{ "cells": [ { "cell_type": "code", "execution_count": null, "metadata": { "collapsed": false }, "outputs": [], "source": [ "# For tips on running notebooks in Google Colab, see\n# https://codelin.vip/beginner/colab\n%matplotlib inline" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "(prototype) Accelerating `torch.save` and `torch.load` with GPUDirect Storage\n=============================================================================\n\nGPUDirect Storage enables a direct data path for direct memory access\ntransfers between GPU memory and storage, avoiding a bounce buffer\nthrough the CPU.\n\nIn version **2.7**, we introduced new prototype APIs to `torch.cuda.gds`\nthat serve as thin wrappers around the [cuFile\nAPIs](https://docs.nvidia.com/gpudirect-storage/api-reference-guide/index.html#cufile-io-api)\nthat can be used with `torch.Tensor` to achieve improved I/O\nperformance.\n\nIn this tutorial, we will demonstrate how to use the `torch.cuda.gds`\nAPIs in conjunction with checkpoints generated by `torch.save` and\n`torch.load` on local filesystem.\n\n```{=html}\n
torch.cuda.gds
APIs in conjunction withcheckpoints generated by torch.save
and torch.load
on local filesystem