Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of Graph Search.
This lecture discusses memory consistency models and their impact on performance when running custom binaries on different processors. It also covers the differences between Instruction-Level Parallelism (ILP) and Thread-Level Parallelism (TLP), as well as the implementation of atomic subroutines like Compare-And-Swap (CAS) in processors. Furthermore, it explores Transactional Memory properties, GPU architecture, shared memory usage, CUDA kernel execution steps, and GPU memory access patterns. The lecture concludes with a comparison of multithreaded workloads and the suitability of different multithreading granularities for processors.