- Apr 07, 2016
-
-
Charles Leggett authored
much more useful if the global static accessor to the thread local EventContext object returns a const ref instead of a copy of the object. See merge request !7
-
Charles Leggett authored
-
- Apr 01, 2016
-
-
Charles Leggett authored
See merge request !6
-
Charles Leggett authored
-
- Mar 25, 2016
-
-
Charles Leggett authored
-
Charles Leggett authored
This set of changes completes the work started by @leggett on thread initialization tools. Such tools are invoked by the ThreadPoolSvc on each worker thread when the pool is initialized and can be used to setup thread-local workspaces in cases where necessary (e.g. for running multi-threaded Geant4). These commits enable the thread-termination part of those tools for corresponding cleanup/finalization of thread-local memory. This is done primarily by adding a terminatePool method to IThreadPoolSvc which is called at the end of ForwardSchedulerSvc::activate. The ThreadPoolSvc underwent some slight refactoring to support the thread-termination tasks. Also included in this merge request are a number of whitespace and indentation fixes, as well as some added class documentation in the affected classes. See merge request !4
-
- Mar 24, 2016
-
-
Steven Andrew Farrell authored
Realized after testing that the IThreadPoolSvc::terminatePool call needs to be executed in the scheduler thread, just like the initPool call. Otherwise, the ThreadInitTools won't be terminated on all threads. This may be due to a subtle feature of TBB which would be good to understand, but this way the implementation probably makes more sense anyway. Added an additional error check in ForwardSchedulerSvc::finalize to catch any problems that may have occurred during the call to terminatePool.
-
Steven Andrew Farrell authored
-
Steven Andrew Farrell authored
-
Steven Andrew Farrell authored
-
Steven Andrew Farrell authored
-
Steven Andrew Farrell authored
-
Steven Andrew Farrell authored
Added the terminatePool functionality to the ThreadPoolSvc to be able to invoke the IThreadInitTool::terminateThread method analogously to how the initThread is called concurrently on each thread. Added terminatePool method to the IThreadPoolSvc. Refactored the ThreadPoolSvc so that the code to launch the tool-wrapping tasks is in a dedicated method, launchTasks, which can be called from both initPool and terminatePool. The boost::barrier is now a function-local instance rather than a class member pointer.
-
Steven Andrew Farrell authored
Adding some more complete documentation to the ThreadPoolSvc, ThreadInitTool, ThreadInitTask, and associated interfaces IThreadPoolSvc and IThreadInitTool. Indentation cleanup in ThreadInitTask.
-
- Mar 22, 2016
-
-
Charles Leggett authored
use std::tie for compound comparisons See merge request !3
-
Charles Leggett authored
-
Charles Leggett authored
remove blank line? See merge request !2
-
Charles Leggett authored
-
Charles Leggett authored
added methods to check if contents of EventIDBase are run/event, timestamp, or lumi/event See merge request !1
-
Charles Leggett authored
-
Charles Leggett authored
Adds an EventIDRange class to GaudiKernel, which holds 2 EventIDBase objects add sort methods to EventIDBase, to sort by Run/Event, Timestamp, and Lumi/Event See merge request !130
-
Benedikt Hegner authored
- added release instructions - updated Jemalloc instructions - fixed issue with duplicated entries - enable automatic Doxygen generation via GitLab-CI See merge request !140
-
Benedikt Hegner authored
See GAUDI-1148. See merge request !78
-
Benedikt Hegner authored
See merge request !144
-
Benedikt Hegner authored
Replaced 'x86_64-slc6-gcc48-opt' with 'x86_64-slc0-gcc99-opt' to avoid possible clashes with real platform strings. See merge request !141
-
Marco Clemencic authored
The error that made it impossible to de-register histograms. With this fix the ATLAS code seems to behave as expected. See merge request !143
-
Marco Clemencic authored
-
Marco Clemencic authored
-
- Mar 21, 2016
-
-
Attila Krasznahorkay authored
Fixing a logic error in THistSvc::deReg that makes it impossible to de-register any histograms at the moment
-
- Mar 18, 2016
-
-
Marco Clemencic authored
Replaced 'x86_64-slc6-gcc48-opt' with 'x86_64-slc0-gcc99-opt' to avoid possible clashes with real platform strings.
-
- Mar 16, 2016
-
-
Marco Clemencic authored
-
Marco Clemencic authored
-
Marco Clemencic authored
-
Marco Clemencic authored
-
Marco Clemencic authored
-
Marco Clemencic authored
-
- Mar 15, 2016
-
-
Marco Clemencic authored
-
Marco Clemencic authored
-
Marco Clemencic authored
-
Marco Clemencic authored
-