OK, I'm on the cusp of success now. I updated to the most recent code and get to a sequence of errors surround math functions. This must be due to linking the MKL libs. I tried various combinations (including those recommended by the MKL docs) but could not get around the errors. Looks like a missing header file(?) In any event, I hope this can be fixed by getting the right into in site.cfg. Currently I have: [mkl] include_dirs = C:\Program Files (x86)\Intel\Compiler\11.0\061\cpp\mkl\include library_dirs = C:\Program Files (x86)\Intel\Compiler\11.0\061\cpp\mkl\em64t\lib mkl_libs = mkl_em64t, mkl_dll, mkl_core # various combinations of other tried... lapack_libs = mkl_lapack Below is the relevant portion of the output. During config the libraries check in as "FOUND" so I'm pretty sure the paths above are good. Thanks, ~Mike C. --------------------------- [edited] numpy\core\src\umath_funcs.inc.src(557): warning #266: function "sinhl" declared implicitly shr = sinhl(xr); ^ numpy\core\src\umath_funcs.inc.src(558): warning #266: function "coshl" declared implicitly chr = coshl(xr); ^ numpy\core\src\umath_loops.inc.src(913): warning #266: function "floorl" declare d implicitly *((longdouble *)op1) = floorl(in1/in2); ^ numpy\core\src\umath_loops.inc.src(923): warning #266: function "fmodl" declared implicitly const longdouble res = fmodl(in1,in2); ^ numpy\core\src\umath_loops.inc.src(1003): warning #266: function "modfl" declare d implicitly *((longdouble *)op1) = modfl(in1, (longdouble *)op2); ^ numpy\core\src\umath_loops.inc.src(1207): warning #266: function "fabsf" declare d implicitly if (fabsf(in1i) <= fabsf(in1r)) { ^ numpy\core\src\umath_loops.inc.src(1110): warning #266: function "floorl" declar ed implicitly ((longdouble *)op1)[0] = floorl((in1r*in2r + in1i*in2i)/d); ^ numpy\core\src\umath_loops.inc.src(1207): warning #266: function "fabsl" declare d implicitly if (fabsl(in1i) <= fabsl(in1r)) { ^ numpy\core\src\umath_loops.inc.src(1246): warning #266: function "sqrtl" declare d implicitly *((longdouble *)op1) = sqrtl(in1r*in1r + in1i*in1i); ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(244): error : identifier "acosl" is undefined arccos_data[2] = (void *) acosl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(253): error : identifier "acoshf" is undefined arccosh_data[0] = (void *) acoshf; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(257): error : identifier "acoshl" is undefined arccosh_data[2] = (void *) acoshl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(270): error : identifier "asinl" is undefined arcsin_data[2] = (void *) asinl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(279): error : identifier "asinhf" is undefined arcsinh_data[0] = (void *) asinhf; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(283): error : identifier "asinhl" is undefined arcsinh_data[2] = (void *) asinhl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(296): error : identifier "atanl" is undefined arctan_data[2] = (void *) atanl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(309): error : identifier "atan2l" is undefined arctan2_data[2] = (void *) atan2l; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(312): error : identifier "atanhf" is undefined arctanh_data[0] = (void *) atanhf; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(316): error : identifier "atanhl" is undefined arctanh_data[2] = (void *) atanhl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(335): error : identifier "ceill" is undefined ceil_data[2] = (void *) ceill; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(343): error : identifier "cosl" is undefined cos_data[2] = (void *) cosl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(356): error : identifier "coshl" is undefined cosh_data[2] = (void *) coshl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(385): error : identifier "expl" is undefined exp_data[2] = (void *) expl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(394): error : identifier "exp2f" is undefined exp2_data[0] = (void *) exp2f; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(396): error : identifier "exp2" is undefined exp2_data[1] = (void *) exp2; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(398): error : identifier "exp2l" is undefined exp2_data[2] = (void *) exp2l; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(401): error : identifier "expm1f" is undefined expm1_data[0] = (void *) expm1f; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(405): error : identifier "expm1l" is undefined expm1_data[2] = (void *) expm1l; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(414): error : identifier "fabsf" is undefined fabs_data[0] = (void *) fabsf; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(418): error : identifier "fabsl" is undefined fabs_data[2] = (void *) fabsl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(425): error : identifier "floorl" is undefined floor_data[2] = (void *) floorl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(438): error : identifier "fmodl" is undefined fmod_data[12] = (void *) fmodl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(441): error : identifier "hypotf" is undefined hypot_data[0] = (void *) hypotf; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(445): error : identifier "hypotl" is undefined hypot_data[2] = (void *) hypotl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(456): error : identifier "logl" is undefined log_data[2] = (void *) logl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(469): error : identifier "log10l" is undefined log10_data[2] = (void *) log10l; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(478): error : identifier "log1pf" is undefined log1p_data[0] = (void *) log1pf; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(482): error : identifier "log1pl" is undefined log1p_data[2] = (void *) log1pl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(491): error : identifier "log2f" is undefined log2_data[0] = (void *) log2f; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(493): error : identifier "log2" is undefined log2_data[1] = (void *) log2; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(495): error : identifier "log2l" is undefined log2_data[2] = (void *) log2l; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(534): error : identifier "powl" is undefined power_data[12] = (void *) powl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(564): error : identifier "rintf" is undefined rint_data[0] = (void *) rintf; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(566): error : identifier "rint" is undefined rint_data[1] = (void *) rint; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(568): error : identifier "rintl" is undefined rint_data[2] = (void *) rintl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(581): error : identifier "sinl" is undefined sin_data[2] = (void *) sinl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(594): error : identifier "sinhl" is undefined sinh_data[2] = (void *) sinhl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(607): error : identifier "sqrtl" is undefined sqrt_data[2] = (void *) sqrtl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(624): error : identifier "tanl" is undefined tan_data[2] = (void *) tanl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(637): error : identifier "tanhl" is undefined tanh_data[2] = (void *) tanhl; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(648): error : identifier "truncf" is undefined trunc_data[0] = (void *) truncf; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(650): error : identifier "trunc" is undefined trunc_data[1] = (void *) trunc; ^ build\src.win-amd64-2.6\numpy\core\include/numpy/__umath_generated.c(652): error : identifier "truncl" is undefined trunc_data[2] = (void *) truncl; ^ numpy\core\include\numpy/ufuncobject.h(372): warning #177: function "generate_ov erflow_error" was declared but never referenced static void generate_overflow_error(void) { ^ numpy\core\src\umath_funcs.inc.src(70): warning #177: function "exp2_1mf" was de clared but never referenced exp2_1mf(float x) ^ numpy\core\src\umath_funcs.inc.src(70): warning #177: function "exp2_1m" was dec lared but never referenced exp2_1m(double x) ^ numpy\core\src\umath_funcs.inc.src(70): warning #177: function "exp2_1ml" was de clared but never referenced exp2_1ml(longdouble x) ^ compilation aborted for build\src.win-amd64-2.6\numpy\core\src\umathmodule.c (co de 2) error: Command "C:\Program Files (x86)\Intel\Compiler\11.0\061\cpp\Bin\intel64\i cl.exe /c /nologo /Ox /MD /W3 /GS- /DNDEBUG -Ibuild\src.win-amd64-2.6\numpy\core \src -Inumpy\core\include -Ibuild\src.win-amd64-2.6\numpy\core\include/numpy -In umpy\core\src -Inumpy\core\include -IC:\Python26\include -IC:\Python26\PC /Tcbui ld\src.win-amd64-2.6\numpy\core\src\umathmodule.c /Fobuild\temp.win-amd64-2.6\Re lease\build\src.win-amd64-2.6\numpy\core\src\umathmodule.obj" failed with exit s tatus 2