PERFORMANCE EVALUATION OF MULTI-CORE PROCESSOR: A CASE STUDY OF QUAD-CORE PROCESSOR

Abstract

Several processor innovations to enhance computer performance have been developed; ranging from those utilized on single core processors like pipelining, increasing clock frequency etc which have hit their limits – leading to increasing the number of processors on single chips to bring about better resultant performance from the combine utilization of the multiple processors, known as chip multiprocessor. As long as there are still demand for better performance the efforts of improving computer performance continues. This study researched into the improvement features of Intel Nehalem Micro-architecture, the microprocessor architecture designed with acclaimed high performance (referred to by Intel as “the Next Generation Microarchitecture) used in the production of Intel Proliant X5570 processor. Using Standard Performance Evaluation Corporation (SPEC) CPU 2006 benchmark suites, the performance evaluations of the multicore processor were carried out. The CINT2006 part of the suite which contains the integer performance benchmarks was employed. CINT2006 focuses on measuring and comparing compute-intensive integer performance of the processors. The CINT2006 benchmarks were also run on AMD Jaguar Microprocessor –a high rated microprocessor architecture used in production of processors for Xbox One (a home video game console developed and marketed by Microsoft) to compare their performances in term of execution speed and throughput. Nehalem micro-architecture proved a considerable better performance over Jaguar micro-architectural design. The enhancement achieved by the Nehalem architecture was based on the improved interconnectivity of the component which was developed from Front Side Bus to Quick Path Interconnect. It was concluded that an increase in the processor performance in term of speed and throughput can be achieved by the integration of fast interconnectivity between the different components/parts of the processor to enhance the speed of transfer of both instructions and data, making it close to the speed of the processor chip.