I have installed the new DriveWork SDK from NVidia and wanted to run the sample_object_detector. When doing so (on Ubuntu Computer) the following error occured:
Platform: Detected Generic x86 Platform
Initialize DriveWorks SDK v1.5.3584
Release build with GNU 4.9.4 from heads/buildbrain-branch-0-g38b3a75
TimeSource: monotonic epoch time offset is 1550192108463707
Driveworks exception thrown: DW_CUDA_ERROR: Platform: cannot acquire cuda context : unkown error
terminate called after throwing an instance of 'std::runtime_error'
what(): [] DW Error DW_CUDA_ERROR executing DW function:
dwInitialize(&m_sdk, DW_VERSION, &SDKParams)
at /dvs/git/dirty/gitlab-master_av/dw/sdk/samples/dnn/sample_object_detector/main.cpp:135
Aborted (core dumped)
CUDA 10.0 is installed and the installed GPU Driver Version is: 410.78.
Does anyone know a solution to this Problem?
Dear HV_ZF,
can you check ldd of sample_object_detector to confirm if it gets linked to right CUDA so files. If not please set LD_LIBRARY_PATH accordingly.
/usr/local/driveworks-1.5/bin$ ./sample_object_detector
[26-02-2019 10:34:26] Platform: Detected Generic x86 Platform
[26-02-2019 10:34:26] Initialize DriveWorks SDK v1.5.3584
[26-02-2019 10:34:26] Release build with GNU 4.9.4 from heads/buildbrain-branch-0-g38b3a75
[26-02-2019 10:34:26] TimeSource: monotonic epoch time offset is 1551171285306328
[26-02-2019 10:34:26] Platform: number of GPU devices detected 1
[26-02-2019 10:34:26] Platform: currently selected GPU device discrete ID 0
[26-02-2019 10:34:26] SDK: Resources mounted from .././data/resources
[26-02-2019 10:34:26] SensorFactory::createSensor() -> camera.virtual, profiling=1,tensorRT_model=,video=.././data/samples/sfm/triangulation/video_0.h264
[26-02-2019 10:34:26] CameraNVCUVID: no seek table found at .././data/samples/sfm/triangulation/video_0.h264.seek, seeking is not available.
SimpleCamera: Camera image: 1280x800
Camera image with 1280x800 at 30 FPS
[26-02-2019 10:34:26] Glob Size is 46033424 bytes.
[26-02-2019 10:34:26] Added linear block of size 30670848
[26-02-2019 10:34:26] Added linear block of size 23003136
[26-02-2019 10:34:26] Added linear block of size 7667712
[26-02-2019 10:34:26] Added linear block of size 7667712
sample_object_detector: caskConvolutionLayer.cpp:145: virtual void nvinfer1::task::caskConvolutionLayer::allocateResources(const nvinfer1::cudnn::CommonContext&): Assertion `configIsValid(context)' failed.
Aborted (core dumped)
/usr/local/cuda-10.0/samples/bin/x86_64/linux/release$ ./deviceQuery
./deviceQuery Starting...
CUDA Device Query (Runtime API) version (CUDART static linking)
Detected 1 CUDA Capable device(s)
Device 0: "Quadro M2000M"
CUDA Driver Version / Runtime Version 10.0 / 10.0
CUDA Capability Major/Minor version number: 5.0
Total amount of global memory: 4042 MBytes (4238802944 bytes)
( 5) Multiprocessors, (128) CUDA Cores/MP: 640 CUDA Cores
GPU Max Clock rate: 1137 MHz (1.14 GHz)
Memory Clock rate: 2505 Mhz
Memory Bus Width: 128-bit
L2 Cache Size: 2097152 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 1 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 supports Compute Preemption: No
Supports Cooperative Kernel Launch: No
Supports MultiDevice Co-op Kernel Launch: No
Device PCI Domain ID / Bus ID / location ID: 0 / 1 / 0
Compute Mode:
< Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >
deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 10.0, CUDA Runtime Version = 10.0, NumDevs = 1
Result = PASS
Dear HV_ZF,
It seems you are having maxwell based GPU. But minimim pascal based GPU is recommended to run driveworks samples on host. I think that is the reason. Let us know if you notice this issue with pascal based cards.
Thank you for your help!
That pascal is the minimum required architecture was not written in the release notes for the latest DriveWorks. Maybe it would be a good idea to add this (I only found the minimum requirements, when I was searching, in the Drive_Software_Docs).
In the Release Notes for DriveWorks1.2 from last October the minimum requirements were given (mininum was Maxwell) and thus I assumed that they didn’t changed, because it was not mentioned.
To test it on a pascal based card, is currently difficult because I do not have access to one.