Open templary opened 2 years ago
Hello and welcome to the Cromwell repo. Three minutes is about what I would expect from personal experience, as a minimum time to run any task.
Consider that Life Sciences allocates, starts, and pulls Docker on a dedicated VM just to print "hello world". It will never look favorable for small tasks whose execution time is short compared to VM setup time.
Okay, thank you so much for the answer.
In this case, then, I would ask if using
docker-image-cache-manifest-file = "gs://xxxxx-xxxxx/xxxxx.json"
is it possible to achieve acceleration in Google Life Sciences or is it possible to use this cache method only for acceleration when running on the local backend?
I am thinking of such a solution, do you think it is in line with cromwell's good practices? -> distribute calculations according to whether I estimate they will be heavy and if so send them to google life sciences for calculation and if not calculate them on the local backend?
Hello,
I'm trying to connect cromwell with Gloud life science (v2beta) everything works as it should except for the problem that it takes almost 3 minutes to complete a simple hello world task. I am using the config file recommended by the documentation to run.
I would like to ask for advice on how to solve this problem. Thank you a lot.
config file:
WDL:
input
Gcloud log (edited):