Python test functions

test.py

class test.Test

class provide helper functions to create test cases

cmp(A, B)

compare if two arrays are totally the same

Parameters

  • A (ndarray) – dense matrix in the host memory

  • B (ndarray) – dense matrix in the host memory

cmpWithinTolerance(A, B)

compare if two arrays are equal within tolerance

Parameters

  • A (ndarray) – dense matrix in the host memory

  • B (ndarray) – dense matrix in the host memory

multiply_and_cmp(C, A, B, X, m, n, post_scale, pRelu_val=[1, 0])

calculate matrix multiply golden result matrix C ( C = relu ((A * B + bias) * postScale >> postShift) ) on CPU and compare result to FPGA result.

will save the results to text files if not equal

Parameters

  • C (ndarray) – output dense matrix in the host memory

  • A (ndarray) – dense matrix in the host memory

  • B (ndarray) – dense matrix in the host memory

  • X (ndarray) – dense matrix in the host memory

  • post_scale ((int,int)) – (postScale, postShift)

  • pRelu_val ((int,int)) – (PReLUScale, PReLUAlpha)

test_basic_randint(PE, xclbin_opts, post_scale, max_dim)

generate random input matrices with random sizes and run the test

Parameters

  • PE (int) – number of kernels - 1

  • post_scale ((int,int)) – (postScale, postShift)

  • max_dim (int) – control the max size generated by the test

test_basic_size(m, k, n, xclbin_opts, PE=0, post_scale=[1, 0])

generate random input matrices with given sizes and run the test

Parameters

  • m (int) – number of rows of first dense matrix

  • k (int) – number of cols of first dense matrix

  • n (int) – number of cols of second dense matrix

  • xclbin_opts (dictionary) – information read from config_info.dat used to build the xclbin

  • PE (int) – number of kernels - 1

  • post_scale ((int,int)) – (postScale, postShift)

get_freq(command)

return frequency of the xclbin

Parameters

command (string) – path to the xbutil in the machine

class test.FcnTest

class provide helper functions to create FCN test cases

test_basic_randint(PE, xclbin_opts, post_scale, RELU_scale, max_dim)

generate random input matrices with random sizes and run the test

Parameters

  • PE (int) – number of kernels - 1

  • post_scale ((int,int)) – (postScale, postShift)

  • RELU_scale ((int,int)) – (PReLUScale, PReLUAlpha)

  • max_dim (int) – control the max size generated by the test

class test.SpmvTest

class provide helper functions to create spmv test cases

multiply_and_cmp_spmv(row, col, data, m, k, nnz, B, C)

calculate sparse matrix multiply golden result matrix C on CPU and compare result to FPGA result.

Parameters

  • row (ndarray) – sparse matrix’s row indices

  • col (ndarray) – sparse matrix’s col indices

  • data (ndarray) – sparse matrix’s non-zero elements

  • m (int) – number of rows for this sparse matrix

  • k (int) – number of cols for this sparse matrix

  • nnz (int) – number of non-zero elements of this sparse matrix

  • C (ndarray) – output dense matrix in the host memory

  • B (ndarray) – dense matrix in the host memory

fillMod(B, size, Max=32764)

fill array with values (not random)

Parameters

  • B (ndarray) – dense matrix in the host memory

  • size (int) – size of the array

  • Max (int) – control the max value filled to the array