{"id":924,"date":"2022-12-09T12:14:52","date_gmt":"2022-12-09T05:14:52","guid":{"rendered":"https:\/\/dgway.com\/blog_E\/?p=924"},"modified":"2022-12-09T12:50:06","modified_gmt":"2022-12-09T05:50:06","slug":"moving-average-accelerator-real-time-data-stream-processing","status":"publish","type":"post","link":"https:\/\/dgway.com\/blog_E\/2022\/12\/09\/moving-average-accelerator-real-time-data-stream-processing\/","title":{"rendered":"Moving Average Accelerator (Real-time data stream processing)"},"content":{"rendered":"
\n
\"\"
Moving Average Accelerator (Real-time data stream processing)
https:\/\/github.com\/dg-hpcdev\/alveo-simple-examples\/tree\/main\/03_moving_average<\/a><\/figcaption><\/figure>\n<\/div>\n\n\n

This example demonstrate a free-running kernel. Free-running kernel is a kernel which always repeatedly execute on its own whearas normal kernel requires user command (such as q.enqueueTask(krnl)<\/code>). One scenario one might use a free-running kernel is processing a data stream from other kernel (both RTL\/HLS) without any CPU involvement.<\/p>\n\n\n\n

    \n
  • The write_stream<\/code> kernel read from a buffer of floats then write them to stream one by one. In real projects, this could be replaced by other data producing kernel such as streaming data from TCP<\/li>\n\n\n\n
  • The simple_moving_average<\/code> kernel read the data stream from write_stream<\/code> and calculate simple moving average with window size = 20 and write the result to another stream. For the sake of simplicity, the edge cases(the beginning and the end of data stream) are ignored: the sum are always divided by 20.<\/li>\n\n\n\n
  • The read_stream<\/code> kernel read from simple_moving_average<\/code> kernel’s output stream and write them to FPGA’s global memory waiting for user command to migrate them to host PC. If you want the kernel to directly write to host memory, see Host Memory.<\/li>\n\n\n\n
  • The exponential_moving_average<\/code> kernel is similar to simple_moving_average<\/code> kernel but calculates exponential moving average instead and could be interchanged. This is provided for demonstrating how to modify configuration file.<\/li>\n<\/ul>\n\n\n\n

    Configuration File<\/h2>\n\n\n\n

    These lines specify which kernels are instantiated and how many. The basic syntax is nk=kernel_name:number_of_kernel:name_1,name_2<\/code> (name is optional)<\/p>\n\n\n\n

    nk=simple_moving_average:1\nnk=write_stream:1\nnk=read_stream:1<\/code><\/pre>\n\n\n\n
    These lines maps kernel port to memory.<\/p>\n\n\n\n
    sp=write_stream_1.buf:DDR[0]\nsp=read_stream_1.out:DDR[0]<\/code><\/pre>\n\n\n\n
    These lines specify stream connections between kernels<\/p>\n\n\n\n
    stream_connect=write_stream_1.data_stream:simple_moving_average_1.in\nstream_connect=simple_moving_average_1.out:read_stream_1.data_stream<\/code><\/pre>\n\n\n\n
    Swapping Kernel<\/h2>\n\n\n\n
    Suppose you want to change from simple_moving_average<\/code> to exponential_moving_average<\/code><\/p>\n\n\n\n
    First, Instantiate exponential_moving_average<\/code> instead of simple_moving_average<\/code>.<\/p>\n\n\n\n
    [connectivity]\nnk=exponential_moving_average:1\nnk=write_stream:1\nnk=read_stream:1<\/code><\/pre>\n\n\n\n
    Then, correctly change the stream connections.<\/p>\n\n\n\n
    stream_connect=write_stream_1.data_stream:exponential_moving_average_1.in\nstream_connect=exponential_moving_average_1.out:read_stream_1.data_stream<\/code><\/pre>\n\n\n\n
    Setup<\/h2>\n\n\n\n
    This must be done everytime a new terminal is opened<\/p>\n\n\n\n
    source \/opt\/xilinx\/xrt\/setup.sh\n# Replace <Vitis install path> and <vesion>\nsource <Vitis install path>\/Vitis\/<version>\/settings64.sh\nexport PLATFORM_REPO_PATHS=\/opt\/xilinx\/platforms\n# Change to appropiate platform\nexport PLATFORM=xilinx_u250_gen3x16_xdma_4_1_202210_1<\/code><\/pre>\n\n\n\n
    For software emulation:<\/p>\n\n\n\n
    export XCL_EMULATION_MODE=sw_emu\ncd sw_emu<\/code><\/pre>\n\n\n\n
    For hardware emulation:<\/p>\n\n\n\n
    export XCL_EMULATION_MODE=hw_emu\ncd hw_emu<\/pre>\n\n\n\n
    For hardware:<\/p>\n\n\n\n
    cd hw<\/code><\/pre>\n\n\n\n
    Build Host Software<\/h2>\n\n\n\n
    g++ -Wall -g -std=c++11 ..\/src\/host.cpp -o moving_average -I${XILINX_XRT}\/include\/ -L${XILINX_XRT}\/lib\/ -lOpenCL -pthread -lrt -lstdc++<\/code><\/pre>\n\n\n\n
    Build And Link Kenel<\/h2>\n\n\n\n
    # Replace ${TARGET} with \/ Set TARGET to:\n#  sw_emu if targeting software emulation\n#  hw_emu if targeting hardware emulation\n#  hw if targeting hardware\nv++ -c -t ${TARGET} --platform ${PLATFORM} -k exponential_moving_average -I..\/src ..\/src\/exponential_moving_average.cpp -o exponential_moving_average.xo \nv++ -c -t ${TARGET} --platform ${PLATFORM} -k simple_moving_average -I..\/src ..\/src\/simple_moving_average.cpp -o simple_moving_average.xo \nv++ -c -t ${TARGET} --platform ${PLATFORM} -k write_stream -I..\/src ..\/src\/write_stream.cpp -o write_stream.xo \nv++ -c -t ${TARGET} --platform ${PLATFORM} -k read_stream -I..\/src ..\/src\/read_stream.cpp -o read_stream.xo \nv++ -l -t ${TARGET} --platform ${PLATFORM} --config ..\/src\/moving_average.cfg .\/simple_moving_average.xo .\/read_stream.xo .\/write_stream.xo -o moving_average.xclbin<\/code><\/pre>\n\n\n\n
    Configure Emulator<\/h2>\n\n\n\n
    Only when targeting software\/hardware emulation<\/p>\n\n\n\n
    emconfigutil --platform ${PLATFORM} --nd 1<\/code><\/pre>\n\n\n\n
    Run the host software<\/h2>\n\n\n\n
    .\/moving_average moving_average.xclbin<\/code><\/pre>\n","protected":false},"excerpt":{"rendered":"
    This example demonstrate a free-running kernel. Free-running kernel is a kernel which always repeatedly execute on its own whearas normal kernel requires user command (such as q.enqueueTask(krnl)). One scenario one might use a free-running kernel is processing a data stream from other kernel (both RTL\/HLS) without any CPU involvement. Configuration File These lines specify which kernels are instantiated and how many....<\/p>\n","protected":false},"author":1,"featured_media":943,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[18],"tags":[],"class_list":["post-924","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-hls-development-series"],"_links":{"self":[{"href":"https:\/\/dgway.com\/blog_E\/wp-json\/wp\/v2\/posts\/924","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/dgway.com\/blog_E\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/dgway.com\/blog_E\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/dgway.com\/blog_E\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/dgway.com\/blog_E\/wp-json\/wp\/v2\/comments?post=924"}],"version-history":[{"count":5,"href":"https:\/\/dgway.com\/blog_E\/wp-json\/wp\/v2\/posts\/924\/revisions"}],"predecessor-version":[{"id":948,"href":"https:\/\/dgway.com\/blog_E\/wp-json\/wp\/v2\/posts\/924\/revisions\/948"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/dgway.com\/blog_E\/wp-json\/wp\/v2\/media\/943"}],"wp:attachment":[{"href":"https:\/\/dgway.com\/blog_E\/wp-json\/wp\/v2\/media?parent=924"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/dgway.com\/blog_E\/wp-json\/wp\/v2\/categories?post=924"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/dgway.com\/blog_E\/wp-json\/wp\/v2\/tags?post=924"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}