Skip to content

Commit 92ab47f

Browse files
authored
Update quickstart example config (#3695)
# Description Minor modifications to the Quickstart documentation. Fixes a typo in `retries`, adds the relevant import statements to the Polaris config, and changes the config to use the debug queue rather than prod. ## Type of change Choose which options apply, and delete the ones which do not apply. - Update to human readable text: Documentation/error messages/comments
1 parent 45d60a8 commit 92ab47f

File tree

1 file changed

+9
-4
lines changed

1 file changed

+9
-4
lines changed

docs/quickstart.rst

Lines changed: 9 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -176,12 +176,17 @@ This script runs on a system that must stay on-line until all of your tasks comp
176176
much computing power, such as the login node for a supercomputer.
177177

178178
The :class:`~parsl.config.Config` object holds definitions of Executors and the Providers and Launchers they rely on.
179-
An example which launches 512 workers on 128 nodes of the Polaris supercomputer looks like
179+
An example which launches 4 workers on 1 node of the Polaris supercomputer looks like
180180

181181
.. code-block:: python
182182
183+
from parsl import Config
184+
from parsl.executors import HighThroughputExecutor
185+
from parsl.providers import PBSProProvider
186+
from parsl.launchers import MpiExecLauncher
187+
183188
config = Config(
184-
retires=1, # Restart task if they fail once
189+
retries=1, # Restart task if they fail once
185190
executors=[
186191
HighThroughputExecutor(
187192
available_accelerators=4, # Maps one worker per GPU
@@ -191,13 +196,13 @@ An example which launches 512 workers on 128 nodes of the Polaris supercomputer
191196
account="example",
192197
worker_init="module load conda; conda activate parsl",
193198
walltime="1:00:00",
194-
queue="prod",
199+
queue="debug",
195200
scheduler_options="#PBS -l filesystems=home:eagle", # Change if data on other filesystem
196201
launcher=MpiExecLauncher(
197202
bind_cmd="--cpu-bind", overrides="--depth=64 --ppn 1"
198203
), # Ensures 1 manger per node and allows it to divide work to all 64 cores
199204
select_options="ngpus=4",
200-
nodes_per_block=128,
205+
nodes_per_block=1,
201206
cpus_per_node=64,
202207
),
203208
),

0 commit comments

Comments
 (0)