Intel® DevCloud
Help for those needing help starting or connecting to the Intel® DevCloud

Cannot run openVINO example on DevCloud FPGA instance



I was trying to run the object detection example from OpenVINO with FPGA support on DevCloud. However, the example dose not work as expected. 

The scripts I used for environment setup is as followed. I used fpga_runtime instance in interactive mode. There are many bitstreams in the installation folder with "SSD" or "RMNET" in their names, I just randomly chose one as the DLA_AOCX. Not sure if this is the cause.

qsub -I -l nodes=1:fpga_runtime:ppn=2 -d . 
source /glob/development-tools/versions/oneapi/beta06/inteloneapi/ sh 
source /glob/deep-learning/versions/openvino_2019R1/openvino_fpga_2019.1.133/bin/ 
export DLA_AOCX=/glob/deep-learning/versions/openvino_2019R1/openvino_2019.1.133/bitstreams/a10_dcp_bitstreams/2019R1_RC_FP11_SSD300.aocx

When I ran the example, I got the following error. The program complains that that it failed to initialize the DMA, and then hangs forever after logging "Start inference (1 iterations)". The Python example does not work either, the Inference Engine throws the same error and everything fallback to CPU execution.

$ ./object_detection_sample_ssd -m open_model_zoo/model_downloader/Retail/object_detection/pedestrian/rmnet_ssd/0013/dldt/person-detection-retail-0013.xml -i demo/car.png -d HETERO:FPGA,CPU 
[ INFO ] InferenceEngine:
	API version ............ 1.6
	Build .................. custom_releases/2019/R1_ebce9728578ef3131f2f282b3fbc3232109c598e
Parsing input parameters
[ INFO ] Files were added: 1
[ INFO ]     openvino_demos/demo/car.png
[ INFO ] Loading plugin

	API version ............ 1.6
	Build .................. heteroPlugin
	Description ....... heteroPlugin
[ INFO ] Loading network files:
[ INFO ] Preparing input blobs
[ INFO ] Batch size is 1
[ INFO ] Preparing output blobs
[ INFO ] Loading model to the plugin
Error initializing DMA: not found
Error initializing DMA
[ WARNING ] Image is resized from (787, 259) to (544, 320)
[ INFO ] Batch size is 1
[ INFO ] Start inference (1 iterations)

Does anyone know hot to solve this issue? 

0 Kudos
1 Reply

Hi Shaojie,

Greetings to you.

From observation, your setvars script seems to have a discrepancy. Please source the script omitting the sh at the end of line, as follows:

source /glob/development-tools/versions/oneapi/beta06/inteloneapi/


Additionally, I am also moving this topic to DevCloud for Edge forum for further assistance.

For future DevCloud related questions, I would suggest posting in the DevCloud forum.

Link :




0 Kudos