[Rdkit-discuss] Parallel conformation generation

Hi, I’ve been using the parallel conformation generation code ( https://buildmedia.readthedocs.org/media/pdf/rdkit/latest/rdkit.pdf page 95) in a number of projects including the open source antibiotics.

Re: [Rdkit-discuss] parallel conformation generation

This is great Andrew (especially the subsequent explanation)! Many Thanks. Considering that this is a task lots of people will want to do - is this code CONTRIB dir material? (perhaps max_workers should be a fourth command line argument defaulting to 1) A few months (years?) back Greg

Re: [Rdkit-discuss] parallel conformation generation

Hi Andrew, Thanks for this. I didn't know about the futures and progressbar modules. You wrote: --- *I have to use the zip because map(f, iterable, [chunksize=None]) only takes a single iterable. This also means I need to change the generateconformations function so that it takes a single

Re: [Rdkit-discuss] parallel conformation generation

Hi again, Greg asked why I used the concurrent.futures module rather than the multiprocessing module which is standard with Python 2.6. There are a few differences in the API which makes the futures module more interesting. First off, here's how you could write the same process pool part using

[Rdkit-discuss] parallel conformation generation

JP mentioned that he would like to use multiple CPUs for his conformation generation code, which he posted during his talk today. Here's a straight-forward parallelization of his code. On my laptop, with 4 processes allocated, I get a 2.5x speedup. I have an Intel Core i7 with 1 CPU and 4 cores.