|
|
@@ -72,6 +72,18 @@ base_kops = [ |
|
|
'x': {'exts': {'_k': {'dir': 'x86_64', 'kernel': 'zdot.S'}}}, |
|
|
'x': {'exts': {'_k': {'dir': 'x86_64', 'kernel': 'zdot.S'}}}, |
|
|
}, |
|
|
}, |
|
|
}, |
|
|
}, |
|
|
|
|
|
{ 'base': '?dotc', |
|
|
|
|
|
'modes': { |
|
|
|
|
|
'c': {'exts': {'_k': {'dir': 'x86_64', 'kernel': 'cdot.c'}}}, |
|
|
|
|
|
'z': {'exts': {'_k': {'dir': 'x86_64', 'kernel': 'zdot.c'}}}, |
|
|
|
|
|
}, |
|
|
|
|
|
}, |
|
|
|
|
|
{ 'base': '?dotu', |
|
|
|
|
|
'modes': { |
|
|
|
|
|
'c': {'exts': {'_k': {'dir': 'x86_64', 'kernel': 'cdot.c'}}}, |
|
|
|
|
|
'z': {'exts': {'_k': {'dir': 'x86_64', 'kernel': 'zdot.c'}}}, |
|
|
|
|
|
}, |
|
|
|
|
|
}, |
|
|
# TODO(rg): Add dsdotkernel conditionals |
|
|
# TODO(rg): Add dsdotkernel conditionals |
|
|
# xDOTU xDOTC xxDOT aren't present |
|
|
# xDOTU xDOTC xxDOT aren't present |
|
|
{ 'base': '?nrm2', |
|
|
{ 'base': '?nrm2', |
|
|
@@ -501,6 +513,8 @@ base_kops = [ |
|
|
'modes': { |
|
|
'modes': { |
|
|
's': {'exts': { |
|
|
's': {'exts': { |
|
|
'_beta': {'dir': 'generic', 'kernel': 'gemm_beta.c'}, |
|
|
'_beta': {'dir': 'generic', 'kernel': 'gemm_beta.c'}, |
|
|
|
|
|
'_small_matrix_permit': {'dir': 'generic', |
|
|
|
|
|
'kernel': 'gemm_small_matrix_permit.c'}, |
|
|
# TODO(rg): the _NUM prefixes are arch dependent |
|
|
# TODO(rg): the _NUM prefixes are arch dependent |
|
|
'_incopy': {'dir': 'generic', 'kernel': 'gemm_ncopy_8.c'}, |
|
|
'_incopy': {'dir': 'generic', 'kernel': 'gemm_ncopy_8.c'}, |
|
|
'_itcopy': {'dir': 'generic', 'kernel': 'gemm_tcopy_8.c'}, |
|
|
'_itcopy': {'dir': 'generic', 'kernel': 'gemm_tcopy_8.c'}, |
|
|
@@ -512,6 +526,8 @@ base_kops = [ |
|
|
}}, |
|
|
}}, |
|
|
'd': {'exts': { |
|
|
'd': {'exts': { |
|
|
'_beta': {'dir': 'generic', 'kernel': 'gemm_beta.c'}, |
|
|
'_beta': {'dir': 'generic', 'kernel': 'gemm_beta.c'}, |
|
|
|
|
|
'_small_matrix_permit': {'dir': 'generic', |
|
|
|
|
|
'kernel': 'gemm_small_matrix_permit.c'}, |
|
|
'_incopy': {'dir': 'generic', 'kernel': 'gemm_ncopy_8.c'}, |
|
|
'_incopy': {'dir': 'generic', 'kernel': 'gemm_ncopy_8.c'}, |
|
|
'_itcopy': {'dir': 'generic', 'kernel': 'gemm_tcopy_8.c'}, |
|
|
'_itcopy': {'dir': 'generic', 'kernel': 'gemm_tcopy_8.c'}, |
|
|
'_oncopy': {'dir': 'generic', 'kernel': 'gemm_ncopy_2.c'}, |
|
|
'_oncopy': {'dir': 'generic', 'kernel': 'gemm_ncopy_2.c'}, |
|
|
@@ -519,6 +535,8 @@ base_kops = [ |
|
|
}}, |
|
|
}}, |
|
|
'c': {'exts': { |
|
|
'c': {'exts': { |
|
|
'_beta': {'dir': 'generic', 'kernel': 'zgemm_beta.c'}, |
|
|
'_beta': {'dir': 'generic', 'kernel': 'zgemm_beta.c'}, |
|
|
|
|
|
'_small_matrix_permit': {'dir': 'generic', |
|
|
|
|
|
'kernel': 'zgemm_small_matrix_permit.c'}, |
|
|
'_incopy': {'dir': 'generic', 'kernel': 'zgemm_ncopy_8.c'}, |
|
|
'_incopy': {'dir': 'generic', 'kernel': 'zgemm_ncopy_8.c'}, |
|
|
'_itcopy': {'dir': 'generic', 'kernel': 'zgemm_tcopy_8.c'}, |
|
|
'_itcopy': {'dir': 'generic', 'kernel': 'zgemm_tcopy_8.c'}, |
|
|
'_oncopy': {'dir': 'generic', 'kernel': 'zgemm_ncopy_2.c'}, |
|
|
'_oncopy': {'dir': 'generic', 'kernel': 'zgemm_ncopy_2.c'}, |
|
|
@@ -526,6 +544,8 @@ base_kops = [ |
|
|
}}, |
|
|
}}, |
|
|
'z': {'exts': { |
|
|
'z': {'exts': { |
|
|
'_beta': {'dir': 'generic', 'kernel': 'zgemm_beta.c'}, |
|
|
'_beta': {'dir': 'generic', 'kernel': 'zgemm_beta.c'}, |
|
|
|
|
|
'_small_matrix_permit': {'dir': 'generic', |
|
|
|
|
|
'kernel': 'zgemm_small_matrix_permit.c'}, |
|
|
'_incopy': {'dir': 'generic', 'kernel': 'zgemm_ncopy_8.c'}, |
|
|
'_incopy': {'dir': 'generic', 'kernel': 'zgemm_ncopy_8.c'}, |
|
|
'_itcopy': {'dir': 'generic', 'kernel': 'zgemm_tcopy_8.c'}, |
|
|
'_itcopy': {'dir': 'generic', 'kernel': 'zgemm_tcopy_8.c'}, |
|
|
'_oncopy': {'dir': 'generic', 'kernel': 'zgemm_ncopy_2.c'}, |
|
|
'_oncopy': {'dir': 'generic', 'kernel': 'zgemm_ncopy_2.c'}, |
|
|
@@ -1003,8 +1023,8 @@ endforeach |
|
|
|
|
|
|
|
|
_kern_libs = [] |
|
|
_kern_libs = [] |
|
|
foreach conf: kernel_confs |
|
|
foreach conf: kernel_confs |
|
|
message(conf['name']) |
|
|
|
|
|
message(conf) |
|
|
|
|
|
|
|
|
# message(conf['name']) |
|
|
|
|
|
# message(conf) |
|
|
_kern_libs += static_library( |
|
|
_kern_libs += static_library( |
|
|
conf['name'], |
|
|
conf['name'], |
|
|
conf['src'], |
|
|
conf['src'], |
|
|
|