CAFFE实验学习笔记（3）——SSD(Single Shot MultiBox Detector)

1.  cuda-7.5 is installed,but itshows 5.5 while checking the nvcc –v

http://blog.csdn.net/xuezhisdc/article/details/48651003 1、将/usr/local/cuda-7.0/bin添加到环境变量PATH路径中，这样一来，就可以在任何路径下调用cuda相关的可执行文件了。 2、将/usr/local/cuda7.0/lib64添加环境变量LD_LIBRARY_PATH 中，作为共享库使用。这样一来，后面编译Cuda Samples 和OpenCV时，就不会提示找不到库的错误了。 操作1：将以下内容添加到文件/etc/profile的最后面，保存后，执行命令source /etc/profile，使配置生效。 PATH=/usr/local/cuda-7.5/bin:$PATH export PATH LD_LIBRARY_PATH=/usr/local/cuda-7.5/lib64 export LD_ LIBRARY_PATH

ruyunli@ruyunli-All-Series:~$ echo $PATH /usr/local/cuda-7.5/bin:/home/ruyunli/anaconda2/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games

ruyunli@ruyunli-All-Series:~$ echo $LD_LIBRARY_PATH /usr/local/cuda-7.5/lib64

 

# Optionally, download your own cudnn; requires registration. if [ -f "cudnn-7.0-linux-x64-v4.0-prod.tgz" ] ; then   tar -xvf cudnn-7.0-linux-x64-v4.0-prod.tgz   sudo cp -P cuda/lib64/libcudnn* /usr/local/cuda/lib64   sudo cp cuda/include/cudnn.h /usr/local/cuda/include fi # Need to put cuda on the linker path.  This may not be the best way, but it works. sudo sh -c "sudo echo '/usr/local/cuda/lib64' > /etc/ld.so.conf.d/cuda_hack.conf" sudo ldconfig /usr/local/cuda/lib64

http://buptldy.github.io/2016/04/09/2016-04-09-Deepin%20CUDA%E5%AE%89%E8%A3%85%E5%8F%8AKeras%E4%BD%BF%E7%94%A8GPU%E6%A8%A1%E5%BC%8F%E8%BF%90%E8%A1%8C/

Device 0: "GeForce GTX 980 Ti"   CUDA Driver Version / Runtime Version          8.0 / 7.5   CUDA Capability Major/Minor version number:    5.2   Total amount of global memory:                 6075 MBytes (6369837056 bytes)   (22) Multiprocessors, (128) CUDA Cores/MP:     2816 CUDA Cores   GPU Max Clock rate:                            1190 MHz (1.19 GHz)   Memory Clock rate:                             3505 Mhz   Memory Bus Width:                              384-bit   L2 Cache Size:                                 3145728 bytes   Maximum Texture Dimension Size (x,y,z)         1D=(65536), 2D=(65536, 65536), 3D=(4096, 4096, 4096)   Maximum Layered 1D Texture Size, (num) layers  1D=(16384), 2048 layers   Maximum Layered 2D Texture Size, (num) layers  2D=(16384, 16384), 2048 layers   Total amount of constant memory:               65536 bytes   Total amount of shared memory per block:       49152 bytes   Total number of registers available per block: 65536   Warp size:                                     32   Maximum number of threads per multiprocessor:  2048   Maximum number of threads per block:           1024   Max dimension size of a thread block (x,y,z): (1024, 1024, 64)   Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)   Maximum memory pitch:                          2147483647 bytes   Texture alignment:                             512 bytes   Concurrent copy and kernel execution:          Yes with 2 copy engine(s)   Run time limit on kernels:                     No   Integrated GPU sharing Host Memory:            No   Support host page-locked memory mapping:       Yes   Alignment requirement for Surfaces:            Yes   Device has ECC support:                        Disabled   Device supports Unified Addressing (UVA):      Yes   Device PCI Domain ID / Bus ID / location ID:   0 / 1 / 0   Compute Mode:      < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >   Device 1: "GeForce GTX 970"   CUDA Driver Version / Runtime Version          8.0 / 7.5   CUDA Capability Major/Minor version number:    5.2   Total amount of global memory:                 4034 MBytes (4229758976 bytes)   (13) Multiprocessors, (128) CUDA Cores/MP:     1664 CUDA Cores   GPU Max Clock rate:                            1317 MHz (1.32 GHz)   Memory Clock rate:                             3505 Mhz   Memory Bus Width:                              256-bit   L2 Cache Size:                                 1835008 bytes   Maximum Texture Dimension Size (x,y,z)         1D=(65536), 2D=(65536, 65536), 3D=(4096, 4096, 4096)   Maximum Layered 1D Texture Size, (num) layers  1D=(16384), 2048 layers   Maximum Layered 2D Texture Size, (num) layers  2D=(16384, 16384), 2048 layers   Total amount of constant memory:               65536 bytes   Total amount of shared memory per block:       49152 bytes   Total number of registers available per block: 65536   Warp size:                                     32   Maximum number of threads per multiprocessor:  2048   Maximum number of threads per block:           1024   Max dimension size of a thread block (x,y,z): (1024, 1024, 64)   Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)   Maximum memory pitch:                          2147483647 bytes   Texture alignment:                             512 bytes   Concurrent copy and kernel execution:          Yes with 2 copy engine(s)   Run time limit on kernels:                     Yes   Integrated GPU sharing Host Memory:            No   Support host page-locked memory mapping:       Yes   Alignment requirement for Surfaces:            Yes   Device has ECC support:                        Disabled   Device supports Unified Addressing (UVA):      Yes   Device PCI Domain ID / Bus ID / location ID:   0 / 5 / 0   Compute Mode:      < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) > > Peer access from GeForce GTX 980 Ti (GPU0) -> GeForce GTX 970 (GPU1) : No > Peer access from GeForce GTX 970 (GPU1) -> GeForce GTX 980 Ti (GPU0) : No   deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 8.0, CUDA Runtime Version = 7.5, NumDevs = 2, Device0 = GeForce GTX 980 Ti, Device1 = GeForce GTX 970

 

The NVIDIA CUDA® Deep Neural Network library(cuDNN) is a GPU-accelerated library of primitives for deep neural networks.