Backport of commit 145e0240 to HltAsyncEventLoopMgr.

Backport commit 2b7217bc to HltAsyncEventLoopMgr.

Backport commit 34b5d520 to HltAsyncEventLoopMgr.

Simone Sottocornola authored 1 year ago and Marija Vranjes Milosavljevic committed 1 year ago

Added maxParallelismExtra setting to AvalancheSchedulerSvcCfg

Propagate the number of forks to `Gaudi::ConcurrencyFlags` so clients
can use this to detect if we are in MP-mode.

Relates to ATR-28997.

Stewart Martin-Haugh authored 1 year ago and John Kenneth Anders committed 1 year ago

Asynchronous I/O in HltEventLoopMgr: copy to new class

Oleg Kuprash authored 1 year ago and Tim Martin committed 1 year ago

ATR-28873: support more workflows with container jobs in Trig_reco_tf.py

Delete the legacy runHLT_standalone job options and related utilities.

Jennifer Rachel Curran authored 1 year ago and Vakhtang Tsulaia committed 1 year ago

Removal of functions to add differences in pre/pos trees as added to a later step and check on if an event status name is present for events that do not have an explicit 0 set

Jennifer Rachel Curran authored 1 year ago and Tadej Novak committed 1 year ago

Fix to read to end of string in line for accepted/rejected events so that for runs where accepted/rejected > 9, the events won't be cut off

Scott Snyder authored 1 year ago and Adam Edward Barton committed 1 year ago

TrigPSC: Update for python 3.11.

Update the passing of command-line arguements to python;
PySys_SetArgv is deprecated as of 3.11.

Jennifer Rachel Curran authored 1 year ago and Adam Edward Barton committed 1 year ago

Debug stream recovery, Fix to trig reco

Jennifer Rachel Curran authored 1 year ago and Adam Edward Barton committed 1 year ago

Addition of new function to add a Tree and histograms that show the difference in the Pre/Pos histograms in dbgAnalysis, modification of a function to add in the HLT_accepted_events histogram along with HLT_rejected_events as well as an additional python file that creates an empty file with the BSFile structure and changes to trigRecoExe to accomodate these additional functions and to rename the empty BSFile to RAW.pool.root if no events are selected for the chosen streams to prevent failures at T0 when events are actually just rejected by the HLT and/or stream selection when running the TriggerTransform for debug recovery.

Mark Stockton authored 1 year ago and Johannes Elmsheuser committed 1 year ago

CA updates for TriggerTransform

Scott Snyder authored 1 year ago and Johannes Elmsheuser committed 1 year ago

TrigPSC: Fix cppcheck warnings.

Pass objects by const reference, not by value.
Prefer to initialize class members in an initialization list.
Blank macro arguments can confuse cppcheck.

Stewart Martin-Haugh authored 1 year ago and Frank Winklmeier committed 1 year ago

Set scheduler defaults to match runHLT in PSC

ALFA does not need to be supported anymore (ATR-28444).

Clean up memory management.

Stewart Martin-Haugh authored 1 year ago and Frank Winklmeier committed 1 year ago

Migrate TimeBurner, MistimeMon to CA (ATR-28359)

Andrii Verbytskyi authored 1 year ago and Johannes Elmsheuser committed 1 year ago

Remove the linkage to boost libraries in case only header-only components are used

The HLTMPPU implementation of MonSvc is not thread-safe. To prevent any
possible issues, add an extra lock also to the `deReg(const std::string&)`
method following what is being done for all other methods.

Maybe related to ATR-28222.

Fix ALFA LB histogram (ATR-27909).

Fix two issues that prevented the LB histograms to be filled correctly:
- Two `Monitored::Scalar` with the name "y" were used. Rename one of
  them to "ay" and adjust `defineHistogram` calls.
- Remove the `_LB` suffixes from the axis variables to match the C++
  code.

Closes ATR-27909.

Andrii Verbytskyi authored 1 year ago and Adam Edward Barton committed 1 year ago

Drop boost::lexical_cast in the places wher std::to_string will work

- `Trigger.DecodeHLT` -> `Trigger.decodeHLT`
- `Trigger.endOfEventProcessing.Enabled` -> `Trigger.enableEndOfEventProcessing`
- `Trigger.HLTSeeding.forceEnableAllChains` -> `Trigger.forceEnableAllChains`

Tadej Novak authored 1 year ago and Frank Winklmeier committed 1 year ago

Trigger: implement rewriteLVL1 feature in CA

See merge request atlas/athena!65052

(cherry picked from commit 018455eb)

e45a3c4f TrigPSC: set writeBS=True by default and support doL1Sim
e008ed83 TrigServices+TriggerJobOpts: implement rewriteLVL1 feature in CA
58d37c70 TriggerJobOpts: deprecate rewriteLVL1 modifier

This reverts merge request !64860

Implement the equivalent of the `rewriteLVL1` Modifier for CA. Instead
of introducing yet another flag, we assume that when the LVL1 simulation
and BS-writing is enabled, we always want to write out the updated
LVL1 result.

Relates to ATR-26582.

Enable BS-writing by default and fully support the `doL1Sim` Modifier of
runHLT_standalone. The corresponding flag needs to be set early on to
avoid merge conflicts in TrigServicesCfg.

Also need to set default `EDMVersion` in order for the
auto-configuration to work correctly.

Joerg Stelzer authored 1 year ago and Michele Faucci Giannelli committed 1 year ago

ALFARobMonitor multi-LB hist publish by default