6.5930/1 Hardware Architecture for Deep Learning

6.5930/1 Spring 2024 Lecture Notes

L-01: Introduction and Applications (pdf)
L-02: Overview of DNN components (pdf)
L-03: Popular Models (pdf)
L-04: Evaluation and Training (pdf)
L-05: Kernel Computation - CPU pdf
L-06: Kernel Computation - Memory (handout) (pdf)
L-07: Kernel Computation - Vectorization (handout)(pdf)
L-08: Transforms (handout) (pdf)
L-09: GPUs (pdf)
L-10: Accelerator Architecture (pdf)
L-11: Dataflows - 1 (handout) (pdf) (animations)
L-12: Dataflows - 2 (handout) (pdf) (animations)
L-13: Convolutional Mappings (handout) (pdf)
L-14: Calculating Data Movement (pdf) (notebook)
L-15: Sparsity (pdf) (Information on project)
L-16: Sparse Architectures - 1 (handout) (pdf)
L-17: Sparse Architectures - 2 (handout) (pdf) (animations)
L-18: Sparse Matrix Multiplication (pdf) (animations)
L-19: Numeric Precision (pdf)
L-20: Advanced Technology (pdf)
L-21: Advanced Technology (pdf)

Note: Lecture notes marked "handouts" are for use during class (without answers to the questions), while notes marked "pdf" are the full lecture notes. Notes marked "animations" show the animated movies of data accesses. Notes marked "notebook" are Jupyter notebooks.