##// END OF EJS Templates
Merge pull request #1369 from minrk/EngineError...
Merge pull request #1369 from minrk/EngineError load header with engine id when engine dies in TaskScheduler This ensures that the metadata dict on the Client has the engine_uuid of the engine on which the task failed. Previously, this entry would remain empty. It is identical to code elsewhere (Hub, Client) for constructing the dummy reply when engines die.

File last commit:

r6035:3077781f
r6098:0291d619 merge
Show More
task1.ipynb
117 lines | 2.1 KiB | text/plain | TextLexer

Simple task farming example

In [3]:
from IPython.parallel import Client

A Client.load_balanced_view is used to get the object used for working with load balanced tasks.

In [4]:
rc = Client()
v = rc.load_balanced_view()

Set the variable d on all engines:

In [5]:
rc[:]['d'] = 30

Define a function that will be our task:

In [6]:
def task(a):
    return a, 10*d, a*10*d

Run the task once:

In [7]:
ar = v.apply(task, 5)

Print the results:

In [8]:
print "a, b, c: ", ar.get()
a, b, c:  [5, 300, 1500]