Run the Application Using the Bloom16x Kernel - 2023.1 English

Vitis Tutorials: Hardware Acceleration (XD099)

Document ID
Release Date
2023.1 English

Navigate to the makefile directory, and run the make command.

cd $LAB_WORK_DIR/makefile; make run STEP=split_buffer TARGET=hw PF=16

The following output displays.

Processing 1398.903 MBytes of data
Split_buffer : Splitting data in 2 sub-buffers of 699.452 MBytes for FPGA processing
Executed FPGA accelerated version  |   733.6324 ms   ( FPGA 233.314 ms )
Executed Software-Only version     |   3133.5186 ms
Verification: PASS

You can see that if the documents are split into two buffers, and the overall application execution time using the Bloom8x kernel and Bloom16x kernel are very close. As expected, using the Bloom16x kernel rather than the Bloom8x kernel has no benefit.

While developing your own application, these attributes can be explored to make trade-offs and pick the optimal kernel implementation optimized for resources/performance.

You will continue this lab with the Bloom8x kernel.