Harnessing Multicore Hardware
In the last few years, the trend in CPU performance improvement has shifted from raw computational speed to parallelism as various memory and power (heat dissipation) limits have been reached at the higher clock frequencies. After surpassing clock speeds of 3 GHz, further CPU performance improvement is now achieved by increasing the number of computational elements (cores) within a given CPU package. This means that in order for software applications to run faster on the latest machines, they need to take advantage of the new multicore architecture.
This becomes especially important for electronic design automation (EDA) applications. With growing design size and complexity, continued EDA tool performance improvements are necessary to help design teams meet their schedules. However, given the intricacy of EDA tools, many factors need to be taken into account when implementing multicore analysis hardware support to ensure fast turnaround time and optimal use of compute resources. In this paper, we discuss the evolution of multicore analysis computer hardware; the various approaches to implement multicore analysis support, specifically for static timing analysis tools, such as PrimeTime; and how to optimize hardware and software settings to improve turnaround time. To download this paper, click here.
Tags: Synopsys