03:31 <Corbin> cfbolz: I'm still working on my problem, but your suggestions have sped up my test by maybe 2x or so. I need to refactor my user-level code to allocate fewer lists now. Thanks again.

06:24 <cfbolz> Cool

06:25 <mattip> I saw a tweet on the pypy feed about clair3 and pypy, so dug around a bit and found this

06:25 <mattip> https://github.com/HKU-BAL/Clair#clair-uses-pypy-for-speedup

06:26 <mattip> "Without a change to the code, using PyPy python interpreter on some tensorflow independent modules such as ExtractVariantCandidates and CreateTensor gives a 5-10 times speed up"

06:29 <mattip> there was also a tweet about antocuni joining Anaconda

06:29 <mattip> https://www.businesswire.com/news/home/20220404005889/en/Anaconda-Welcomes-Python-Thought-Leaders-to-Advance-Technical-Frontier

06:29 <LarstiQ> also notes "Tensorflow is optimized using Cython thus not compatible with pypy3"

07:28 <cfbolz> https://github.com/HKU-BAL/Clair3

07:28 <cfbolz> The successor project uses it too

07:34 <mattip> maybe we could get them to write a guest blog post. The documentation of Clair3 is quite domain specific