【CUDA开发】 Check failed: error == cudaSuccess (8 vs. 0) invalid device function

最近在复现R-CNN一系列的实验时，配置代码环境真是花费了不少时间。由于对MATLAB不熟悉，实验采用的都是github上rbg大神的Python版本。在配置Faster R-CNN时，编译没有问题，一运行 ./tools/demo.py --net zf 就会出现如下错误：

<span style="font-size:14px;">Loaded network ./data/faster_rcnn_models/ZF_faster_rcnn_final.caffemodel
F1008 roi_pooling_layer.cu:91] Check failed: error == cudaSuccess (8 vs. 0) invalid device function
*** Check failure stack trace: *** </span>

但是采用CPU mode运行时可以成功。

最后在https://github.com/rbgirshick/py-faster-rcnn/issues/2 找到了我想要的答案，有兴趣的可以慢慢阅读。

不想看的话，就直接按照我下面的方式修改。

一般情况下都是因为显卡的计算能力不同而导致的，修改 py-faster-rcnn/lib/setup.py 的第135行，将arch改为与你显卡相匹配的数值，（比如我的GTX 760，计算能力是3.0，就将sm_35改成了sm_30）然后删除utils/bbox.c，nms/cpu_nms.c ，nms/gpu_nms.cpp 重新编译即可

我看到有些人说还有其他的问题，那么可以在最开始的makefile.config文件中就开始修改，不过我没有试过,具体步骤如下

<span style="font-size:14px;">As below, there is my solution (thress steps):
1 if you're using the GPU instance on AWS, then please change the architecture setting into:
# CUDA architecture setting: going with all of them.
# For CUDA < 6.0, comment the *_50 lines for compatibility.
CUDA_ARCH := -gencode arch=compute_30,code=sm_30 \
-gencode arch=compute_50,code=sm_50 \
-gencode arch=compute_50,code=compute_50
Because the GPU in AWS does not support compute_35
2 I changed sm_35 into sm_30 in lib/setup.py file
3 cd lib, remove these files: utils/bbox.c nms/cpu_nms.c nms/gpu_nms.cpp, if they exist.
And then make && cd ../caffe/ && make clean && make -j8 && make pycaffe -j8 </span>

转载于:https://www.cnblogs.com/huty/p/8517108.html

【CUDA开发】 Check failed: error == cudaSuccess (8 vs. 0) invalid device function相关推荐

Check failed: error == cudaSuccess (8 vs. 0) invalid device function
我是编译caffe 例子make runtest时报的错: 1 if you're using the GPU instance on AWS, then please change the arch ...
Check failed: error == cudaSuccess (10 vs. 0) invalid device ordinal
ubuntu14.04 caffe 运行别人的程序,报错Check failed: error == cudaSuccess (10 vs. 0) invalid device ordinal.根据 ...
配置SSD-caffe测试时出现“Check failed: error == cudaSuccess (10 vs. 0) invalid device ordinal”解决
这是由于GPU数量不匹配造成的,如果训练自己的数据,那么我们只需要将solver.prototxt文件中的device_id项改为自己的GPU块数,一块就是0,两块就是1,以此类推. 但是SSD配置时 ...
Check failed: error == cudaSuccess (35 vs. 0) CUDA driver version is insufficient for CUDA runtime
Check failed: error == cudaSuccess (35 vs. 0) CUDA driver version is insufficient for CUDA runtime ...
【caffe】 Check failed: error == cudaSuccess (30 vs. 0) unknown error
解决办法 : 加 sudo 用caffe,不报错,那一定要的时假caffe. ubuntu16.04 ,写了一个.py文件调用resnet.caffemodel 来对图片进行分类. 但是在运行时出错, ...
caffe 报错 Check failed: error == cudaSuccess (77 vs. 0) an illegal memory access was encounteredcaffe
caffe 报错 Check failed: error == cudaSuccess (77 vs. 0) an illegal memory access was encountered 训练时候 ...
Check failed: error == cudaSuccess (74 vs. 0) misaligned address的解决方法
今天训练网络的时候加入了Deconvolution Layer,结果一运行报Check failed: error == cudaSuccess (74 vs. 0) misaligned addre ...
caffe 报错 Check failed: error == cudaSuccess (77 vs. 0) an illegal memory access was encountered
之前从没有遇到这样的问题,找了三个小时的bug和资料后,在此处发现了解决方案,他是这么说的: In my case the "top" and "bottom" ...
Check failed: error == cudaSuccess (74 vs. 0) misaligned address
以下摘自github hoszbh的回答 https://github.com/BVLC/caffe/issues/5729 modifying cudnn_conv_layer.cpp and al ...

【CUDA开发】 Check failed: error == cudaSuccess (8 vs. 0) invalid device function

【CUDA开发】 Check failed: error == cudaSuccess (8 vs. 0) invalid device function相关推荐

最新文章

热门文章