Repository navigation
Closed
Description
I have successfully built and tested BLIS with icc (2015) using make check. However, when I use gcc 7.3 together with the binutils 2.30 (on the same machine), some tests fail. As far as I can tell, only real test cases are affected. I ran configure with the options --enable-shared --enable-static --enable-blas --enable-cblas. Attached to this message you can find the blocks of output.testsuite where test cases fail.
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_sdotxf_nn_ccc 100 8 2.857 5.68e+00 FAILURE
blis_sdotxf_nc_ccc 100 8 3.433 1.21e+01 FAILURE
blis_sdotxf_cn_ccc 100 8 4.134 6.43e+00 FAILURE
blis_sdotxf_cc_ccc 100 8 3.883 7.20e+00 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_ddotxf_nn_ccc 100 8 2.883 1.01e+01 FAILURE
blis_ddotxf_nc_ccc 100 8 3.800 5.14e+00 FAILURE
blis_ddotxf_cn_ccc 100 8 4.061 5.18e+00 FAILURE
blis_ddotxf_cc_ccc 100 8 4.290 7.56e+00 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_sdotxaxpyf_nnnn_ccccc 100 8 2.832 6.23e+00 FAILURE
blis_sdotxaxpyf_nnnc_ccccc 100 8 4.051 1.05e+01 FAILURE
blis_sdotxaxpyf_nncn_ccccc 100 8 2.988 8.14e+00 FAILURE
blis_sdotxaxpyf_nncc_ccccc 100 8 3.902 1.10e+01 FAILURE
blis_sdotxaxpyf_ncnn_ccccc 100 8 4.488 8.43e+00 FAILURE
blis_sdotxaxpyf_ncnc_ccccc 100 8 4.469 5.85e+00 FAILURE
blis_sdotxaxpyf_nccn_ccccc 100 8 4.390 6.90e+00 FAILURE
blis_sdotxaxpyf_nccc_ccccc 100 8 4.476 9.72e+00 FAILURE
blis_sdotxaxpyf_cnnn_ccccc 100 8 4.494 8.90e+00 FAILURE
blis_sdotxaxpyf_cnnc_ccccc 100 8 4.408 6.50e+00 FAILURE
blis_sdotxaxpyf_cncn_ccccc 100 8 4.476 8.56e+00 FAILURE
blis_sdotxaxpyf_cncc_ccccc 100 8 4.488 7.03e+00 FAILURE
blis_sdotxaxpyf_ccnn_ccccc 100 8 4.396 2.20e+00 FAILURE
blis_sdotxaxpyf_ccnc_ccccc 100 8 4.501 8.39e+00 FAILURE
blis_sdotxaxpyf_cccn_ccccc 100 8 4.501 1.26e+01 FAILURE
blis_sdotxaxpyf_cccc_ccccc 100 8 4.402 7.32e+00 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_sgemv_nn_rcc 100 50 1.547 3.06e+00 FAILURE
blis_sgemv_nc_rcc 100 50 8.598 1.51e+00 FAILURE
blis_sgemv_cn_rcc 100 50 9.141 1.53e+00 FAILURE
blis_sgemv_cc_rcc 100 50 9.443 2.29e+00 FAILURE
blis_sgemv_tn_rcc 100 50 4.602 0.00e+00 PASS
blis_sgemv_tc_rcc 100 50 11.099 0.00e+00 PASS
blis_sgemv_hn_rcc 100 50 10.764 0.00e+00 PASS
blis_sgemv_hc_rcc 100 50 11.161 0.00e+00 PASS
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_dgemv_nn_rcc 100 50 4.808 1.57e+00 FAILURE
blis_dgemv_nc_rcc 100 50 6.468 1.38e+00 FAILURE
blis_dgemv_cn_rcc 100 50 6.734 1.56e+00 FAILURE
blis_dgemv_cc_rcc 100 50 6.671 1.46e+00 FAILURE
blis_dgemv_tn_rcc 100 50 2.066 0.00e+00 PASS
blis_dgemv_tc_rcc 100 50 6.523 0.00e+00 PASS
blis_dgemv_hn_rcc 100 50 6.757 0.00e+00 PASS
blis_dgemv_hc_rcc 100 50 6.906 0.00e+00 PASS
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_sgemv_nn_ccc 100 50 7.358 0.00e+00 PASS
blis_sgemv_nc_ccc 100 50 11.312 0.00e+00 PASS
blis_sgemv_cn_ccc 100 50 11.211 0.00e+00 PASS
blis_sgemv_cc_ccc 100 50 11.287 0.00e+00 PASS
blis_sgemv_tn_ccc 100 50 7.519 3.19e+00 FAILURE
blis_sgemv_tc_ccc 100 50 8.921 2.68e+00 FAILURE
blis_sgemv_hn_ccc 100 50 9.099 1.39e+00 FAILURE
blis_sgemv_hc_ccc 100 50 8.467 1.41e+00 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_dgemv_nn_ccc 100 50 5.325 0.00e+00 PASS
blis_dgemv_nc_ccc 100 50 6.821 0.00e+00 PASS
blis_dgemv_cn_ccc 100 50 6.789 0.00e+00 PASS
blis_dgemv_cc_ccc 100 50 6.831 0.00e+00 PASS
blis_dgemv_tn_ccc 100 50 5.963 1.49e+00 FAILURE
blis_dgemv_tc_ccc 100 50 6.757 1.41e+00 FAILURE
blis_dgemv_hn_ccc 100 50 6.784 1.34e+00 FAILURE
blis_dgemv_hc_ccc 100 50 6.878 2.70e+00 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_sger_nn_ccr 100 50 4.794 3.61e-02 FAILURE
blis_sger_nc_ccr 100 50 7.102 4.06e-03 FAILURE
blis_sger_cn_ccr 100 50 7.133 2.03e-02 FAILURE
blis_sger_cc_ccr 100 50 7.179 2.60e-05 MARGINAL
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_dger_nn_ccr 100 50 4.990 9.27e-03 FAILURE
blis_dger_nc_ccr 100 50 5.869 1.58e-02 FAILURE
blis_dger_cn_ccr 100 50 5.903 1.47e-01 FAILURE
blis_dger_cc_ccr 100 50 5.992 1.58e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_sger_nn_jjr 100 50 2.094 5.45e-02 FAILURE
blis_sger_nc_jjr 100 50 2.128 2.08e-04 FAILURE
blis_sger_cn_jjr 100 50 2.130 2.31e-02 FAILURE
blis_sger_cc_jjr 100 50 2.120 2.37e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_dger_nn_jjr 100 50 1.808 2.67e-02 FAILURE
blis_dger_nc_jjr 100 50 1.836 8.93e-02 FAILURE
blis_dger_cn_jjr 100 50 1.850 1.63e-02 FAILURE
blis_dger_cc_jjr 100 50 1.835 6.78e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_shemv_lnn_rcc 100 1.894 3.07e-02 FAILURE
blis_shemv_lnc_rcc 100 3.762 3.85e-02 FAILURE
blis_shemv_lcn_rcc 100 3.374 3.65e-02 FAILURE
blis_shemv_lcc_rcc 100 3.915 3.24e-02 FAILURE
blis_shemv_unn_rcc 100 1.627 3.39e-02 FAILURE
blis_shemv_unc_rcc 100 3.546 3.05e-02 FAILURE
blis_shemv_ucn_rcc 100 3.241 2.98e-02 FAILURE
blis_shemv_ucc_rcc 100 3.650 2.98e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dhemv_lnn_rcc 100 1.328 4.87e-02 FAILURE
blis_dhemv_lnc_rcc 100 1.426 3.79e-02 FAILURE
blis_dhemv_lcn_rcc 100 1.490 4.32e-02 FAILURE
blis_dhemv_lcc_rcc 100 1.759 5.20e-02 FAILURE
blis_dhemv_unn_rcc 100 1.235 5.14e-02 FAILURE
blis_dhemv_unc_rcc 100 1.410 4.61e-02 FAILURE
blis_dhemv_ucn_rcc 100 1.786 5.61e-02 FAILURE
blis_dhemv_ucc_rcc 100 1.917 4.69e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_shemv_lnn_ccc 100 2.800 3.04e-02 FAILURE
blis_shemv_lnc_ccc 100 3.618 2.64e-02 FAILURE
blis_shemv_lcn_ccc 100 3.197 2.54e-02 FAILURE
blis_shemv_lcc_ccc 100 3.584 3.18e-02 FAILURE
blis_shemv_unn_ccc 100 3.144 2.91e-02 FAILURE
blis_shemv_unc_ccc 100 3.573 2.95e-02 FAILURE
blis_shemv_ucn_ccc 100 3.132 3.26e-02 FAILURE
blis_shemv_ucc_ccc 100 3.632 3.12e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_sher_ln_cr 100 1.912 1.99e-02 FAILURE
blis_sher_lc_cr 100 0.787 3.73e-02 FAILURE
blis_sher_un_cr 100 3.613 4.55e-02 FAILURE
blis_sher_uc_cr 100 3.550 9.41e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dher_ln_cr 100 2.917 6.38e-05 FAILURE
blis_dher_lc_cr 100 3.030 5.98e-03 FAILURE
blis_dher_un_cr 100 3.200 2.95e-02 FAILURE
blis_dher_uc_cr 100 3.413 1.50e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_sher_ln_jr 100 1.977 1.94e-02 FAILURE
blis_sher_lc_jr 100 2.114 3.03e-02 FAILURE
blis_sher_un_jr 100 2.066 2.62e-02 FAILURE
blis_sher_uc_jr 100 2.089 1.57e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dher_ln_jr 100 1.756 3.70e-02 FAILURE
blis_dher_lc_jr 100 1.815 1.46e-02 FAILURE
blis_dher_un_jr 100 1.759 2.66e-03 FAILURE
blis_dher_uc_jr 100 1.760 4.56e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_sher2_lnn_ccr 100 3.657 2.72e-02 FAILURE
blis_sher2_lnc_ccr 100 4.433 2.75e-02 FAILURE
blis_sher2_lcn_ccr 100 4.660 3.31e-02 FAILURE
blis_sher2_lcc_ccr 100 5.227 2.20e-02 FAILURE
blis_sher2_unn_ccr 100 3.905 4.53e-02 FAILURE
blis_sher2_unc_ccr 100 4.302 2.61e-02 FAILURE
blis_sher2_ucn_ccr 100 4.800 4.10e-02 FAILURE
blis_sher2_ucc_ccr 100 4.672 4.97e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dher2_lnn_ccr 100 1.093 3.35e-02 FAILURE
blis_dher2_lnc_ccr 100 3.860 2.13e-02 FAILURE
blis_dher2_lcn_ccr 100 4.185 2.62e-02 FAILURE
blis_dher2_lcc_ccr 100 4.381 1.50e-02 FAILURE
blis_dher2_unn_ccr 100 3.118 2.02e-02 FAILURE
blis_dher2_unc_ccr 100 3.814 1.74e-02 FAILURE
blis_dher2_ucn_ccr 100 4.100 2.41e-02 FAILURE
blis_dher2_ucc_ccr 100 4.398 1.79e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_sher2_lnn_jjr 100 2.028 2.61e-02 FAILURE
blis_sher2_lnc_jjr 100 2.140 3.66e-02 FAILURE
blis_sher2_lcn_jjr 100 2.201 3.04e-02 FAILURE
blis_sher2_lcc_jjr 100 2.201 3.83e-02 FAILURE
blis_sher2_unn_jjr 100 2.115 2.10e-02 FAILURE
blis_sher2_unc_jjr 100 1.939 3.86e-02 FAILURE
blis_sher2_ucn_jjr 100 1.958 2.82e-02 FAILURE
blis_sher2_ucc_jjr 100 0.920 3.69e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dher2_lnn_jjr 100 1.803 3.04e-02 FAILURE
blis_dher2_lnc_jjr 100 1.827 1.75e-02 FAILURE
blis_dher2_lcn_jjr 100 1.855 1.11e-02 FAILURE
blis_dher2_lcc_jjr 100 1.868 4.22e-02 FAILURE
blis_dher2_unn_jjr 100 1.817 2.78e-02 FAILURE
blis_dher2_unc_jjr 100 1.795 2.27e-02 FAILURE
blis_dher2_ucn_jjr 100 1.857 5.87e-02 FAILURE
blis_dher2_ucc_jjr 100 1.836 1.52e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_ssymv_lnn_rcc 100 2.734 3.52e-02 FAILURE
blis_ssymv_lnc_rcc 100 3.791 3.03e-02 FAILURE
blis_ssymv_lcn_rcc 100 3.419 3.64e-02 FAILURE
blis_ssymv_lcc_rcc 100 3.861 3.36e-02 FAILURE
blis_ssymv_unn_rcc 100 2.579 2.84e-02 FAILURE
blis_ssymv_unc_rcc 100 3.420 3.12e-02 FAILURE
blis_ssymv_ucn_rcc 100 2.988 2.93e-02 FAILURE
blis_ssymv_ucc_rcc 100 3.702 3.00e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dsymv_lnn_rcc 100 1.465 4.29e-02 FAILURE
blis_dsymv_lnc_rcc 100 1.628 4.36e-02 FAILURE
blis_dsymv_lcn_rcc 100 1.533 4.71e-02 FAILURE
blis_dsymv_lcc_rcc 100 1.642 4.30e-02 FAILURE
blis_dsymv_unn_rcc 100 1.543 4.71e-02 FAILURE
blis_dsymv_unc_rcc 100 1.833 4.50e-02 FAILURE
blis_dsymv_ucn_rcc 100 1.788 4.00e-02 FAILURE
blis_dsymv_ucc_rcc 100 1.822 5.36e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_ssymv_lnn_ccc 100 2.767 2.23e-02 FAILURE
blis_ssymv_lnc_ccc 100 3.494 2.97e-02 FAILURE
blis_ssymv_lcn_ccc 100 2.998 2.96e-02 FAILURE
blis_ssymv_lcc_ccc 100 3.284 2.69e-02 FAILURE
blis_ssymv_unn_ccc 100 2.874 3.41e-02 FAILURE
blis_ssymv_unc_ccc 100 3.595 2.89e-02 FAILURE
blis_ssymv_ucn_ccc 100 3.399 3.24e-02 FAILURE
blis_ssymv_ucc_ccc 100 4.002 3.35e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_ssyr_ln_cr 100 3.423 1.56e-02 FAILURE
blis_ssyr_lc_cr 100 3.862 3.70e-02 FAILURE
blis_ssyr_un_cr 100 4.065 5.02e-02 FAILURE
blis_ssyr_uc_cr 100 3.860 9.55e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dsyr_ln_cr 100 3.312 9.01e-03 FAILURE
blis_dsyr_lc_cr 100 3.331 2.18e-02 FAILURE
blis_dsyr_un_cr 100 3.446 1.14e-04 FAILURE
blis_dsyr_uc_cr 100 3.155 1.83e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_ssyr_ln_jr 100 1.932 8.40e-03 FAILURE
blis_ssyr_lc_jr 100 2.084 2.83e-04 FAILURE
blis_ssyr_un_jr 100 2.064 1.21e-02 FAILURE
blis_ssyr_uc_jr 100 2.092 3.76e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dsyr_ln_jr 100 1.779 9.28e-03 FAILURE
blis_dsyr_lc_jr 100 1.821 4.14e-02 FAILURE
blis_dsyr_un_jr 100 1.805 8.72e-03 FAILURE
blis_dsyr_uc_jr 100 1.795 4.25e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_ssyr2_lnn_ccr 100 3.958 1.48e-02 FAILURE
blis_ssyr2_lnc_ccr 100 4.312 3.09e-02 FAILURE
blis_ssyr2_lcn_ccr 100 4.646 4.09e-02 FAILURE
blis_ssyr2_lcc_ccr 100 4.984 5.58e-02 FAILURE
blis_ssyr2_unn_ccr 100 4.356 3.64e-02 FAILURE
blis_ssyr2_unc_ccr 100 4.472 1.34e-02 FAILURE
blis_ssyr2_ucn_ccr 100 4.718 6.07e-03 FAILURE
blis_ssyr2_ucc_ccr 100 5.044 2.50e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dsyr2_lnn_ccr 100 3.668 2.23e-02 FAILURE
blis_dsyr2_lnc_ccr 100 3.922 2.17e-02 FAILURE
blis_dsyr2_lcn_ccr 100 4.256 4.42e-02 FAILURE
blis_dsyr2_lcc_ccr 100 3.937 2.93e-02 FAILURE
blis_dsyr2_unn_ccr 100 3.283 8.34e-02 FAILURE
blis_dsyr2_unc_ccr 100 3.534 1.87e-02 FAILURE
blis_dsyr2_ucn_ccr 100 3.675 3.92e-02 FAILURE
blis_dsyr2_ucc_ccr 100 4.119 5.99e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_ssyr2_lnn_jjr 100 2.039 7.09e-02 FAILURE
blis_ssyr2_lnc_jjr 100 2.163 3.33e-02 FAILURE
blis_ssyr2_lcn_jjr 100 2.181 1.55e-02 FAILURE
blis_ssyr2_lcc_jjr 100 2.199 2.81e-02 FAILURE
blis_ssyr2_unn_jjr 100 2.109 3.90e-02 FAILURE
blis_ssyr2_unc_jjr 100 2.114 4.25e-02 FAILURE
blis_ssyr2_ucn_jjr 100 2.155 3.01e-02 FAILURE
blis_ssyr2_ucc_jjr 100 2.146 4.03e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dsyr2_lnn_jjr 100 1.825 4.57e-02 FAILURE
blis_dsyr2_lnc_jjr 100 1.807 5.10e-03 FAILURE
blis_dsyr2_lcn_jjr 100 1.841 2.66e-02 FAILURE
blis_dsyr2_lcc_jjr 100 1.854 1.62e-02 FAILURE
blis_dsyr2_unn_jjr 100 1.811 2.40e-02 FAILURE
blis_dsyr2_unc_jjr 100 1.836 4.39e-02 FAILURE
blis_dsyr2_ucn_jjr 100 1.847 1.14e-02 FAILURE
blis_dsyr2_ucc_jjr 100 1.827 1.37e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_strmv_lnn_rc 100 3.724 3.05e-02 FAILURE
blis_strmv_lnu_rc 100 5.609 2.43e-02 FAILURE
blis_strmv_lcn_rc 100 4.423 2.88e-02 FAILURE
blis_strmv_lcu_rc 100 5.094 3.11e-02 FAILURE
blis_strmv_ltn_rc 100 2.408 1.12e-09 PASS
blis_strmv_ltu_rc 100 5.126 3.73e-09 PASS
blis_strmv_lhn_rc 100 5.605 1.21e-09 PASS
blis_strmv_lhu_rc 100 5.734 3.75e-09 PASS
blis_strmv_unn_rc 100 4.864 3.06e-02 FAILURE
blis_strmv_unu_rc 100 5.624 3.28e-02 FAILURE
blis_strmv_ucn_rc 100 5.435 3.01e-02 FAILURE
blis_strmv_ucu_rc 100 5.510 2.78e-02 FAILURE
blis_strmv_utn_rc 100 4.953 9.04e-09 PASS
blis_strmv_utu_rc 100 4.739 1.51e-07 PASS
blis_strmv_uhn_rc 100 4.819 8.97e-09 PASS
blis_strmv_uhu_rc 100 4.604 9.66e-08 PASS
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dtrmv_lnn_rc 100 3.779 3.14e-02 FAILURE
blis_dtrmv_lnu_rc 100 4.581 3.01e-02 FAILURE
blis_dtrmv_lcn_rc 100 4.421 2.82e-02 FAILURE
blis_dtrmv_lcu_rc 100 4.585 4.04e-02 FAILURE
blis_dtrmv_ltn_rc 100 4.462 2.21e-18 PASS
blis_dtrmv_ltu_rc 100 4.013 1.55e-17 PASS
blis_dtrmv_lhn_rc 100 4.080 3.64e-18 PASS
blis_dtrmv_lhu_rc 100 4.329 1.39e-17 PASS
blis_dtrmv_unn_rc 100 2.404 3.23e-02 FAILURE
blis_dtrmv_unu_rc 100 4.975 3.87e-02 FAILURE
blis_dtrmv_ucn_rc 100 4.572 3.68e-02 FAILURE
blis_dtrmv_ucu_rc 100 4.715 3.48e-02 FAILURE
blis_dtrmv_utn_rc 100 4.331 1.68e-17 PASS
blis_dtrmv_utu_rc 100 4.695 1.92e-16 PASS
blis_dtrmv_uhn_rc 100 4.617 1.34e-17 PASS
blis_dtrmv_uhu_rc 100 4.742 2.30e-16 PASS
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_strmv_lnn_cc 100 4.401 1.03e-08 PASS
blis_strmv_lnu_cc 100 4.888 1.02e-07 PASS
blis_strmv_lcn_cc 100 4.627 7.59e-09 PASS
blis_strmv_lcu_cc 100 4.782 1.24e-07 PASS
blis_strmv_ltn_cc 100 4.529 3.21e-02 FAILURE
blis_strmv_ltu_cc 100 5.653 3.01e-02 FAILURE
blis_strmv_lhn_cc 100 5.653 3.34e-02 FAILURE
blis_strmv_lhu_cc 100 5.534 2.99e-02 FAILURE
blis_strmv_unn_cc 100 5.858 1.24e-09 PASS
blis_strmv_unu_cc 100 5.851 4.66e-10 PASS
blis_strmv_ucn_cc 100 5.981 5.73e-10 PASS
blis_strmv_ucu_cc 100 5.834 7.68e-09 PASS
blis_strmv_utn_cc 100 4.926 3.29e-02 FAILURE
blis_strmv_utu_cc 100 5.698 2.40e-02 FAILURE
blis_strmv_uhn_cc 100 5.376 3.24e-02 FAILURE
blis_strmv_uhu_cc 100 4.762 2.99e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dtrmv_lnn_cc 100 3.858 1.37e-17 PASS
blis_dtrmv_lnu_cc 100 4.137 2.25e-16 PASS
blis_dtrmv_lcn_cc 100 4.566 1.47e-17 PASS
blis_dtrmv_lcu_cc 100 4.583 1.93e-16 PASS
blis_dtrmv_ltn_cc 100 4.496 3.36e-02 FAILURE
blis_dtrmv_ltu_cc 100 4.852 2.63e-02 FAILURE
blis_dtrmv_lhn_cc 100 4.773 2.89e-02 FAILURE
blis_dtrmv_lhu_cc 100 4.985 3.20e-02 FAILURE
blis_dtrmv_unn_cc 100 4.771 1.96e-18 PASS
blis_dtrmv_unu_cc 100 4.796 1.55e-17 PASS
blis_dtrmv_ucn_cc 100 4.869 2.17e-18 PASS
blis_dtrmv_ucu_cc 100 0.739 1.73e-18 PASS
blis_dtrmv_utn_cc 100 3.687 2.66e-02 FAILURE
blis_dtrmv_utu_cc 100 4.027 3.75e-02 FAILURE
blis_dtrmv_uhn_cc 100 3.814 2.70e-02 FAILURE
blis_dtrmv_uhu_cc 100 4.062 3.14e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_strsv_lnn_rc 100 2.112 1.45e-02 FAILURE
blis_strsv_lnu_rc 100 5.345 3.58e-02 FAILURE
blis_strsv_lcn_rc 100 4.829 1.50e-02 FAILURE
blis_strsv_lcu_rc 100 5.531 3.28e-02 FAILURE
blis_strsv_ltn_rc 100 1.698 1.69e-08 PASS
blis_strsv_ltu_rc 100 5.583 2.13e-08 PASS
blis_strsv_lhn_rc 100 5.005 2.39e-08 PASS
blis_strsv_lhu_rc 100 5.760 6.36e-08 PASS
blis_strsv_unn_rc 100 2.984 1.51e-02 FAILURE
blis_strsv_unu_rc 100 4.990 2.83e-02 FAILURE
blis_strsv_ucn_rc 100 4.411 1.58e-02 FAILURE
blis_strsv_ucu_rc 100 4.591 3.57e-02 FAILURE
blis_strsv_utn_rc 100 4.585 1.13e-07 PASS
blis_strsv_utu_rc 100 4.869 1.19e-07 PASS
blis_strsv_uhn_rc 100 4.221 1.15e-07 PASS
blis_strsv_uhu_rc 100 5.168 1.27e-07 PASS
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dtrsv_lnn_rc 100 2.982 1.41e-02 FAILURE
blis_dtrsv_lnu_rc 100 4.263 3.04e-02 FAILURE
blis_dtrsv_lcn_rc 100 3.546 1.25e-02 FAILURE
blis_dtrsv_lcu_rc 100 4.303 2.94e-02 FAILURE
blis_dtrsv_ltn_rc 100 2.184 8.82e-17 PASS
blis_dtrsv_ltu_rc 100 4.572 6.27e-17 PASS
blis_dtrsv_lhn_rc 100 3.470 3.32e-17 PASS
blis_dtrsv_lhu_rc 100 4.023 6.16e-17 PASS
blis_dtrsv_unn_rc 100 2.473 1.42e-02 FAILURE
blis_dtrsv_unu_rc 100 3.778 3.44e-02 FAILURE
blis_dtrsv_ucn_rc 100 3.101 1.53e-02 FAILURE
blis_dtrsv_ucu_rc 100 3.792 3.33e-02 FAILURE
blis_dtrsv_utn_rc 100 3.549 2.40e-16 PASS
blis_dtrsv_utu_rc 100 4.450 1.95e-16 PASS
blis_dtrsv_uhn_rc 100 3.693 2.48e-16 PASS
blis_dtrsv_uhu_rc 100 4.318 2.31e-16 PASS
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_strsv_lnn_cc 100 4.110 1.27e-07 PASS
blis_strsv_lnu_cc 100 5.811 1.03e-07 PASS
blis_strsv_lcn_cc 100 4.914 1.18e-07 PASS
blis_strsv_lcu_cc 100 5.371 1.25e-07 PASS
blis_strsv_ltn_cc 100 2.884 1.56e-02 FAILURE
blis_strsv_ltu_cc 100 4.798 2.74e-02 FAILURE
blis_strsv_lhn_cc 100 3.743 1.47e-02 FAILURE
blis_strsv_lhu_cc 100 5.028 2.74e-02 FAILURE
blis_strsv_unn_cc 100 4.780 1.39e-08 PASS
blis_strsv_unu_cc 100 5.714 1.61e-08 PASS
blis_strsv_ucn_cc 100 4.950 3.22e-08 PASS
blis_strsv_ucu_cc 100 5.656 5.11e-08 PASS
blis_strsv_utn_cc 100 4.550 1.63e-02 FAILURE
blis_strsv_utu_cc 100 5.473 2.88e-02 FAILURE
blis_strsv_uhn_cc 100 4.869 1.87e-02 FAILURE
blis_strsv_uhu_cc 100 5.501 3.49e-02 FAILURE
% blis_<dt><op>_<params>_<stor> m gflops resid result
blis_dtrsv_lnn_cc 100 3.801 2.12e-16 PASS
blis_dtrsv_lnu_cc 100 4.803 1.63e-16 PASS
blis_dtrsv_lcn_cc 100 4.090 1.88e-16 PASS
blis_dtrsv_lcu_cc 100 4.847 1.87e-16 PASS
blis_dtrsv_ltn_cc 100 3.446 1.57e-02 FAILURE
blis_dtrsv_ltu_cc 100 4.407 2.80e-02 FAILURE
blis_dtrsv_lhn_cc 100 3.658 1.58e-02 FAILURE
blis_dtrsv_lhu_cc 100 4.486 2.95e-02 FAILURE
blis_dtrsv_unn_cc 100 3.788 1.05e-17 PASS
blis_dtrsv_unu_cc 100 4.581 3.77e-17 PASS
blis_dtrsv_ucn_cc 100 3.693 3.57e-17 PASS
blis_dtrsv_ucu_cc 100 4.384 8.56e-17 PASS
blis_dtrsv_utn_cc 100 3.788 1.41e-02 FAILURE
blis_dtrsv_utu_cc 100 4.808 2.51e-02 FAILURE
blis_dtrsv_uhn_cc 100 3.943 1.44e-02 FAILURE
blis_dtrsv_uhu_cc 100 4.769 3.67e-02 FAILURE
% blis_<dt><op>_<stor> m n k gflops resid result
blis_sgemm_ukr_r 6 16 100 3.638 1.17e-02 FAILURE
% blis_<dt><op>_<stor> m n k gflops resid result
blis_dgemm_ukr_r 6 8 100 10.823 1.69e-02 FAILURE
% blis_<dt><op>_<stor> m n k gflops resid result
blis_cgemm_ukr_r 3 8 100 23.501 3.01e-08 PASS
% blis_<dt><op>_<stor> m n k gflops resid result
blis_zgemm_ukr_r 3 4 100 7.987 2.88e-17 PASS
% blis_<dt><op>_<stor> m n k gflops resid result
blis_sgemm_ukr_c 6 16 100 14.201 5.65e-03 FAILURE
% blis_<dt><op>_<stor> m n k gflops resid result
blis_dgemm_ukr_c 6 8 100 20.645 3.11e-03 FAILURE
% blis_<dt><op>_<stor> m n k gflops resid result
blis_cgemm_ukr_c 3 8 100 43.936 2.98e-08 PASS
% blis_<dt><op>_<stor> m n k gflops resid result
blis_zgemm_ukr_c 3 4 100 23.821 3.40e-17 PASS
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_sgemmtrsm_ukr_l_r 6 16 100 7.683 2.04e-03 FAILURE
blis_sgemmtrsm_ukr_u_r 6 16 100 5.439 2.04e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_dgemmtrsm_ukr_l_r 6 8 100 7.078 1.05e-03 FAILURE
blis_dgemmtrsm_ukr_u_r 6 8 100 5.854 1.86e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_cgemmtrsm_ukr_l_r 3 8 100 11.396 1.22e-08 PASS
blis_cgemmtrsm_ukr_u_r 3 8 100 11.719 1.39e-08 PASS
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_zgemmtrsm_ukr_l_r 3 4 100 12.030 2.11e-17 PASS
blis_zgemmtrsm_ukr_u_r 3 4 100 14.204 1.36e-17 PASS
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_sgemmtrsm_ukr_l_c 6 16 100 41.200 2.08e-03 FAILURE
blis_sgemmtrsm_ukr_u_c 6 16 100 39.083 3.16e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_dgemmtrsm_ukr_l_c 6 8 100 19.426 1.15e-03 FAILURE
blis_dgemmtrsm_ukr_u_c 6 8 100 18.552 1.82e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_cgemmtrsm_ukr_l_c 3 8 100 27.029 4.75e-09 PASS
blis_cgemmtrsm_ukr_u_c 3 8 100 29.982 3.13e-09 PASS
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_zgemmtrsm_ukr_l_c 3 4 100 15.178 1.55e-17 PASS
blis_zgemmtrsm_ukr_u_c 3 4 100 14.436 2.06e-17 PASS
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_sgemm_nn_rrr 100 100 100 19.979 4.50e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_dgemm_nn_rrr 100 100 100 19.495 4.23e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_cgemm_nn_rrr 100 100 100 58.010 2.41e-08 PASS
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_zgemm_nn_rrr 100 100 100 13.274 2.27e-17 PASS
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_sgemm_nn_ccc 100 100 100 45.638 1.96e-08 PASS
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_dgemm_nn_ccc 100 100 100 28.110 3.36e-17 PASS
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_cgemm_nn_ccc 100 100 100 62.698 2.60e-08 PASS
% blis_<dt><op>_<params>_<stor> m n k gflops resid result
blis_zgemm_nn_ccc 100 100 100 35.898 2.32e-17 PASS
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_shemm_llnn_ccc 100 100 43.239 3.90e-03 FAILURE
blis_shemm_lunn_ccc 100 100 45.403 3.39e-03 FAILURE
blis_shemm_rlnn_ccc 100 100 45.662 4.79e-03 FAILURE
blis_shemm_runn_ccc 100 100 44.332 3.20e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m k gflops resid result
blis_sherk_ln_rr 100 100 15.065 6.59e-03 FAILURE
blis_sherk_un_rr 100 100 24.422 4.50e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m k gflops resid result
blis_dherk_ln_rr 100 100 18.382 4.27e-03 FAILURE
blis_dherk_un_rr 100 100 18.340 5.82e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m k gflops resid result
blis_sher2k_lnn_rrr 100 100 29.417 4.86e-03 FAILURE
blis_sher2k_unn_rrr 100 100 29.300 5.11e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m k gflops resid result
blis_dher2k_lnn_rrr 100 100 20.368 4.43e-03 FAILURE
blis_dher2k_unn_rrr 100 100 20.136 4.77e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_ssymm_llnn_rrr 100 100 41.576 1.18e-03 FAILURE
blis_ssymm_lunn_rrr 100 100 45.099 2.14e-03 FAILURE
blis_ssymm_rlnn_rrr 100 100 44.704 3.13e-03 FAILURE
blis_ssymm_runn_rrr 100 100 45.411 2.87e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_dsymm_llnn_rrr 100 100 24.428 6.08e-04 FAILURE
blis_dsymm_lunn_rrr 100 100 27.147 7.66e-04 FAILURE
blis_dsymm_rlnn_rrr 100 100 27.445 2.81e-03 FAILURE
blis_dsymm_runn_rrr 100 100 27.217 3.61e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_ssymm_llnn_ccc 100 100 43.540 1.91e-03 FAILURE
blis_ssymm_lunn_ccc 100 100 39.789 1.94e-03 FAILURE
blis_ssymm_rlnn_ccc 100 100 44.787 1.52e-03 FAILURE
blis_ssymm_runn_ccc 100 100 45.945 1.84e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m k gflops resid result
blis_ssyrk_ln_rr 100 100 24.835 4.96e-03 FAILURE
blis_ssyrk_un_rr 100 100 27.755 5.82e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m k gflops resid result
blis_dsyrk_ln_rr 100 100 17.905 5.52e-03 FAILURE
blis_dsyrk_un_rr 100 100 19.488 5.75e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m k gflops resid result
blis_ssyrk_ln_cc 100 100 26.978 5.87e-03 FAILURE
blis_ssyrk_un_cc 100 100 30.095 7.23e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m k gflops resid result
blis_dsyrk_ln_cc 100 100 17.679 5.61e-03 FAILURE
blis_dsyrk_un_cc 100 100 19.533 6.10e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m k gflops resid result
blis_ssyr2k_lnn_rrr 100 100 30.041 4.06e-03 FAILURE
blis_ssyr2k_unn_rrr 100 100 29.014 3.98e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m k gflops resid result
blis_dsyr2k_lnn_rrr 100 100 20.131 3.68e-03 FAILURE
blis_dsyr2k_unn_rrr 100 100 19.409 4.03e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m k gflops resid result
blis_ssyr2k_lnn_ccc 100 100 28.311 4.00e-03 FAILURE
blis_ssyr2k_unn_ccc 100 100 26.463 3.87e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m k gflops resid result
blis_dsyr2k_lnn_ccc 100 100 20.009 5.91e-03 FAILURE
blis_dsyr2k_unn_ccc 100 100 19.136 4.59e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_strmm_llnn_rr 100 100 22.065 1.91e-03 FAILURE
blis_strmm_llnu_rr 100 100 28.846 1.39e-03 FAILURE
blis_strmm_lunn_rr 100 100 22.762 1.93e-03 FAILURE
blis_strmm_lunu_rr 100 100 28.680 1.73e-03 FAILURE
blis_strmm_rlnn_rr 100 100 26.187 1.92e-03 FAILURE
blis_strmm_rlnu_rr 100 100 30.305 2.26e-03 FAILURE
blis_strmm_runn_rr 100 100 24.478 1.69e-03 FAILURE
blis_strmm_runu_rr 100 100 27.479 1.46e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_dtrmm_llnn_rr 100 100 16.638 1.27e-03 FAILURE
blis_dtrmm_llnu_rr 100 100 18.322 1.46e-03 FAILURE
blis_dtrmm_lunn_rr 100 100 16.263 1.83e-03 FAILURE
blis_dtrmm_lunu_rr 100 100 18.009 1.32e-03 FAILURE
blis_dtrmm_rlnn_rr 100 100 17.064 1.83e-03 FAILURE
blis_dtrmm_rlnu_rr 100 100 18.600 1.85e-03 FAILURE
blis_dtrmm_runn_rr 100 100 14.241 1.69e-03 FAILURE
blis_dtrmm_runu_rr 100 100 15.083 1.42e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_strsm_llnn_rr 100 100 22.600 1.21e-03 FAILURE
blis_strsm_llnu_rr 100 100 28.844 3.79e-03 FAILURE
blis_strsm_lunn_rr 100 100 22.806 9.46e-04 FAILURE
blis_strsm_lunu_rr 100 100 28.039 3.32e-03 FAILURE
blis_strsm_rlnn_rr 100 100 27.456 1.15e-03 FAILURE
blis_strsm_rlnu_rr 100 100 27.592 4.01e-03 FAILURE
blis_strsm_runn_rr 100 100 28.372 9.41e-04 FAILURE
blis_strsm_runu_rr 100 100 28.738 5.51e-03 FAILURE
% blis_<dt><op>_<params>_<stor> m n gflops resid result
blis_dtrsm_llnn_rr 100 100 17.346 1.02e-03 FAILURE
blis_dtrsm_llnu_rr 100 100 18.108 3.53e-03 FAILURE
blis_dtrsm_lunn_rr 100 100 15.789 1.31e-03 FAILURE
blis_dtrsm_lunu_rr 100 100 17.264 3.66e-03 FAILURE
blis_dtrsm_rlnn_rr 100 100 17.067 9.44e-04 FAILURE
blis_dtrsm_rlnu_rr 100 100 14.562 4.78e-03 FAILURE
blis_dtrsm_runn_rr 100 100 17.793 9.90e-04 FAILURE
blis_dtrsm_runu_rr 100 100 18.148 4.70e-03 FAILURE
Metadata
Metadata
Assignees
Labels
No labels
Activity
mrader1248 commentedon Mar 23, 2018
As gcc 7.3 is rather new, I also tested gcc 7.2 -- same problem.
devinamatthews commentedon Mar 23, 2018
Could you please tell us:
mrader1248 commentedon Mar 23, 2018
./configure --prefix=${INSTALL_DIR} --enable-shared --enable-static --enable-blas --enable-cblas CC=gcc-7.3 haswell, output belowdevinamatthews commentedon Mar 23, 2018
OK, I can replicate this and will look into it.
mrader1248 commentedon Mar 23, 2018
Thank you very much for your quick response!
devinamatthews commentedon Mar 23, 2018
No problem, thanks for the bug report. Look like a doozy!
fgvanzee commentedon Mar 23, 2018
My default assumption is that this is a compiler bug, given that so many compilers before/other than gcc 7.2/7.3 do not show any problems. (Granted, it's always possible we're using some sloppy syntax that is no longer forgiven in those versions.)
@devinamatthews Thanks for looking into this.
devinamatthews commentedon Mar 23, 2018
@mrader1248 we've found the problem and I believe @fgvanzee is working on a fix (unless you want me to do it).
fgvanzee commentedon Mar 23, 2018
@devinamatthews I'll commit a fix to the issue @mrader1248 pointed out. Thanks for your help identifying the root cause.
fgvanzee commentedon Mar 23, 2018
@mrader1248 Please try e2192a8 and let us know if the test failures persist or go away. Thanks for your feedback.
mrader1248 commentedon Mar 24, 2018
Thank you very much for your great help! Now everything works as expected. Looking forward to benchmarking.
Just for curiosity's sake: Why is the zen code needed for the haswell config?
Will this commit be included into the latest version?
fgvanzee commentedon Mar 24, 2018
@mrader1248 Glad we were able to fix your problem. (Devin tracked down the issue quickly.)
While the Haswell, Broadwell, Skylake, Kabylake, Coffeelake, and Zen microarchitectures are all distinct, they actually share roughly the same instruction set (at least for our purposes) and number of vector registers. The net effect of this is that we can recycle kernel code across those microarchitectures, even if each of those systems would ideally use different cache blocksizes (which can be encoded in the individual sub-configurations, e.g.
haswellvs.zen) .So, we could have labeled the kernels as
haswelland imported them into thezenconfiguration, and the result would have been the same.I'm not sure if I understand the question. If you are asking if this commit will be merged into
master, the answer is absolutely. (In fact, I just did.)If you are asking if this commit will make it into the next version number, the answer is the same (yes). Please note that when we bump the version number is somewhat arbitrary. But rest assured that there is nothing special about using a commit that has a fresh version number; it's just an arbitrary milestone. We are constantly making improvements, in part thanks to great feedback from contributors such as yourself, and so we strongly encourage our users, especially those engaged with developers, to use the head commit of the
master(ordev) branch, rather than the latest tagged version.mrader1248 commentedon Mar 25, 2018
Thank you for the clarification.
Yes, to be 'git-precise': I wanted to know whether this commit will be merged into
master, such that I have to work with a specific commit no longer.Thank you all again for your help!
fgvanzee commentedon Apr 4, 2018
@mrader1248 Would you care to reveal your real name so I can acknowledge your contribution in our upcoming 0.3.1 announcement? No worries if you'd rather stay anonymous.
mrader1248 commentedon Apr 5, 2018
It's Michael Rader, but in the end you did the job, I just found a bug.
fgvanzee commentedon Apr 5, 2018
@mrader1248 Thanks Michael. Nevertheless, we like to acknowledge such contributions. (The ability to fix a bug does us no good if we don't know the bug exists.)
Fixing the coverity issues with CID: 23269 and CID: 137049 (flame#180)
Merge commit 'db3134ed6d239a7962a2b7470d8c46611b9d17ef' into AOCL-5.2-RC