周健康 阅读(40) 评论(0)

Jetson TX2上的demo

一、快速傅里叶-海动图 sample

The CUDA samples directory is copied to the home directory on the device by JetPack. The built binaries are in the following directory:

/home/ubuntu/NVIDIA_CUDA-<version>_Samples/bin/armv7l/linux/release/gnueabihf/

这里的version需要看你自己安装的CUDA版本而定

Run the samples at the command line or by double-clicking on them in the file browser. For example, when you run the oceanFFT sample, the following screen is displayed.

 

二、车辆识别加框sample

nvidia@tegra-ubuntu:~/tegra_multimedia_api/samples/backend$

 

./backend 1 ../../data/Video/sample_outdoor_car_1080p_10fps.h264 H264

 

--trt-deployfile ../../data/Model/GoogleNet_one_class/GoogleNet_modified_oneClass_halfHD.prototxt

 

--trt-modelfile ../../data/Model/GoogleNet_one_class/GoogleNet_modified_oneClass_halfHD.caffemodel --trt-forcefp32 0 --trt-proc-interval 1 -fps 10

 

三、GEMM(通用矩阵乘法)测试

nvidia@tegra-ubuntu:/usr/local/cuda/samples/7_CUDALibraries/batchCUBLAS$ ./batchCUBLAS -m1024 -n1024 -k1024

batchCUBLAS Starting...

GPU Device 0: "NVIDIA Tegra X2" with compute capability 6.2

 

 ==== Running single kernels ====

Testing sgemm#### args: ta=0 tb=0 m=1024 n=1024 k=1024  alpha = (0xbf800000, -1) beta= (0x40000000, 2)#### args: lda=1024 ldb=1024 ldc=1024

^^^^ elapsed = 0.00372291 sec  GFLOPS=576.83@@@@ sgemm test OK

Testing dgemm#### args: ta=0 tb=0 m=1024 n=1024 k=1024  alpha = (0x0000000000000000, 0) beta= (0x0000000000000000, 0)#### args: lda=1024 ldb=1024 ldc=1024

^^^^ elapsed = 0.10940003 sec  GFLOPS=19.6296@@@@ dgemm test OK

 

 ==== Running N=10 without streams ====

Testing sgemm#### args: ta=0 tb=0 m=1024 n=1024 k=1024  alpha = (0xbf800000, -1) beta= (0x00000000, 0)#### args: lda=1024 ldb=1024 ldc=1024

^^^^ elapsed = 0.03462315 sec  GFLOPS=620.245@@@@ sgemm test OK

Testing dgemm#### args: ta=0 tb=0 m=1024 n=1024 k=1024  alpha = (0xbff0000000000000, -1) beta= (0x0000000000000000, 0)#### args: lda=1024 ldb=1024 ldc=1024

^^^^ elapsed = 1.09212208 sec  GFLOPS=19.6634@@@@ dgemm test OK

 

 ==== Running N=10 with streams ====

Testing sgemm#### args: ta=0 tb=0 m=1024 n=1024 k=1024  alpha = (0x40000000, 2) beta= (0x40000000, 2)#### args: lda=1024 ldb=1024 ldc=1024

^^^^ elapsed = 0.03504515 sec  GFLOPS=612.776@@@@ sgemm test OK

Testing dgemm#### args: ta=0 tb=0 m=1024 n=1024 k=1024  alpha = (0xbff0000000000000, -1) beta= (0x0000000000000000, 0)#### args: lda=1024 ldb=1024 ldc=1024

^^^^ elapsed = 1.09177494 sec  GFLOPS=19.6697@@@@ dgemm test OK

 

 ==== Running N=10 batched ====

Testing sgemm#### args: ta=0 tb=0 m=1024 n=1024 k=1024  alpha = (0x3f800000, 1) beta= (0xbf800000, -1)#### args: lda=1024 ldb=1024 ldc=1024

^^^^ elapsed = 0.03766394 sec  GFLOPS=570.17@@@@ sgemm test OK

Testing dgemm#### args: ta=0 tb=0 m=1024 n=1024 k=1024  alpha = (0xbff0000000000000, -1) beta= (0x4000000000000000, 2)#### args: lda=1024 ldb=1024 ldc=1024

^^^^ elapsed = 1.09389901 sec  GFLOPS=19.6315@@@@ dgemm test OK

 

Test Summary0 error(s)

四、内存带宽测试

nvidia@tegra-ubuntu:/usr/local/cuda/samples/1_Utilities/bandwidthTest$ ./bandwidthTest

[CUDA Bandwidth Test] - Starting...

Running on...

 

 Device 0: NVIDIA Tegra X2

 Quick Mode

 

 Host to Device Bandwidth, 1 Device(s)

 PINNED Memory Transfers

   Transfer Size (Bytes)    Bandwidth(MB/s)

   33554432            20215.8

 

 Device to Host Bandwidth, 1 Device(s)

 PINNED Memory Transfers

   Transfer Size (Bytes)    Bandwidth(MB/s)

   33554432            20182.2

 

 Device to Device Bandwidth, 1 Device(s)

 PINNED Memory Transfers

   Transfer Size (Bytes)    Bandwidth(MB/s)

   33554432            35742.8

 

Result = PASS

 

NOTE: The CUDA Samples are not meant for performance measurements. Results may vary when GPU Boost is enabled.

五、设备查询

nvidia@tegra-ubuntu:~/work/TensorRT/tmp/usr/src/tensorrt$ cd /usr/local/cuda/samples/1_Utilities/deviceQuery

nvidia@tegra-ubuntu:/usr/local/cuda/samples/1_Utilities/deviceQuery$ ls

deviceQuery  deviceQuery.cpp  deviceQuery.o  Makefile  NsightEclipse.xml  readme.txt

nvidia@tegra-ubuntu:/usr/local/cuda/samples/1_Utilities/deviceQuery$ ./deviceQuery

./deviceQuery Starting...

 

 CUDA Device Query (Runtime API) version (CUDART static linking)

 

Detected 1 CUDA Capable device(s)

 

Device 0: "NVIDIA Tegra X2"

  CUDA Driver Version / Runtime Version          8.0 / 8.0

  CUDA Capability Major/Minor version number:    6.2

  Total amount of global memory:                 7851 MBytes (8232062976 bytes)

  ( 2) Multiprocessors, (128) CUDA Cores/MP:     256 CUDA Cores

  GPU Max Clock rate:                            1301 MHz (1.30 GHz)

  Memory Clock rate:                             1600 Mhz

  Memory Bus Width:                              128-bit

  L2 Cache Size:                                 524288 bytes

  Maximum Texture Dimension Size (x,y,z)         1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)

  Maximum Layered 1D Texture Size, (num) layers  1D=(32768), 2048 layers

  Maximum Layered 2D Texture Size, (num) layers  2D=(32768, 32768), 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: 32768

  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 1 copy engine(s)

  Run time limit on kernels:                     No

  Integrated GPU sharing Host Memory:            Yes

  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 / 0 / 0

  Compute Mode:

     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

 

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 8.0, CUDA Runtime Version = 8.0, NumDevs = 1, Device0 = NVIDIA Tegra X2Result = PASS

 

六、大型项目的测试

详情查看https://developer.nvidia.com/embedded/jetpack

这里面还有一些项目