[PyInstaller] Re: numpy using mkl libraries

Thanks everyone for the replies and suggestions. I was planning to add to the *binaries*, but followed what others did by setting directly in *data*. That worked. I didn't try adding numpy to hiddenimports though, because the rest of numpy works and is detected properly. RaC On Monday, July

Re: [PyInstaller] Re: numpy using mkl libraries

The other mkl dlls are manually specified in a hook file inside PyInstaller here . It looks like they just missed that one simply because libiomp5md doesn’t have mkl in its name. In which case it

Re: [PyInstaller] Re: numpy using mkl libraries

> *Question*: why the mkl libraries are not properly detected/used during >>> runtime even if they are included in the executable? >>> >> MKL does some run-time magic to detect the processor and the like, so it may be linking in an unusual way that PyInstaller can not detect. Adding it by hand

[PyInstaller] Re: numpy using mkl libraries

*UPDATE 2 :* In the meantime ... I added the dll to the data variable in specs (as stated here https://stackoverflow.com/questions/62625517/pyinstaller-exe-file-terminates-early-without-an-error-message and here

[PyInstaller] Re: numpy using mkl libraries

I can't answer the question I'm afraid. But did want to say thank you for asking it. I had the exact same error but didn't realise it. When I tried copying the required dll into the folder my compiled program is now working. So thanks, but yeah this seems to be two of us with this bug. On

[PyInstaller] Re: numpy using mkl libraries

*UPDATE:* I checked one by one the dlls I had in the folder where the executable was running and found that there was only one that was required to run the executable properly: Seems like this omp library is not included in the executable. Any reason why ? Thanks in advance! RaC On