Merge pull request #279 from LLNL/v0.13.3-RC

v0.13.3 RC
LLNL · Jul 31, 2024 · 9385369 · 9385369
2 parents b25dcd2 + 30adbc6
commit 9385369
Show file tree

Hide file tree

Showing 8 changed files with 24 additions and 17 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -17,7 +17,7 @@ cmake_minimum_required(VERSION 3.23)
 
 project(CARE
         LANGUAGES C CXX
-        VERSION 0.13.2)
+        VERSION 0.13.3)
 
 include(${PROJECT_SOURCE_DIR}/cmake/Setup.cmake)
 

diff --git a/README.md b/README.md
@@ -5,7 +5,7 @@
 [comment]: # (SPDX-License-Identifier: BSD-3-Clause)
 [comment]: # (#################################################################)
 
-# CARE v0.13.2
+# CARE v0.13.3
 
 CARE: CHAI and RAJA Extensions
 ===============================

diff --git a/RELEASE_NOTES.md b/RELEASE_NOTES.md
@@ -12,6 +12,12 @@ in this file.
 
 The format of this file is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
+## [Version 0.13.3] - Release date 2024-07-31
+
+### Fixed
+- Replaced loop\_work alias with seq\_work (loop\_work was removed in RAJA v2024.02.2)
+- Fixed CHUNKED loop macro implementations
+
 ## [Version 0.13.2] - Release date 2024-07-29
 
 ### Changed

diff --git a/docs/sphinx/conf.py b/docs/sphinx/conf.py
@@ -57,7 +57,7 @@
 # The short X.Y version.
 version = '0.13'
 # The full version, including alpha/beta/rc tags.
-release = '0.13.2'
+release = '0.13.3'
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.

diff --git a/scripts/make_release_tarball.sh b/scripts/make_release_tarball.sh
@@ -8,7 +8,7 @@
 ##############################################################################
 
 TAR_CMD=gtar
-VERSION=0.13.2
+VERSION=0.13.3
 
 git archive --prefix=care-${VERSION}/ -o care-${VERSION}.tar HEAD 2> /dev/null
 

diff --git a/src/care/DefaultMacros.h b/src/care/DefaultMacros.h
@@ -313,7 +313,7 @@ OMP_FOR_BEGIN for (auto INDEX = _care_openmp_for_loop_chunk_begin_ndx; INDEX < _
 /// @arg[in] CHECK The variable to check that the start and end macros match
 ///
 ////////////////////////////////////////////////////////////////////////////////
-#define CARE_CHECKED_CHUNKED_GPU_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, CHECK) CARE_CHECKED_OPENMP_FOR_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, CHECK)
+#define CARE_CHECKED_CHUNKED_GPU_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, CHECK) CARE_CHECKED_CHUNKED_OPENMP_FOR_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, CHECK)
 
 #define CARE_CHECKED_CHUNKED_GPU_LOOP_END(CHECK) CARE_CHECKED_CHUNKED_OPENMP_FOR_LOOP_END(CHECK)
 
@@ -366,12 +366,12 @@ OMP_FOR_BEGIN for (auto INDEX = _care_openmp_for_loop_chunk_begin_ndx; INDEX < _
 /// @arg[in] CHECK The variable to check that the start and end macros match
 ///
 ////////////////////////////////////////////////////////////////////////////////
-#define CARE_CHECKED_CHUNKED_PARALLEL_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, CHECK) CARE_CHECKED_OPENMP_FOR_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, CHECK)
+#define CARE_CHECKED_CHUNKED_PARALLEL_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, CHECK) CARE_CHECKED_CHUNKED_OPENMP_FOR_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, CHECK)
 
-#define CARE_CHECKED_CHUNKED_PARALLEL_LOOP_END(CHECK) CARE_CHECKED_OPENMP_FOR_LOOP_END(CHECK)
+#define CARE_CHECKED_CHUNKED_PARALLEL_LOOP_END(CHECK) CARE_CHECKED_CHUNKED_OPENMP_FOR_LOOP_END(CHECK)
 
 #define CARE_CHECKED_CHUNKED_REDUCE_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, CHECK) \
-   CARE_CHECKED_PARALLEL_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, CHECK)
+   CARE_CHECKED_CHUNKED_PARALLEL_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, CHECK)
 
 #define CARE_CHECKED_CHUNKED_REDUCE_LOOP_END(CHECK) CARE_CHECKED_CHUNKED_PARALLEL_LOOP_END(CHECK)
 
@@ -901,9 +901,9 @@ OMP_FOR_BEGIN for (auto INDEX = _care_openmp_for_loop_chunk_begin_ndx; INDEX < _
 /// @arg[in] CHUNK_SIZE Maximum kernel size
 ///
 ////////////////////////////////////////////////////////////////////////////////
-#define CARE_CHUNKED_LOOP(POLICY, INDEX, START_INDEX, END_INDEX, CHUNK_SIZE) CARE_CHECKED_LOOP_START(POLICY, INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, care_loop_chunked_check)
+#define CARE_CHUNKED_LOOP(POLICY, INDEX, START_INDEX, END_INDEX, CHUNK_SIZE) CARE_CHECKED_CHUNKED_LOOP_START(POLICY, INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, care_loop_chunked_check)
 
-#define CARE_CHUNKED_LOOP_END CARE_CHECKED_LOOP_END(care_loop_chunked_check)
+#define CARE_CHUNKED_LOOP_END CARE_CHECKED_CHUNKED_LOOP_END(care_loop_chunked_check)
 
 ////////////////////////////////////////////////////////////////////////////////
 ///
@@ -1041,7 +1041,7 @@ OMP_FOR_BEGIN for (auto INDEX = _care_openmp_for_loop_chunk_begin_ndx; INDEX < _
 /// @arg[in] CHUNK_SIZE Maximum kernel size
 ///
 ////////////////////////////////////////////////////////////////////////////////
-#define CARE_CHUNKED_GPU_LOOP(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE) CARE_CHECKED_GPU_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, care_gpu_loop_chunked_check)
+#define CARE_CHUNKED_GPU_LOOP(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE) CARE_CHECKED_CHUNKED_GPU_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, care_gpu_loop_chunked_check)
 
 #define CARE_CHUNKED_GPU_LOOP_END CARE_CHECKED_CHUNKED_GPU_LOOP_END(care_gpu_loop_chunked_check)
 
@@ -1149,7 +1149,7 @@ OMP_FOR_BEGIN for (auto INDEX = _care_openmp_for_loop_chunk_begin_ndx; INDEX < _
 /// @arg[in] CHUNK_SIZE Maximum kernel size
 ///
 ////////////////////////////////////////////////////////////////////////////////
-#define CARE_CHUNKED_WORK_LOOP(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE) CARE_CHECKED_PARALLEL_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, care_work_loop_chunked_check)
+#define CARE_CHUNKED_WORK_LOOP(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE) CARE_CHECKED_CHUNKED_PARALLEL_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, care_work_loop_chunked_check)
 
 #define CARE_CHUNKED_WORK_LOOP_END CARE_CHECKED_CHUNKED_PARALLEL_LOOP_END(care_work_loop_chunked_check)
 
@@ -1223,7 +1223,7 @@ OMP_FOR_BEGIN for (auto INDEX = _care_openmp_for_loop_chunk_begin_ndx; INDEX < _
 /// @arg[in] CHUNK_SIZE Maximum kernel size
 ///
 ////////////////////////////////////////////////////////////////////////////////
-#define CARE_CHUNKED_REDUCE_LOOP(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE) CARE_CHECKED_REDUCE_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, care_reduce_loop_chunked_check)
+#define CARE_CHUNKED_REDUCE_LOOP(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE) CARE_CHECKED_CHUNKED_REDUCE_LOOP_START(INDEX, START_INDEX, END_INDEX, CHUNK_SIZE, care_reduce_loop_chunked_check)
 
 #define CARE_CHUNKED_REDUCE_LOOP_END CARE_CHECKED_CHUNKED_REDUCE_LOOP_END(care_reduce_loop_chunked_check)
 

diff --git a/src/care/LoopFuser.h b/src/care/LoopFuser.h
@@ -26,6 +26,8 @@ constexpr double CARE_DEFAULT_PHASE = -FLT_MAX/2.0;
 
 #if CARE_ENABLE_LOOP_FUSER
 
+#include "RAJA/RAJA.hpp"
+
 #include "umpire/Allocator.hpp"
 #include "umpire/TypedAllocator.hpp"
 
@@ -581,11 +583,11 @@ class LoopFuser : public FusedActions {
                                  RAJA::constant_stride_array_of_objects >;
 #else
       using workgroup_policy = RAJA::WorkGroupPolicy <
-                                 RAJA::loop_work,
+                                 RAJA::seq_work,
                                  RAJA::ordered,
                                  RAJA::ragged_array_of_objects >;
       using workgroup_ordered_policy = RAJA::WorkGroupPolicy <
-                                 RAJA::loop_work,
+                                 RAJA::seq_work,
                                  RAJA::ordered,
                                  RAJA::ragged_array_of_objects >;
 #endif

diff --git a/test/TestForall.cpp b/test/TestForall.cpp
@@ -79,7 +79,7 @@ CPU_TEST(forall, chunked_dynamic_policy)
    const int length = 10;
    care::host_device_ptr<int> temp(length, "temp");
 
-   CARE_LOOP(care::Policy::sequential, i, 0, length, batch_size) {
+   CARE_CHUNKED_LOOP(care::Policy::sequential, i, 0, length, batch_size) {
       temp[i] = i;
    } CARE_CHUNKED_LOOP_END
 
@@ -185,6 +185,5 @@ GPU_TEST(forall, chunked_dynamic_policy)
    temp.free();
 }
 
-
 #endif // CARE_GPUCC