Tesla P100 Deepstream 4.0.1 run failed on some video files

Accelerated Computing Intelligent Video Analytics DeepStream SDK
#1

I am not having able to run the “/deepstream_sdk_v4.0.1_x86_64/sources/objectDetector_Yolo” sample using default deepstream-app with my own video on Tesla P100. I have only changed the uri path in the source.

I am using

Below is the error

**PERF: FPS 0 (Avg)
**PERF: 0.00 (0.00)
** INFO: <bus_callback:189>: Pipeline ready

** INFO: <bus_callback:175>: Pipeline running

ERROR from src_bin_muxer: Input buffer number of surfaces (-1229539658) must be equal to mux->num_surfaces_per_frame (1)
        Set nvstreammux property num-surfaces-per-frame appropriately

Debug info: gstnvstreammux.c(309): gst_nvstreammux_chain (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstNvStreamMux:src_bin_muxer
ERROR from qtdemux0: Internal data stream error.
Debug info: qtdemux.c(6073): gst_qtdemux_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin0/GstURIDecodeBin:src_elem/GstDecodeBin:decodebin0/GstQTDemux:qtdemux0:
streaming stopped, reason error (-5)
Creating LL OSD context new
Quitting
App run failed
#2

HI
Please check streammux part in config you used, you should keep batch-size same as your source input number.

#3

Hi amycao,

I tried it but still the same error.

#4

Can you run with builtin stream successfully?

#5

The builtin streams are working. I ran the same on the Jetson tx2, it was able to run the video but on P100 it gave an error. In both the cases of P100 and Tx2 the only change was different docker containers.

#6

you run with builtin stream on tx2 working，same stream but not work on p100，right？or your own video on the two platform？

#7

The builtin streams are both working on Tx2 and P100.

The new video is working on Tx2. But is not working on P100 giving the above error.

#8

Can you share us the video for further repro and check?

#9

Hi mohit21sharma.ms,

We haven’t heard back from you in a couple weeks, so marking this issue resolved by “accepting” this comment as answer.
Please open a new forum issue when you are ready and we’ll pick it up there.

#10