Use mkl to generate the dataset; the usage of this script is: ./run_gemm_mkl.sh number_of_thread datatype g(generate)/b(benchmark).
Then use run_gemm_bench.sh
to run the benchmark.
cd ../gemm_mkl ./run_gemm_mkl.sh 16 float g ./run_gemm_bench.sh build_dir.hw.xilinx_u250_gen3x16_xdma_3_1_202020_1/blas.xclbin build_dir.hw.xilinx_u250_gen3x16_xdma_3_1_202020_1/config_info.dat