[英]CUDA segmentation fault for trivial tutorial example
I am trying to run the example add.cu
(see below) from this official nvidia tutorial using nvcc add.cu -o add_cuda; ./add_cuda
我正在尝试使用
nvcc add.cu -o add_cuda; ./add_cuda
从此官方nvidia教程中运行示例add.cu
(请参见下文) nvcc add.cu -o add_cuda; ./add_cuda
nvcc add.cu -o add_cuda; ./add_cuda
and get Segmentation fault (core dumped)
. nvcc add.cu -o add_cuda; ./add_cuda
并获得Segmentation fault (core dumped)
。
I installed the nvidia cuda toolkit on Ubuntu 18 using sudo apt install nvidia-cuda-toolkit
. 我使用
sudo apt install nvidia-cuda-toolkit
在Ubuntu 18上安装了nvidia cuda工具sudo apt install nvidia-cuda-toolkit
。 I have a NVIDIA GF100GL Quadro 5000 and am using NVIDIA driver metapackage from nvidia-driver-390 (proprietary, tested)
我有一个NVIDIA GF100GL Quadro 5000,并且正在使用
NVIDIA driver metapackage from nvidia-driver-390 (proprietary, tested)
软件包NVIDIA driver metapackage from nvidia-driver-390 (proprietary, tested)
I have little C++ experience, but the pure C++ code from the beginning of the tutorial compiled and ran correctly. 我几乎没有C ++经验,但是从本教程开始的纯C ++代码可以正确编译并运行。
Following a comment, I added a check for the return of cudaMallocManaged
and got operation not supported
. 发表评论后,我添加了一个检查
cudaMallocManaged
返回的信息,并获得了operation not supported
。
#include <iostream>
#include <math.h>
// Kernel function to add the elements of two arrays
__global__
void add(int n, float *x, float *y)
{
for (int i = 0; i < n; i++)
y[i] = x[i] + y[i];
}
int main(void)
{
int N = 1<<20;
float *x, *y;
// Allocate Unified Memory – accessible from CPU or GPU
cudaMallocManaged(&x, N*sizeof(float));
cudaMallocManaged(&y, N*sizeof(float));
// initialize x and y arrays on the host
for (int i = 0; i < N; i++) {
x[i] = 1.0f;
y[i] = 2.0f;
}
// Run kernel on 1M elements on the GPU
add<<<1, 1>>>(N, x, y);
// Wait for GPU to finish before accessing on host
cudaDeviceSynchronize();
// Check for errors (all values should be 3.0f)
float maxError = 0.0f;
for (int i = 0; i < N; i++)
maxError = fmax(maxError, fabs(y[i]-3.0f));
std::cout << "Max error: " << maxError << std::endl;
// Free memory
cudaFree(x);
cudaFree(y);
return 0;
}
Your card belongs to fermi family with compute capability version 2.0. 您的卡属于具有计算功能2.0版的fermi系列。 It does not support the Unified Memory as stated here:
它不支持统一内存,如下所示:
K.1.1.
K.1.1。 System Requirements
系统要求
Unified Memory has two basic requirements:
统一内存有两个基本要求:
a GPU with SM architecture 3.0 or higher (Kepler class or newer)
具有SM架构3.0或更高版本(Kepler类或更高版本)的GPU
a 64-bit host application and non-embedded operating system (Linux, Windows, macOS)
64位主机应用程序和非嵌入式操作系统(Linux,Windows,macOS)
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.