How to set buffer-pool-size and batched-push-timeout in new nvstreammux

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU)
GPU
• DeepStream Version
6.2
• Issue Type( questions, new requirements, bugs)
questions
• Requirement details( This is for new requirement. Including the module name-for which plugin or for which sample application, the function description)

I have a pipeline with 32 rtsp sources, a nvstreammux and a detector(nvinfer); Detector is set with “intervel=15” and “batch-size=32”. Each rtsp source produces one frame every 40ms (fps 25), and it takes about 400ms for detector to deal with a batch. To make streammux push buffer to downstream seamlessly, I set “buffer-pool-size=16” so that it won’t stuck in detector. (streammux batch-size=32)
Since buffer-pool-size has been deprecated in new nvstreammux, how should I set configurations to get the same effect?
I’ve tried use a config-file like this:

[property]
adaptive-batching=1

algorithm-type=1

max-fps-control=0

overall-max-fps-n=640

overall-max-fps-d=1

overall-min-fps-n=100

overall-min-fps-d=1

max-same-source-frames= 1

And detector will receive buffer every 10ms. That’s not what I expected. What I want is that the nvstreammux should wait as long as possible to get enough frames in a batch until it reaches a timeout.And it can’t stuck even though the detector takes a long period of time. (The same as batched-push-timeout=40 and buffer-pool-size=16).What should I do?

There are already detailed explanation of the new nvstreammux parameters. Gst-nvstreammux New — DeepStream 6.2 Release documentation (nvidia.com)

You set the minimux fps to be 100 fps, so the nvstreammux do as you set.
Please follow the instructions.

How should I set configurations?

min-overall-fps and max-overall-fps need to be properly set.

a). The min-overall-fps shall be set to the highest framerate of all sources.

b) max-overall-fps shall be >= min-overall-fps Check Mux Config Properties for more information.

Please read Gst-nvstreammux New — DeepStream 6.2 Release documentation (nvidia.com)

Could you read my whole question again? I don’t think we are talking about the same question.

That is what you want. And the parameters should be set as what you want. You want the actaul FPS to be 25 or lower while you set the minimum FPS to be 100.

Please set the parameters according to your requirement.

I see that overall-min-fps-n=100 and overall-min-fps-d=1 means batched-push-timeout=10, what about buffer-pool-size? What should I do to set enough output buffer size so it won’t get stuck in the detector?

No. There is no such settings with the new nvstreammux.

Dose that mean if I add a queue between the nvstreammux and the detector, the nvstreammux will always push data downstream whenever it gets a full batch or reaches timeout?No such things like output buffers in old streammux?

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

There are output buffers. But they are not managed by a pool. So there is no such parameter with the new nvstreammux.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.