Pytorch Compiler Introduction
Pytorch Compiler Introduction
Pytorch Cuda Streams Introduction
This article explores the basic concepts of Cuda Stream, parallel execution, and multi-GPU synchronization strategies. We analyze the advantages of using multiple CUDA streams and how to ensure task synchronization through Cuda Event, utilizing Cuda streams to optimize program performance.
Overview of PyTorch Distributed Training
This document provides a comprehensive overview of distributed training capabilities within PyTorch. Covering the core components of
torch.distributed, it delves into Distributed Data-Parallel Training (DDP), RPC-Based Distributed Training, and Collective Communication (c10d).MallocLab
In malloc lab, we will implement our own versions of
malloc, free, and realloc.Introducton to Pytorch Broadcast
This article introduces the implementation details of pytorch broadcast mechanism, including the forward and backward calculation.
Unraveling PyTorch: Tensor Indexing and Assignment
This article dissects PyTorch’s C++ core to uncover the mechanics of tensor indexing and assignment. From translating Python indices to C++ TensorIndex to the nuances of
handleDimInMultiDimIndexing, we explore both basic and advanced tensor operations.Deep Dive to Pytorch AutoGrad(2)
This article introduces the implementation details of pytorch autograd mechanism.
Deep Dive to Pytorch AutoGrad(1)
This article introduces the implementation details of pytorch autograd mechanism.
ShellLab
In shell lab, we’ll become more familiar with the concepts of process control and signal by writing a simple Unix shell program that supports job control. Source: [https://github.com/yewentao256/CSAPP_15213/tree/main/shelllab]
Deep Dive to Pytorch Contiguous Operator(3)
Uncover the inner workings of PyTorch through a deep dive into the
contiguous operator, from its Python interface to its dispatching and registration process, and finally how it is executed.