Bits for the cross-experiment (ATLAS, CMS, LHCb) study on heterogeneous workflows

This is needed for this study.

• extend the Avalanche Scheduler test suite with a CMS reconstruction scenario assembled by Christopher Jones (CMS);
• tweak the Gaudi execution flow simulator to be able to run the CMS scenario;
• update the LHCb Brunel scenario (new precedence rules and timing provided by @clemenci);
• add script which determines the critical path by a precedence trace;
• ... expecting more changes to come.

CC @leggett @clemenci

marked as a Work In Progress

changed the description

added task scheduling tests labels

@ishapova It seems that the description link either is dead or points to something which is not public yet. Maybe you meant to point here?

The link actually works for me, but the abstract section of the page is probably accessible to authors only. Thanks for spotting this, I will change the link to point to the contribution section shortly.

changed the description

added 43 commits

• a96038d6...8a169955 - 36 commits from branch gaudi:master
• 0666bc5c - add first version of CMS reconstruction scenario
• 40915408 - make the scheduler fail @initialize if it doesn't find any algorithms to run in a workflow
• 4011f80f - fix the CF topology and DF object ids in the CMS reco scenario rules (it now runs)
• 4f8bdf84 - use the full name (don't strip the type) of Algorithm nodes when building...
• 55e916d3 - Revert "use the full name (don't strip the type) of Algorithm nodes when...
• c9a8239d - convert compound algorithm names to simple ones in the CMS reco scenario
• 5f3cd87d - enable efficient scheduling of offloading algorithms in the CMS scenario

Compare with previous version

added 1 commit

• 88640399 - fix formatting

Compare with previous version

added 1 commit

• f30ab368 - write time info for graphml in microseconds

Compare with previous version

added 1 commit

• f4f599e9 - added new ATLAS q431 datafile

Compare with previous version

added 1 commit

• 7a7e86d3 - spelling

Compare with previous version

added 1 commit

• 0ad18394 - update LHCb reco scenario and make more flexible graph node type deduction in...

Compare with previous version

changed the description

added 36 commits

• 0ad18394...aeb156f0 - 24 commits from branch gaudi:master
• cddc5917 - add first version of CMS reconstruction scenario
• 50ff96ec - make the scheduler fail @initialize if it doesn't find any algorithms to run in a workflow
• d935282f - fix the CF topology and DF object ids in the CMS reco scenario rules (it now runs)
• df68810c - use the full name (don't strip the type) of Algorithm nodes when building...
• 5da6aa8d - Revert "use the full name (don't strip the type) of Algorithm nodes when...
• 432ecdbc - convert compound algorithm names to simple ones in the CMS reco scenario
• 2d8e239c - enable efficient scheduling of offloading algorithms in the CMS scenario
• 9acc46ec - fix formatting
• 5ae5fb97 - write time info for graphml in microseconds
• 54d8a27d - added new ATLAS q431 datafile
• e7ca06ff - spelling
• a7d511ce - update LHCb reco scenario and make more flexible graph node type deduction in...

Compare with previous version

added 1 commit

• c4e39e11 - add script to determine critical path by a precedence trace

Compare with previous version

changed the description

added 1 commit

• a31ce38a - update dataDeps to account for more objects in AOD file

Compare with previous version

unmarked as a Work In Progress

I'm about to close and reopen this MR to get around the "could not retrieve the pipeline status" bug.

closed