Concurrent execution of independent blocks


#1

According to this comment, Datalore maintains a computation graph of dependent cells. Is it possible to parallelize the evaluation of independent blocks? For example, suppose I have some long-running computation:

for i in range(99999999):
    pass

And this is followed by a block that does not depend on the results of the previous computation, eg.:

print("Hello")

Then I need to wait for the first block to complete before evaluating the second block. I noticed that if the order is switched, then it is possible to evaluate two blocks concurrently. I would like to continue working on the notepad or launch a new block of code while the previous block is still running. Thanks!


#2

Right now, it’s not so easy for us, because our dependency graph based not only on python code, but on python environment analysis after cell execution also. Nevertheless, we are looking forward to find a way to implement this feature, and I’m sure we will find it sooner or later. Will keep you updated! Many thanks for your feedback!