问题描述 Issue Description

如图，用官方命令检测是正常的。跑MLP也没问题。但是一运行卷积，内核就挂了。
python 3.7 paddle 2.5

版本&环境信息 Version & Environment Information

Paddle version: N/A
Paddle With CUDA: N/A

OS: Windows 10
GCC version: N/A
Clang version: N/A
CMake version: N/A
Libc version: N/A
Python version: 3.7.16

CUDA version: 11.7.99
Build cuda_11.7.r11.7/compiler.31442593_0
cuDNN version: N/A
Nvidia driver version: 531.79
Nvidia driver List:
GPU 0: NVIDIA GeForce RTX 3090