Thanks for this great "meta"-kernel!
One thing though, when executing two consecutive cells which use a different kernel, the second cell won't wait for the first one to finish, if the kernel of the second cell is not busy. In other words, the execution state is per kernel.
While that may be useful in some situations, would it be possible to optionally block the execution until all kernels are idle?
Thanks for this great "meta"-kernel! One thing though, when executing two consecutive cells which use a different kernel, the second cell won't wait for the first one to finish, if the kernel of the second cell is not busy. In other words, the execution state is per kernel. While that may be useful in some situations, would it be possible to optionally block the execution until all kernels are idle?