Calculating Henry coefficients for multiple crystals and/or gases in a @distributed for loop, using julia -p X with X > 4 does not work. The result is an error: Use 5 cores or less for Henry coefficient calculations to match the number of blocks
So, it seems that for a single crystal you can only use 4 processes (inconsistent with the error message) and for a single script execution you can also only use 4 processes (as opposed to 4 procs per crystal). This can be worked around by using a shell script to drive the loop and fire off julia -p 4 jobs for each crystal, but should be looked into for better usability and efficiency.
Calculating Henry coefficients for multiple crystals and/or gases in a
@distributed for
loop, usingjulia -p X
with X > 4 does not work. The result is an error:Use 5 cores or less for Henry coefficient calculations to match the number of blocks
So, it seems that for a single crystal you can only use 4 processes (inconsistent with the error message) and for a single script execution you can also only use 4 processes (as opposed to 4 procs per crystal). This can be worked around by using a shell script to drive the loop and fire off
julia -p 4
jobs for each crystal, but should be looked into for better usability and efficiency.