Chris will be holding two sets of office hours to get started on this. One will be Monday 7-9pm. The other will be Wednesday 8-10pm to best accommodate all schedules.
This is the starter code for this project. It loads a vector from file and sums each element into a single result. It then saves the final result.
The purpose of this portion of the lab is to look into loop unrolling and how it can impact the performance of the code executed. This is an optimization that can be perfromed by gcc and therefore may not have a visible impact.
Using the previous code from the loop unrolling section, we should update this to use the ARM Neon intrinsics. Please see the resources.
The goal of this portion of the assignment is to provide a baseline, simple implementation of a matrix multiply.
The goal of this portion of the assignment is to improve the performance of this by using block matrix multiplication.
The goal of this portion of the assignment is to perform some loop unrolling in preparation for ARM Neon implementation of block matrix multiplication.
The goal of this portion of the assignment is to using ARM Neon instead of loop unrolling for vectorization.