search

microsoft / nnfusion

A flexible and efficient deep neural network (DNN) compiler that generates high-performance executable from a DNN model description.
MIT License
955 stars 159 forks source link

whether supported on other platforms #519

Open wxthu opened 1 year ago

wxthu commented 1 year ago

I have tried to deploy nnfusion on various GPU card such as GTX 1660Ti, Tesla T4 and RTX 4090. However, all trials end with failure. And I found in your tutorial that nnfusion are supported on some cards as following. Can I understand that we cannot deploy nnfusion on GPU cards except that on your list?

image
wxthu commented 1 year ago

@LeiWang1990

LeiWang1999 commented 1 year ago

Actually, the nnfusion codegen doesn't depend on the gpu platform, could you kindly provide the error message that you have encountered?

wxthu commented 1 year ago

Thank you. When I compiled model in docker container _nnfusion /home/frozen_lstm_l8s8h256bs1.pb , I met the following error image

Version: Ubuntu 20.04 Cuda in system: 11.8 Driver: 520.61.05 @LeiWang1999

LeiWang1999 commented 1 year ago

This seems to be some environment issue, have you tried nvidia cuda samples to work?