OpDiLib release v1.7

Restructure data of parallel regions and implicit tasks
- associate tapes, positions, and adjoint access modes with tasks instead of parallel regions
- extend data of parallel regions by pointers to the parent task and child tasks
- backend support for querying the current task's data
- create task data for initial implicit tasks
- track adjoint access mode entirely via tasks
- support resetting tasks according to a tape position, to recover previous adjoint access modes

Features
- track adjoint access mode of initial implicit task
- support additional sync region and worksharing types
- support use of mutexes before tool initialization (OMPT backend)
- support posting messages that are printed during the reverse pass

Instrument
- implement onSetAdjointAccessMode in OmpLogicOutputInstrument
- adapt OmpLogicInstrumentInterface and OmpLogicOutputInstrument according to the other changes

Tests
- run build tests with enabled output instrument
- tests treat non-empty output on stderr as error (can be disabled)
- add tests' error output files to gitignore
- add assertions to adjoint access control tests
- add TestAdjointAccessControlNested2, TestTaskReset

Fixes
- fix transport of adjoint access mode into and out of parallel regions
- fix using directive in second order test driver
- fix placement of skipParallelHandling decrement

Chore
- update license headers for 2025
- update OpDiLib-related publications in README

Merge branch 'develop'

Author: jblueh

.gitignore

@@ -1 +1,2 @@
 **/*.out
+**/*.err

.gitlab-ci.yml

@@ -39,3 +39,16 @@ test-clang-ompt:
     - export BACKEND=OMPT
     - make all
 
+test-output:
+  image: ubuntu:22.04
+  timeout: 2h
+  script:
+    - apt update && apt install -y build-essential binutils git
+    - git clone --depth 1 --branch develop https://github.com/SciCompKL/CoDiPack.git
+    - export CODI_DIR=$(pwd)/CoDiPack/include
+    - export OPDI_DIR=$(pwd)/include
+    - cd tests
+    - export CXX=g++
+    - export OUTPUT_INSTRUMENT=yes
+    - export STDERR_OUTPUT_IS_ERROR=no
+    - make all

README.md

@@ -23,31 +23,48 @@ If you have a code that is differentiated with a serial AD tool and parallelize
 2. **Obtain a first parallel differentiated version of your code.** If your compiler supports OMPT, it suffices to add a few lines of code for the initialization and finalization of OpDiLib. Otherwise, you have to use OpDiLib's macro backend, which involves rewriting your OpenMP constructs according to OpDiLib's macro interface. Both approaches are demonstrated in the minimal example below.
 3. **Optimize the performance of the parallel reverse pass.** Check your parallel forward code for parts that do not involve shared reading. Use OpDiLib's adjoint access control tools to disable atomic adjoints for these parts. You may also revise your data access patterns to eliminate additional instances of shared reading.
 
-## Publication
+## Publications
 
-For further details about OpDiLib's design, features and modes of operation, please refer to our publication
+For further details about OpDiLib's design, features, and modes of operation, please refer to our publication
 [Event-Based Automatic Differentiation of OpenMP with OpDiLib](https://doi.org/10.1145/3570159).
-If you use OpDiLib in one of your applications and write a paper, please cite us!
 
 ~~~~{.txt}
 @article{BluehdornSG2023,
-  author = {Bl\"{u}hdorn, Johannes and Sagebaum, Max and Gauger, Nicolas R.},
   title = {{Event-Based Automatic Differentiation of OpenMP with OpDiLib}},
+  author = {Bl\"{u}hdorn, Johannes and Sagebaum, Max and Gauger, Nicolas R.},
   year = {2023},
-  issue_date = {March 2023},
-  publisher = {Association for Computing Machinery},
-  address = {New York, NY, USA},
+  month = {03},
+  journal = {ACM Trans. Math. Softw.},
   volume = {49},
   number = {1},
-  url = {https://doi.org/10.1145/3570159},
-  doi = {10.1145/3570159},
-  journal = {ACM Trans. Math. Softw.},
-  month = {mar},
-  articleno = {3},
-  numpages = {31}
+  note = {Article No.: 3},
+  pages = {1--31},
+  publisher = {Association for Computing Machinery},
+  address = {New York, NY, USA},
+  doi = {10.1145/3570159}
+}
+~~~~
+
+Additional details on applying OpDiLib in a large software with an advanced AD workflow are given in our publication
+[Hybrid parallel discrete adjoints in SU2](https://doi.org/10.1016/j.compfluid.2024.106528).
+
+~~~~{.txt}
+@article{BluehdornGAG2025,
+  title = {Hybrid parallel discrete adjoints in {SU2}},
+  author = {Bl\"{u}hdorn, Johannes and Gomes, Pedro and Aehle, Max and Gauger, Nicolas R.},
+  year = {2025},
+  month = {01},
+  journal = {Computers & Fluids},
+  volume = {289},
+  note = {Article 106528},
+  pages = {1--18},
+  publisher = {Elsevier},
+  doi = {10.1016/j.compfluid.2024.106528}
 }
 ~~~~
 
+If you use OpDiLib in one of your applications and write a paper, please cite us!
+
 ## Minimal Example
 
 The following minimal example assumes that [CoDiPack](https://scicomp.rptu.de/software/codi/) is used as the underlying AD tool. You need CoDiPack [version 2.1](https://github.com/SciCompKL/CoDiPack/releases/tag/v2.1.0) or newer. For additional examples, please refer to OpDiLib's test suite.

include/opdi.cpp

@@ -2,7 +2,7 @@
  * OpDiLib, an Open Multiprocessing Differentiation Library
  *
  * Copyright (C) 2020-2022 Chair for Scientific Computing (SciComp), TU Kaiserslautern
- * Copyright (C) 2023-2024 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
+ * Copyright (C) 2023-2025 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
  * Homepage: https://scicomp.rptu.de
  * Contact:  Prof. Nicolas R. Gauger (opdi@scicomp.uni-kl.de)
  *

include/opdi.hpp

@@ -2,7 +2,7 @@
  * OpDiLib, an Open Multiprocessing Differentiation Library
  *
  * Copyright (C) 2020-2022 Chair for Scientific Computing (SciComp), TU Kaiserslautern
- * Copyright (C) 2023-2024 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
+ * Copyright (C) 2023-2025 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
  * Homepage: https://scicomp.rptu.de
  * Contact:  Prof. Nicolas R. Gauger (opdi@scicomp.uni-kl.de)
  *

include/opdi/backend/backendInterface.hpp

@@ -2,7 +2,7 @@
  * OpDiLib, an Open Multiprocessing Differentiation Library
  *
  * Copyright (C) 2020-2022 Chair for Scientific Computing (SciComp), TU Kaiserslautern
- * Copyright (C) 2023-2024 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
+ * Copyright (C) 2023-2025 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
  * Homepage: https://scicomp.rptu.de
  * Contact:  Prof. Nicolas R. Gauger (opdi@scicomp.uni-kl.de)
  *
@@ -42,6 +42,9 @@ namespace opdi {
       virtual std::size_t getNestedLockIdentifier(omp_nest_lock_t* lock) = 0;
 
       virtual void* getParallelData() = 0;
+      virtual void* getTaskData() = 0;
+
+      virtual void setInitialImplicitTaskData(void* data) = 0;
   };
 
   extern BackendInterface* backend;

include/opdi/backend/macro/dataTools.hpp

@@ -2,7 +2,7 @@
  * OpDiLib, an Open Multiprocessing Differentiation Library
  *
  * Copyright (C) 2020-2022 Chair for Scientific Computing (SciComp), TU Kaiserslautern
- * Copyright (C) 2023-2024 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
+ * Copyright (C) 2023-2025 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
  * Homepage: https://scicomp.rptu.de
  * Contact:  Prof. Nicolas R. Gauger (opdi@scicomp.uni-kl.de)
  *
@@ -34,6 +34,9 @@ namespace opdi {
       static std::stack<void*> parallelData;
       #pragma omp threadprivate(parallelData)
 
+      static std::stack<void*> taskData;
+      #pragma omp threadprivate(taskData)
+
     public:
       static void pushParallelData(void* parallelData) {
         DataTools::parallelData.push(parallelData);
@@ -50,5 +53,21 @@ namespace opdi {
 
         return DataTools::parallelData.top();
       }
+
+      static void pushTaskData(void* taskData) {
+        DataTools::taskData.push(taskData);
+      }
+
+      static void popTaskData() {
+        DataTools::taskData.pop();
+      }
+
+      static void* getTaskData() {
+        if (DataTools::taskData.empty()) {
+          return nullptr;
+        }
+
+        return DataTools::taskData.top();
+      }
   };
 }

include/opdi/backend/macro/implicitBarrierTools.hpp

@@ -2,7 +2,7 @@
  * OpDiLib, an Open Multiprocessing Differentiation Library
  *
  * Copyright (C) 2020-2022 Chair for Scientific Computing (SciComp), TU Kaiserslautern
- * Copyright (C) 2023-2024 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
+ * Copyright (C) 2023-2025 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
  * Homepage: https://scicomp.rptu.de
  * Contact:  Prof. Nicolas R. Gauger (opdi@scicomp.uni-kl.de)
  *

include/opdi/backend/macro/macroBackend.cpp

@@ -2,7 +2,7 @@
  * OpDiLib, an Open Multiprocessing Differentiation Library
  *
  * Copyright (C) 2020-2022 Chair for Scientific Computing (SciComp), TU Kaiserslautern
- * Copyright (C) 2023-2024 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
+ * Copyright (C) 2023-2025 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
  * Homepage: https://scicomp.rptu.de
  * Contact:  Prof. Nicolas R. Gauger (opdi@scicomp.uni-kl.de)
  *
@@ -30,6 +30,7 @@
 // static macro backend members
 
 std::stack<void*> opdi::DataTools::parallelData;
+std::stack<void*> opdi::DataTools::taskData;
 
 std::stack<bool> opdi::ImplicitBarrierTools::implicitBarrierStack;
 

include/opdi/backend/macro/macroBackend.hpp

@@ -2,7 +2,7 @@
  * OpDiLib, an Open Multiprocessing Differentiation Library
  *
  * Copyright (C) 2020-2022 Chair for Scientific Computing (SciComp), TU Kaiserslautern
- * Copyright (C) 2023-2024 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
+ * Copyright (C) 2023-2025 Chair for Scientific Computing (SciComp), University of Kaiserslautern-Landau
  * Homepage: https://scicomp.rptu.de
  * Contact:  Prof. Nicolas R. Gauger (opdi@scicomp.uni-kl.de)
  *
@@ -25,6 +25,8 @@
 
 #pragma once
 
+#include <cassert>
+
 #include "../../config.hpp"
 
 #ifdef OPDI_BACKEND
@@ -57,9 +59,15 @@ namespace opdi {
 
       void init() {
         opdi_init_lock(&ReductionTools::globalReducerLock);
+
+        // task data for initial implicit task is created in the logic layer
       }
 
       void finalize() {
+        // pop task data associated with initial implicit task
+        DataTools::popTaskData();
+        assert(DataTools::getTaskData() == nullptr);
+
         opdi_set_lock(&ReductionTools::globalReducerLock);
 
         for (auto lock : ReductionTools::individualReducerLocks) {
@@ -74,6 +82,15 @@ namespace opdi {
       void* getParallelData() {
         return DataTools::getParallelData();
       }
+
+      void* getTaskData() {
+        return DataTools::getTaskData();
+      }
+
+      void setInitialImplicitTaskData(void* data) {
+        assert(DataTools::getTaskData() == nullptr);
+        DataTools::pushTaskData(data);
+      }
   };
 
 }