HomeTopics in Manufacturing
Topics in Manufacturing
ME 290R
Instructor: MCMAINS, S
Location: TuTh 11-1230P, 108 WHEELER
Email: mcmains@me.berkeley.edu
Course Description: Modern programmable Graphics Processing Units (GPUs) dramatically outperform CPUs on a
variety of arithmetically intense computations. Furthermore, the performance gap is growing, with GPU performance doubling roughly every 6 months instead of the 18 months it takes CPU performance to double. Unlike earlier graphics cards, today’s programmable GPUs allow the userto perform general purpose computations on the GPU, such as dense matrix multiplication or simulations of fluid dynamics. Making maximum use of this power requires understanding theunique programming model of the GPU, and creatively mapping a problem onto the graphics computation steps. This course will cover the fundamentals of the graphics pipeline, how to program the GPU, and the latest research on general-purpose GPU computation, with an emphasis on applications in Computer Aided Design, Computer Aided
Manufacturing, and Computer Aided Engineering analysis.
variety of arithmetically intense computations. Furthermore, the performance gap is growing, with GPU performance doubling roughly every 6 months instead of the 18 months it takes CPU performance to double. Unlike earlier graphics cards, today’s programmable GPUs allow the userto perform general purpose computations on the GPU, such as dense matrix multiplication or simulations of fluid dynamics. Making maximum use of this power requires understanding theunique programming model of the GPU, and creatively mapping a problem onto the graphics computation steps. This course will cover the fundamentals of the graphics pipeline, how to program the GPU, and the latest research on general-purpose GPU computation, with an emphasis on applications in Computer Aided Design, Computer Aided
Manufacturing, and Computer Aided Engineering analysis.