Radim Řehůřek

> Does this mean that I have not been using BLAS even from the previous system? No, it means the BLAS lib was using default threading. I don't know what...

Instead, can you try running the `timeit` in the exact same way you run your doc2vec code?

And this (`--ntasks=1 --cpus-per-task=24 as arguments in SBATCH`) is how you got the Doc2Vec timings in your original post?

Thank you for the detailed reports. I'm not sure what the problem is – this will have to wait until someone has the time to reproduce & dig deeper.

Well that's exactly what we were checking above – whether your BLAS is parallelized. I'm not familiar with `slurm`, but my understanding of the experiments you reported is that your...

Sure, makes sense to me. I don't know if this breaks anything deeper down – but hopefully our tests would catch that.

@mpenkov could you run the test suite? And if all passes, let's merge & release as part of 4.3.3.

Great! Can you test this out please & open a PR? Thanks.

Thanks for reporting! @mpenkov Is fuzzyTM really a hard dependency? If so that's terrible, definitely an omission / bug (or if intentional, done in very bad taste). Let's release a...

I tracked the change in `setup.py` down to https://github.com/RaRe-Technologies/gensim/pull/3398. @ERijck why do you think this was needed, why did you add that line?