Whether there will be a result output when vxGraphParameterEnqueueReadyRef is enqueued at a slower speed than the TIDL model calculation speed. For example, the queue time is 1 second, and the model calculation time is 0.1 seconds.
This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Whether there will be a result output when vxGraphParameterEnqueueReadyRef is enqueued at a slower speed than the TIDL model calculation speed. For example, the queue time is 1 second, and the model calculation time is 0.1 seconds.
For example, in the multi-camera demo,
1680.07> GRAPH: app_multi_cam_graph (#nodes = 3, #executions = 3934)
1680.07> NODE: CAPTURE1: capture_node: avg = 33255 usecs, min/max = 32172 / 54692 usecs, #executions = 3934
1680.07> NODE: DISPLAY1: DisplayNode: avg = 11348 usecs, min/max = 70 / 16930 usecs, #executions = 3934
1680.09> NODE: VPAC_LDC1: ldc_node: avg = 13993 usecs, min/max = 13826 / 14570 usecs, #executions = 3934
NODE: CAPTURE1 takes the most time,
NODE: VPAC_LDC1 takes less time,
If captureObj.raw_image_arr enqueued for 1 second, and LDC processing is much less than 1 second, will there be a result output in this case?
Hi,
If a producer node is running slower than the consumer node, the consumer node would eventually get starved for buffers and be slowed down by the producer node.
However, it would still produce outputs, it is just that the compute resource of the consumer node may be idle while the producer is creating outputs.
It might be helpful to generate a timing diagram using the RT logging feature like the below to visualize:
software-dl.ti.com/.../TIOVX_DEBUG.html