Programming Parallel Computers 2019

Exercise NN: neural networks

Individual tasks

Please see the grading tool for details on benchmark instances and time limits.

NN9a: fast CPU solution [challenging]

Using all resources that you have in the CPU, solve the task as fast as possible. You are encouraged to exploit instruction-level parallelism, multithreading, and vector instructions whenever possible, and also to optimize the memory access pattern.

Deadline week 6. Maximum points 5.

NN9b: fast GPU solution [challenging]

Using all resources that you have in the GPU, solve the task as fast as possible.

Deadline week 6. Maximum points 5.