CUDA cask failure at execution for trt_volta_scudnn_128x32_relu_small_nn_v1

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

Hi all,

I have some models in TensorFlow and then optimized by TensorRT. They are running well at fp32 but I am having some problems when using half-precision. I am pasting the error just below:

ERROR: cuda/reformat.cu (452) - Cuda Error in NCHWToNCHW: 77
ERROR: cuda/reformat.cu (452) - Cuda Error in NCHWToNCHW: 77
ERROR: cuda/cudaTopKLayer.cpp (115) - Cuda Error in execute: 77
ERROR: cuda/cudaTopKLayer.cpp (115) - Cuda Error in execute: 77
ERROR: CUDA cask failure at execution for trt_volta_scudnn_128x32_relu_small_nn_v1.

I’ve generated the PLAN file from a turing uarchitecture (GeForce 2080 Ti) so it sounds weird for me that part that sais “failure at execution for trt_volta_scudnn_128x32_relu_small_nn_v1”. Does it have any sense?

Thanks!

2

Hello, It’d help us debug if you can provide a small repro that contains the source and model you are trying to convert to UFF that exhibit this error. Also, can you provide details on the platforms you are using?

Linux distro and version
GPU type
nvidia driver version
CUDA version
CUDNN version
Python version [if using python]
Tensorflow version
TensorRT version

3

@NVES It was my fault. I think the problem was I was trying to run fp32 models as fp16.
Thanks!