xfopencv error in makefile

[dhruvchamania]

using 2017.4, getting this error while making the file(included the code. in [makefile] as well as the what we input in terminal) for the sobel detector.

ERROR: [SYNCHK 200-79] Cannot find the top function 'w0_xf_Sobel' in the design. Possible causes are: (1) the top function name is misspelled; (2) the top function is nonexistent or declared as static. ERROR: [HLS 200-70] Synthesizability check failed. command 'ap_source' returned error code while executing "source /home/kushalchaudhari/Projects/_project/SDSoC_platform/xfopencv/examples/sobelfilter/_sds/vhls/w0_xf_Sobel_run.tcl" invoked from within "hls::main /home/kushalchaudhari/Projects/_project/SDSoC_platform/xfopencv/examples/sobelfilter/_sds/vhls/w0_xf_Sobel_run.tcl -l w0_xf_Sobel_viva..." ("uplevel" body line 1) invoked from within "uplevel 1 hls::main {*}$args" (procedure "hls_proc" line 5) invoked from within "hls_proc $argv" INFO: [Common 17-206] Exiting vivado_hls at Wed Jul 11 12:18:03 2018... ERROR: [SdsCompiler 83-5019] Exiting sds++ : Error when calling '/home/kushalchaudhari/Xilinx/Vivado/2017.4/bin/vivado_hls /home/kushalchaudhari/Projects/_project/SDSoC_platform/xfopencv/examples/sobelfilter/_sds/vhls/w0_xf_Sobel_run.tcl -l w0_xf_Sobel_vivado_hls.log' sds++ log file saved as /home/kushalchaudhari/Projects/_project/SDSoC_platform/xfopencv/examples/sobelfilter/_sds/reports/sds_xf_sobel_accel.log ERROR: [SdsCompiler 83-5004] Build failed sds++ completed at Wed Jul 11 12:18:03 IST 2018

MAKEFILE

APPSOURCES = xf_sobel_tb.cpp xf_sobel_accel.cpp EXECUTABLE = sobel.elf VERBOSE = #-verbose PLATFORM = /home/kushalchaudhari/Projects/_project/example_desing/zip_file/zcu102_rv_ss SDSCC_FLAGS = -sds-pf ${PLATFORM} -std=c++11 ${VERBOSE} \ -sds-hw "xf::Sobel<0,3,0,1,1080, 1920,1>" xf_sobel_accel.cpp -files ../../include/imgproc/xf_sobel.hpp -clkid 2 -sds-end -dmclkid 2

sobel<XF_BORDER_CONSTANT,XF_8UC1,XF_16UC1,HEIGHT, WIDTH,XF_NPPC8>(imgInput, imgOutputx,imgOutputy);

CC = sds++ ${SDSCC_FLAGS}

IDIRS = -I. -I${SYSROOT}/usr/include -I ../../include

LDIRS = --sysroot=${SYSROOT} -L=/lib -L=/usr/lib -Wl,-rpath-link=${SYSROOT}/lib,-rpath-link=${SYSROOT}/usr/lib

LLIBS = -lopencv_imgcodecs -lopencv_core -llzma -ltiff -lpng16 -lz -ljpeg -lopencv_imgproc -ldl -lrt -lwebp -lopencv_features2d -lopencv_flann

LLIBS = -lopencv_core -lopencv_imgproc -lopencv_imgcodecs

OLEVEL = -ggdb

CFLAGS = -O3 -Wall ${OLEVEL} -c ${IDIRS} -DHLS_NO_XIL_FPO_LIB LFLAGS = ${OLEVEL} ${LDIRS} ${LLIBS}

SOURCES := $(patsubst %,./%,$(APPSOURCES)) OBJECTS := $(APPSOURCES:.cpp=.o) DEPS := $(OBJECTS:.o=.d) .PHONY: all all: ${EXECUTABLE} ${EXECUTABLE}: ${OBJECTS} ${CC} ${OBJECTS} ${LFLAGS} -o $@ -include ${DEPS} %.o: %.cpp ${CC} ${CFLAGS} $< -o $@ clean: ${RM} ${EXECUTABLE} ${OBJECTS} ultraclean: clean ${RM} -rf _sds sd_card ${EXECUTABLE}.bit ${EXECUTABLE} ${OBJECTS} copy: /bin/cp -prf ${XLNX_SRC_PATH}/sobelfilter/example . /bin/cp -prf ${XLNX_SRC_PATH}/sobelfilter/src .

TERMINAL source /home/kushalchaudhari/Xilinx/Vivado/2017.4/settings64.sh

export SYSROOT=/home/kushalchaudhari/Projects/_project/example_desing/zip_file/zcu102_rv_ss/sw/sysroot/

Hope this helps for you to resolve the issue.

[bgouthamb]

@dhruvchamania Please check whether you have changed the configuration through the macros in _xf_sobelconfig.h and _xf_configparams.h. If yes, you need to update the parameters in the hardware function in the Makefile accordingly.

[dhruvchamania]

Thanks for the reply, had not updated parameters. Now synthesis has begin, will update and close issue as soon as we test on ZC102. Another doubt till then, like we use caffe with python in opencv, can the same implementation be done with xfopencv?. Especially in a video file. If yes any detailed guides avaliable to implement that can you please provide.

[varunsanthaseelan]

@dhruvchamania , in order for you to run Python on zcu102, you will need a platform with a OS that supports python on it. The platforms for zcu102 that Xilinx ships today do not have that capability. Having said that, I would strongly encourage you to check out the PYNQ board and the computer vision overlays that can be used from within Python (https://github.com/Xilinx/PYNQ-ComputerVision).