pylops
pylops copied to clipboard
PyLops
Added Intel `mkl_fft`
Thank you for this PR.
So far Mac tests are failing, do you understand why?
Reason
According to this, the,mkl is not able to locate the Volumes/localdisk/tools/tc/agent1/work/636dcb48f7ee4116/base/conda-bld/numpy_and_dev_1651366109538/_build_env/lib/libmkl_rt.2.dylib' (no such file)
Alright can you try to investigate a bit deeper? Either it failed to install or maybe there is no bindings for mac (probably if this is the case, the mkl-fft library should say this)? Worst case scenario we can check whether it’s Mac or Linux and not run the test for Mac - we did this before, I can point you to the code if needed..
More importantly also readthedocs fails, this one we need to be able to fix before we proceed.
I’ll get back with a more detailed review soon, and @cako will probably do the same as he created the issue :)
Alright can you try to investigate a bit deeper? Either it failed to install or maybe there is no bindings for mac (probably if this is the case, the mkl-fft library should say this)? Worst case scenario we can check whether it’s Mac or Linux and not run the test for Mac - we did this before, I can point you to the code if needed..
More importantly also readthedocs fails, this one we need to be able to fix before we proceed.
I’ll get back with a more detailed review soon, and @cako will probably do the same as he created the issue :)
Yeah I am trying it, will let you know
@rohanbabbar04 what is the status of this PR? I see that rtd still fails, did you find out why?
I see also now an issue with Python 3.10 in our Github Action. Seems like mkl-fft is not shipped for python 3.10 yet, is that possible?
I was able to work it out for Mac mkl-fft is not supported for 3.10 The docs are working locally but getting seg faults here
I'll make a new PR explaining my changes, still need to think of the docs
Alright, as far as python 3.10 is concerned we could make two separate build processes with two separate requirements-dev for now...
For the doc, we definitely need to understand why before we can proceed.
Sure go ahead and make a clean PR, we can continue there :)
I'm a bit out of the loop on this PR, what is the current status? Should I review it, or another one will be submitted? If this one is not done, I would suggest changing it to draft. Thanks!
@rohanbabbar04 what is the status of this PR. I plan to make a new release in one week or so, shall we aim to get this one in or not?
Let's not include this one in the next release, currently working on my GSOC proposal I'll finish this one after that...