Elements missing in DeepStream 6.2-devel container

[willemvdkletersteeg]

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU)
GPU, RTX 2080 Ti
• DeepStream Version
6.2
• JetPack Version (valid for Jetson only)
• TensorRT Version
• NVIDIA GPU Driver Version (valid for GPU only)
525.60.13, CUDA 12.0
• Issue Type( questions, new requirements, bugs)
Bug?
• How to reproduce the issue ? (This is for bugs. Including which sample app is using, the configuration files content, the command line used and other details for reproducing)
I’ve built a container using Dockerfile:

FROM nvcr.io/nvidia/deepstream:6.2-devel

# Extra libaries need to be installed manually as of 6.1.1
RUN /opt/nvidia/deepstream/deepstream-6.2/user_additional_install.sh
RUN /opt/nvidia/deepstream/deepstream-6.2/user_deepstream_python_apps_install.sh -v 1.1.6
[...]

Just like I would’ve done with prior versions. But somehow I’m missing a certain amount of deepstream elements?

root@a4df0b27390e:~# gst-inspect-1.0 nvinfer
No such element or plugin 'nvinfer'
root@a4df0b27390e:~# gst-inspect-1.0 nvstreammux
No such element or plugin 'nvstreammux'
root@a4df0b27390e:~# gst-inspect-1.0 | grep nvdsgst
nvdsgst_eglglessink:  nveglglessink: EGL/GLES vout Sink
nvdsgst_metautils:  nvdsmetainsert: nvdsmetainsert
nvdsgst_metautils:  nvdsmetaextract: nvdsmetaextract
nvdsgst_ofvisual:  nvofvisual: nvofvisual
nvdsgst_of:  nvof: nvof
nvdsgst_text_to_speech:  nvds_text_to_speech: DS Text To Speech Plugin for Conversational AI use cases
nvdsgst_ucx:  nvdsucxserversink: UCX server sink
nvdsgst_ucx:  nvdsucxclientsrc: UCX client source
nvdsgst_ucx:  nvdsucxclientsink: UCX client sink
nvdsgst_ucx:  nvdsucxserversrc: UCX server source
nvdsgst_audiotemplate:  nvdsaudiotemplate: DS AUDIO template Plugin for Transform IP use-cases
nvdsgst_msgconv:  nvmsgconv: Message Converter
nvdsgst_videotemplate:  nvdsvideotemplate: NvDsVideoTemplate plugin for Transform/In-Place use-cases
nvdsgst_postprocess:  nvdspostprocess: NvDsPostProcess plugin for Transform/In-Place use-cases
nvdsgst_3dfilter:  nvds3dfilter: DS 3D filter custom plugin
nvdsgst_dsanalytics:  nvdsanalytics: DsAnalytics plugin
nvdsgst_msgbroker:  nvmsgbroker: Message Broker
nvdsgst_speech:  nvdsasr: DS ASR Plugin for speech use-cases
nvdsgst_segvisual:  nvsegvisual: nvsegvisual
nvdsgst_metamux:  nvdsmetamux: meta muxer
nvdsgst_inferaudio:  nvinferaudio: NvInfer Audio plugin
nvdsgst_jpegdec:  nvjpegdec: JPEG image decoder
nvdsgst_xfer:  nvdsxfer: NvDsXfer plugin
nvdsgst_nvmultiurisrcbin:  nvmultiurisrcbin: NvMultiUri Bin
nvdsgst_logger:  nvdslogger: Nvdslogger
nvdsgst_videotestsrc:  nvvideotestsrc: NVIDIA GPU Video Test Source
nvdsgst_osd:  nvdsosd: NvDsOsd plugin

Where did they all go? Isn’t this container supposed to be an “all-in-one” solution?

[yuweiw]

Yes, it’s an “all-in-one” solution. But if your environment is not configured properly, some related plugins may not work. You can check your environment by the following link first:
https://docs.nvidia.com/metropolis/deepstream/dev-guide/text/DS_Quickstart.html?highlight=compatibility#id6

[willemvdkletersteeg]

Ah, thanks for pointing this out. I hadn’t seen this table yet. But the interesting thing is: we’ve been running DS6.1 and DS6.1.1 without any problems on our set of RTX2080’s. It says RTX Ampere (3xxx) from 6.1 onwards and this hasn’t changed with 6.2. Strange.

Also I think it’s just unexpected behaviour to switch off certain capabilities without any warning or message. I’ve been looking for the problem for hours. Maybe this should be in the README at least, that certain elements will be switched off or invisible when running on incompatible hardware. I would expect some kind of explanatory error message.

[yuweiw]

There is no update from you for a period, assuming this is not an issue anymore. Hence we are closing this topic. If need further support, please open a new one. Thanks

Basically, we sugget you install the relevant dependencies according to the compatibility table on your host. Like DS 6.2 needs CUDA 11.8, but your env is CUDA 12.0. It may cause some compatibility problems. Thanks

[rch23]

Is it the nvcc --version CUDA version that matters or nvidia-smi CUDA version?
My nvcc shows 11.8, while nvidia-smi is 12.1