uncomplicate

2017-09-12T14:44:45.000600Z

Hi, When I do a dot product of double vectors I get the expected answer: (n is neanderthal core and nn is neaderthal native). (n/dot (nn/dv 1 2 3) (nn/dv 1 2 3)) => 14.0 but the dot of float vectors gives me 0: (n/dot (nn/fv 1 2 3) (nn/fv 1 2 3)) => 0.0 Am I missing something? Thanks.

2017-09-12T15:25:57.000039Z

@juliobarros This is a known issue of MKL on OSX only. https://github.com/uncomplicate/neanderthal/issues/30

2017-09-12T15:26:32.000432Z

I haven't had time to investigate this further yet.

2017-09-12T15:29:27.000252Z

Ok. Thanks. Also, FYI I stumbled around trying to install MKL and figured out setting DYLD_LIBRARY_PATH on MAC OS X doesn’t work as expected anymore because because of system integrity protection (SIP) http://osxdaily.com/2015/10/05/disable-rootless-system-integrity-protection-mac-os-x/ I instead created links to an exisiting directory. ie. ln -s /opt/intel/lib* /usr/local/lib (and /opt/intel/mkl/lib) and that seems to work. There is a mention of this in the docs (thanks!) but maybe a clearer sentence would alter most people that this is an option. Thanks again.

2017-09-12T15:30:57.000007Z

Maybe the best would be if you open an issue on github with this, since there is a slim chance that anyone would find this information on Slack.