pytorch/vision

ROCM version is slow

Open

#2,682 建立於 2020年9月17日

在 GitHub 查看
 (7 留言) (0 反應) (0 負責人)Python (15,050 star) (6,858 fork)batch import
help wantedmodule: rocm

描述

🐛 Bug

Inference speed is abysmal - way slower than cpu. It works as expected on NVidia. I'm not really sure whether this is torch vision or torch itself at fault.

To Reproduce

Steps to reproduce the behavior:

  1. Dockerfile: https://gist.github.com/boniek83/0bbf858a24961816557c522007a11c11
  2. Lab: https://github.com/timesler/facenet-pytorch/blob/master/examples/infer.ipynb
  3. Data for lab: https://github.com/timesler/facenet-pytorch/tree/master/data/test_images

Expected behavior

Performance should be competitive.

Environment

  • PyTorch / torchvision Version (e.g., 1.0 / 0.4.0): 1.6.0 / 0.7.0
  • OS (e.g., Linux): CentOS 7.8
  • How you installed PyTorch / torchvision (conda, pip, source): source
  • Build command you used (if compiling from source): see Dockerfile above
  • Python version: 3.6
  • GPU models and configuration: Radeon VII
  • Any other relevant information: ROCK 3.7 during runtime

cc @iotamudelta @ashishfarmer

貢獻者指南

技術棧
pythonpytorch
領域
performancemachine learning
議題類型
bug
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
4
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
over 1 week
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
stale
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
needs investigation
前置要求
Understanding of PyTorch internalsExperience with AMD ROCm stackPerformance profiling skills
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
10
研究方向
The issue reports slow inference on AMD ROCm compared to CPU and Nvidia GPUs. Investigate whether the bottleneck is in torchvision or PyTorch core by profiling ROCm kernels. The user provided a Dockerfile and a facenet notebook for reproduction. First, attempt to reproduce the performance issue in a controlled environment. Then compare execution traces with and without ROCm to identify underperforming operations. Look for recent fixes or discussions in PyTorch's ROCm GitHub issues and relevant modules like torchvision's ops.
ROCM version is slow · pytorch/vision#2682 | Good First Issue