is a common operation in image- and video-editing applications. By implementing a full-width, single-pass, 128-bit-wide shuffle unit, a processor from the Penryn family of processors can perform full-width shuffles in a single cycle and is 3 times faster than previous-generation processors. The Super Shuffle Engine improves the performance of Intel Streaming Single Instruction Multiple Data (SIMD) Extensions (SSE), Streaming SIMD Extensions 2 (SSE2), Supplemental Streaming SIMD Extensions 3 (SSSE3), and Streaming SIMD Extensions 4 (SSE4) instructions, and this will benefit a wide range of applications including imaging and video applications, games, 3D modeling, and high-performance computing.
• Inclusion Filter: An Inclusion Filter was added in the Penryn family of processors to enhance the existing inclusion logic that was limiting server performance.
• Renamed RSB: The Renamed Return Stack Buffer (RRSB) increases return prediction accuracy and improves performance.
• CLI STI Performance Tuning: In the Penryn family of processors, the Clear Interrupt Flag (CLI) and Set Interrupt Flag (STI) macroinstructions were optimized to perform an execution pipeline serialization only when a new IF value is consumed and only if the new value is not yet updated, instead of post-serializing on every CLI or STI. This improves throughput of CLI-STI pairs by 2.5 times over previous-generation technology.
• Enhanced Intel® Dynamic Acceleration Technology (EDAT): EDAT is a power-management feature added to mobile processors that improves energy efficiency by dynamically increasing the performance of active core(s) when not all cores are utilized.
• Enhanced Intel® Virtualization Technology: Virtualization partitions or compartmentalizes a single computer so that it can run separate operating systems and software. This virtual partitioning better leverages multi-core processing power, increases efficiency, and cuts costs by letting a single machine act as many virtual ‘mini’ computers. The Penryn family of processors speeds up virtual machine transition (entry exit) times by an average of 25 percent to 75 percent. This is all done through microarchitecture improvements and requires no virtual machine software changes.

  Section 3 of 10  

Back to Top