New executor for Fujitsu Technical Computing Suite (TCS) #5928
Conversation
✅ Deploy Preview for nextflow-docs-staging canceled.
|
There was a problem hiding this comment.
Thanks a lot for this contribution. Please consider adding some unit tests. You can reference to other executors for some examples. Also please sign your commits as example in the contributing docs
| I referred to the other executors groovy implementations. | ||
| Resource requests and other job characteristics can be controlled via the following process directives: | ||
| - time | ||
| - clusterOptions | ||
| queue (resource group), cpu, node, and other options should be indicated by clusterOptions. | ||
| This is because they depend on target systems (required options are not the same) and can be controlled by "-L" options in arguments of pjsub command. | ||
|
|
||
| example of nextflow.config on Genkai (Kyushu University) | ||
| process { | ||
| executor = 'tcs' | ||
| time = '00:30:00' | ||
| clusterOptions = '-L rscgrp=a-batch -L vnode-core=4' | ||
| } | ||
|
|
||
| example of nextflow.config on Flow (Nagoya University) | ||
| process { | ||
| executor = 'tcs' | ||
| time = '00:30:00' | ||
| clusterOptions = '-L rscunit=cx -L rscgrp=cx-share -L gpu=1' | ||
| } |
There was a problem hiding this comment.
Let's move this to the docs. Can you create a new section in docs/executor.md? You can use existing sections as a template
There was a problem hiding this comment.
Let's move this to the docs. Can you create a new section in
docs/executor.md? You can use existing sections as a template
Thank you, I'm updating the documents.
Thanks for the advice. I'll try to add unit tests and add sign. |
Signed-off-by: Satoshi Ohshima <[email protected]>>
This is a new Executor for Fujitsu Technical Computing Suite (TCS).
Some supercomputers build by Fujitsu are operated (mainly in Japan).
Most of they use Technical Computing Suite (TCS) job management system, they are not supported by current Nextflow.
Therefore, I wrote new Executor citing existing executors.
We tested simple jobs and confirmed the correct behaviors on on Fugaku (Riken), Flow (Naogya Univ.), and Genkai (Kyushu Univ.) supercomputers.
Note that I am not familiar with all the features of Nextflow, so there may be some lack of implementation.