ROCm · samjwu · Aug 7, 2024 · Jun 17, 2024 · Jun 20, 2024 · Jun 24, 2024
diff --git a/.Doxyfile b/.Doxyfile
@@ -960,16 +960,16 @@ INPUT                  = README.md \
                         include/rppi_logical_operations.h \
                         include/rppi_morphological_transforms.h \
                         include/rppi_statistical_operations.h \
+                        include/rppt_tensor_arithmetic_operations.h \
+                        include/rppt_tensor_audio_augmentations.h \
                         include/rppt_tensor_color_augmentations.h \
                         include/rppt_tensor_data_exchange_operations.h \
                         include/rppt_tensor_effects_augmentations.h \
                         include/rppt_tensor_filter_augmentations.h \
                         include/rppt_tensor_geometric_augmentations.h \
+                        include/rppt_tensor_logical_operations.h \
                         include/rppt_tensor_morphological_operations.h \
-                        include/rppt_tensor_statistical_operations.h \
-                        include/rppt_tensor_arithmetic_operations.h \
-                        include/rppt_tensor_audio_augmentations.h \
-                        include/rppt_tensor_logical_operations.h
+                        include/rppt_tensor_statistical_operations.h
 
 
 # This tag can be used to specify the character encoding of the source files
@@ -2381,7 +2381,7 @@ INCLUDE_FILE_PATTERNS  =
 # recursively expanded use the := operator instead of the = operator.
 # This tag requires that the tag ENABLE_PREPROCESSING is set to YES.
 
-PREDEFINED             = GPU_SUPPORT RPP_BACKEND_HIP HIP_COMPILE
+PREDEFINED             = GPU_SUPPORT RPP_BACKEND_HIP HIP_COMPILE AUDIO_SUPPORT
 
 # If the MACRO_EXPANSION and EXPAND_ONLY_PREDEF tags are set to YES then this
 # tag can be used to specify a list of macro names that should be expanded. The

diff --git a/.azuredevops/rocm-ci.yml b/.azuredevops/rocm-ci.yml
@@ -13,6 +13,8 @@ trigger:
   batch: true
   branches:
     include:
+    - develop
+    - mainline
     - master
   paths:
     exclude:
@@ -27,8 +29,9 @@ pr:
   autoCancel: true
   branches:
     include:
-    - master
     - develop
+    - mainline
+    - master
   paths:
     exclude:
     - .github

@@ -1,8 +1,8 @@
 # Changelog for RPP
 
-Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/rpp/en/latest/).
+Full documentation for RPP is available at [https://rocm.docs.amd.com/projects/rpp/en/latest](https://rocm.docs.amd.com/projects/rpp/en/latest)
 
-### RPP 1.8.0 (unreleased)
+## (Unreleased) RPP 1.8.0
 
 ### Changes
 
@@ -25,7 +25,7 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 * CMake - Version `3.22.3`
 * IEEE 754-based half-precision floating-point library - Version `1.12.0`
 
-### RPP 1.5.0
+### RPP 1.5.0 for ROCm 6.1.1
 
 ### Changes
 
@@ -42,7 +42,7 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 * CMake - Version `3.22.3`
 * IEEE 754-based half-precision floating-point library - Version `1.12.0`
 
-## RPP 1.4.0
+## RPP 1.4.0 for ROCm 6.0.0
 
 ### Additions
 
@@ -76,7 +76,7 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 * CMake - Version `3.22.3`
 * IEEE 754-based half-precision floating-point library - Version `1.12.0`
 
-## RPP 1.3.0
+## RPP 1.3.0 for ROCm 5.7.1
 
 ### Additions
 
@@ -106,7 +106,7 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 * Boost - Version `1.72`
 * IEEE 754-based half-precision floating-point library - Version `1.12.0`
 
-## RPP 1.2.0
+## RPP 1.2.0 for ROCm 5.7.1
 
 ### Additions
 
@@ -137,7 +137,7 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 * Boost - Version `1.72`
 * IEEE 754-based half-precision floating-point library - Version `1.12.0`
 
-## RPP 1.1.0
+## RPP 1.1.0 for ROCm 5.7.0
 
 ### Additions
 
@@ -172,7 +172,7 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 * SLES - the Clang package is missing in the latest updates, which means Clang must be manually
   installed.
 
-## RPP 1.0.0
+## RPP 1.0.0 for ROCm 5.7.0
 
 ### Additions
 
@@ -212,7 +212,7 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 * SLES - the Clang package is missing in the latest updates, which means Clang must be manually
   installed.
 
-## RPP 0.99
+## RPP 0.99 for ROCm 5.7.0
 
 ### Additions
 
@@ -241,7 +241,7 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 * Boost - Version `1.72`
 * IEEE 754-based half-precision floating-point library - Version `1.12.0`
 
-## RPP 0.98
+## RPP 0.98 for ROCm 5.7.0
 
 ### Additions
 * Dockers
@@ -251,11 +251,11 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 
 * Readme updates
 
-#### Changes
+### Changes
 
 * CMakeList
 
-#### Fixes
+### Fixes
 
 * Minor bugs and warnings
 
@@ -270,7 +270,7 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 * Boost - Version `1.72`
 * IEEE 754-based half-precision floating-point library - Version `1.12.0`
 
-## RPP 0.97
+## RPP 0.97 for ROCm 5.7.0
 
 ### Additions
 
@@ -301,7 +301,7 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 * Boost - Version `1.72`
 * IEEE 754-based half-precision floating-point library - Version `1.12.0`
 
-## RPP 0.96
+## RPP 0.96 for ROCm 5.7.0
 
 ### Additions
 
@@ -334,7 +334,7 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 
 * RPP is not supported on CentOS 7 and SLES SP2
 
-## RPP 0.95
+## RPP 0.95 for ROCm 5.7.0
 
 ### Additions
 
@@ -368,7 +368,7 @@ Full documentation for RPP is available at (https://rocm.docs.amd.com/projects/r
 
 * ROCm reorganization: install updates no longer match ROCm specifications
 
-## RPP 0.93
+## RPP 0.93 for ROCm 5.7.0
 
 ### Additions
 

diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -58,6 +58,9 @@ endif(CMAKE_INSTALL_PREFIX_INITIALIZED_TO_DEFAULT)
 
 # RPP Default Options
 set(DEFAULT_BUILD_TYPE "Release")
+### RPP_AUDIO_SUPPORT - default = ON, NOTE: support currently only on Ubuntu - user to set to OFF otherwise
+option(RPP_AUDIO_SUPPORT      "Build RPP with Audio Support"         ON)
+option(BUILD_WITH_AMD_ADVANCE "Build RPP for advanced AMD GPU Architecture"    OFF)
 
 # Set message options
 if(NOT WIN32)
@@ -77,6 +80,7 @@ endif()
 if(APPLE)
     set(CMAKE_MACOSX_RPATH 1)
     set(BACKEND "CPU")
+    set(RPP_AUDIO_SUPPORT OFF)
     message("-- ${Magenta}Apple macOS Detected -- GPU Support turned OFF${ColourReset}")
 endif()
 
@@ -134,9 +138,16 @@ list(APPEND CMAKE_MODULE_PATH ${PROJECT_SOURCE_DIR}/cmake)
 find_package(HALF REQUIRED)
 include_directories(${HALF_INCLUDE_DIRS})
 
+if (RPP_AUDIO_SUPPORT)
+    add_definitions(-DAUDIO_SUPPORT)        # For compile flags in RPP
+    set(RPP_AUDIO_AUGMENTATIONS_SUPPORT 1)  # For cmakedefine01 in rpp_audio_augmentations_support.h.in
+endif()
+
 message("-- ${Cyan}RPP Developer Options${ColourReset}")
 message("-- ${Cyan}     -D BACKEND=${BACKEND} [Select RPP Backend [options:CPU/OPENCL/HIP](default:HIP)]${ColourReset}")
 message("-- ${Cyan}     -D CMAKE_BUILD_TYPE=${CMAKE_BUILD_TYPE} [Select RPP build type [options:Debug/Release](default:Release)]${ColourReset}")
+message("-- ${Cyan}     -D RPP_AUDIO_SUPPORT=${RPP_AUDIO_SUPPORT} [Select RPP audio support [options:ON/OFF](default:ON)]${ColourReset}")
+message("-- ${Cyan}     -D BUILD_WITH_AMD_ADVANCE=${BUILD_WITH_AMD_ADVANCE} [Turn ON/OFF Build for AMD advanced GPUs(default:OFF)]${ColourReset}")
 
 # OpenMP
 find_package(OpenMP REQUIRED)
@@ -207,8 +218,23 @@ if("${BACKEND}" STREQUAL "HIP")
     endif()
 
     list(APPEND CMAKE_PREFIX_PATH ${ROCM_PATH} ${ROCM_PATH}/hip)
+
+    # Set supported GPU Targets
     set(DEFAULT_AMDGPU_TARGETS "gfx908;gfx90a;gfx940;gfx941;gfx942;gfx1030;gfx1031;gfx1032;gfx1100;gfx1101;gfx1102")
-    set(AMDGPU_TARGETS "${DEFAULT_AMDGPU_TARGETS}" CACHE STRING "List of specific machine types for library to target")
+    if (BUILD_WITH_AMD_ADVANCE)
+        set(DEFAULT_AMDGPU_TARGETS ${DEFAULT_AMDGPU_TARGETS} "gfx1200;gfx1201")
+    endif()
+
+    # Set AMDGPU_TARGETS
+    if(DEFINED ENV{AMDGPU_TARGETS})
+        set(AMDGPU_TARGETS $ENV{AMDGPU_TARGETS} CACHE STRING "List of specific machine types for library to target")
+    elseif(AMDGPU_TARGETS)
+        message("-- ${White}${PROJECT_NAME} -- AMDGPU_TARGETS set with -D option${ColourReset}")
+    else()
+        set(AMDGPU_TARGETS "${DEFAULT_AMDGPU_TARGETS}" CACHE STRING "List of specific machine types for library to target")
+    endif()
+    message("-- ${White}${PROJECT_NAME} -- AMDGPU_TARGETS: ${AMDGPU_TARGETS}${ColourReset}")
+
     find_package(HIP QUIET)
     if(HIP_FOUND)
         message("-- ${White}${PROJECT_NAME} -- Using HIP - Path:" ${HIP_PATH} "\tVersion:" ${HIP_VERSION} "\tCompiler:" ${HIP_COMPILER}${ColourReset})
@@ -273,6 +299,7 @@ if("${BACKEND}" STREQUAL "CPU")
 endif()
 
 configure_file("${PROJECT_SOURCE_DIR}/include/rpp_backend.h.in" "${PROJECT_BINARY_DIR}/include/rpp_backend.h")
+configure_file("${PROJECT_SOURCE_DIR}/include/rpp_audio_augmentations_support.h.in" "${PROJECT_BINARY_DIR}/include/rpp_audio_augmentations_support.h")
 
 # Enable SIMD for HOST code (in both OpenCL and HIP backends)
 if(NOT DEFINED SIMD_ENABLE)
@@ -294,7 +321,12 @@ message("-- ${White}${PROJECT_NAME} -- Link Libraries: ${LINK_LIBRARY_LIST}${Col
 target_link_libraries(${PROJECT_NAME} ${LINK_LIBRARY_LIST})
 set_target_properties(${PROJECT_NAME} PROPERTIES POSITION_INDEPENDENT_CODE ON)
 set_target_properties(${PROJECT_NAME} PROPERTIES LINKER_LANGUAGE CXX)
-target_link_libraries(${PROJECT_NAME} ${PROJECT_SOURCE_DIR}/libs/third_party/ffts/libffts.a)
+if(RPP_AUDIO_SUPPORT)
+    target_link_libraries(${PROJECT_NAME} ${PROJECT_SOURCE_DIR}/libs/third_party/ffts/libffts.a)
+    message("-- ${Green}${PROJECT_NAME} set to build with RPP_AUDIO_SUPPORT${ColourReset}")
+else()
+    message("-- ${Yellow}${PROJECT_NAME} set to build without RPP_AUDIO_SUPPORT${ColourReset}")
+endif()
 set_target_properties(${PROJECT_NAME} PROPERTIES VERSION ${PROJECT_VERSION} SOVERSION ${PROJECT_VERSION_MAJOR})
 
 target_include_directories(${PROJECT_NAME}
@@ -335,6 +367,9 @@ install(DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/include/ DESTINATION ${CMAKE_INSTA
 install(FILES ${PROJECT_BINARY_DIR}/include/rpp_backend.h
         DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}/rpp
         COMPONENT dev)
+install(FILES ${PROJECT_BINARY_DIR}/include/rpp_audio_augmentations_support.h
+        DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}/rpp
+        COMPONENT dev)
 
 # install Test
 install(DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}/cmake DESTINATION ${CMAKE_INSTALL_DATADIR}/${PROJECT_NAME}/test COMPONENT test)

diff --git a/docs/data/doxygenInputs/lens_img640x480.png b/docs/data/doxygenInputs/lens_img640x480.png
diff --git a/docs/data/doxygenOutputs/effects_augmentations_glitch_img150x150.png b/docs/data/doxygenOutputs/effects_augmentations_glitch_img150x150.png
diff --git a/docs/data/doxygenOutputs/effects_augmentations_jitter_150x150.png b/docs/data/doxygenOutputs/effects_augmentations_jitter_150x150.png
diff --git a/docs/data/doxygenOutputs/geometric_augmentations_lens_correction_img_640x480.png b/docs/data/doxygenOutputs/geometric_augmentations_lens_correction_img_640x480.png
diff --git a/docs/doxygen/Doxyfile b/docs/doxygen/Doxyfile
@@ -962,14 +962,16 @@ INPUT                  = ../../README.md \
                         ../../include/rppi_logical_operations.h \
                         ../../include/rppi_morphological_transforms.h \
                         ../../include/rppi_statistical_operations.h \
+                        ../../include/rppt_tensor_arithmetic_operations.h \
+                        ../../include/rppt_tensor_audio_augmentations.h \
                         ../../include/rppt_tensor_color_augmentations.h \
                         ../../include/rppt_tensor_data_exchange_operations.h \
                         ../../include/rppt_tensor_effects_augmentations.h \
                         ../../include/rppt_tensor_filter_augmentations.h \
                         ../../include/rppt_tensor_geometric_augmentations.h \
+                        ../../include/rppt_tensor_logical_operations.h \
                         ../../include/rppt_tensor_morphological_operations.h \
-                        ../../include/rppt_tensor_statistical_operations.h \
-                        ../../include/rppt_tensor_logical_operations.h
+                        ../../include/rppt_tensor_statistical_operations.h
 
 
 # This tag can be used to specify the character encoding of the source files
@@ -2381,7 +2383,7 @@ INCLUDE_FILE_PATTERNS  =
 # recursively expanded use the := operator instead of the = operator.
 # This tag requires that the tag ENABLE_PREPROCESSING is set to YES.
 
-PREDEFINED             = GPU_SUPPORT RPP_BACKEND_HIP HIP_COMPILE
+PREDEFINED             = GPU_SUPPORT RPP_BACKEND_HIP HIP_COMPILE AUDIO_SUPPORT
 
 # If the MACRO_EXPANSION and EXPAND_ONLY_PREDEF tags are set to YES then this
 # tag can be used to specify a list of macro names that should be expanded. The

diff --git a/docs/sphinx/requirements.in b/docs/sphinx/requirements.in
@@ -1 +1 @@
-rocm-docs-core[api_reference]==1.4.0
+rocm-docs-core[api_reference]==1.5.1
diff --git a/docs/sphinx/requirements.txt b/docs/sphinx/requirements.txt
@@ -110,7 +110,7 @@ requests==2.28.2
     # via
     #   pygithub
     #   sphinx
-rocm-docs-core[api-reference]==1.4.0
+rocm-docs-core[api-reference]==1.5.1
     # via -r requirements.in
 smmap==5.0.0
     # via gitdb

diff --git a/include/rpp_audio_augmentations_support.h.in b/include/rpp_audio_augmentations_support.h.in
@@ -0,0 +1,6 @@
+#ifndef GUARD_RPP_AUDIO_AUGMENTATIONS_SUPPORT_H_IN
+#define GUARD_RPP_AUDIO_AUGMENTATIONS_SUPPORT_H_IN
+
+#cmakedefine01 RPP_AUDIO_AUGMENTATIONS_SUPPORT
+
+#endif
diff --git a/include/rppdefs.h b/include/rppdefs.h
@@ -64,6 +64,7 @@ SOFTWARE.
 const float ONE_OVER_6 = 1.0f / 6;
 const float ONE_OVER_3 = 1.0f / 3;
 const float ONE_OVER_255 = 1.0f / 255;
+const uint MMS_MAX_SCRATCH_MEMORY = 76800000; // maximum scratch memory size (number of floats) needed for MMS buffer in RNNT training
 
 /******************** RPP typedefs ********************/
 
@@ -136,7 +137,15 @@ typedef enum
     /*! \brief src and dst layout mismatch \ingroup group_rppdefs */
     RPP_ERROR_LAYOUT_MISMATCH           = -18,
     /*! \brief Number of channels is invalid. (Needs to adhere to function specification.) \ingroup group_rppdefs */
-    RPP_ERROR_INVALID_CHANNELS          = -19
+    RPP_ERROR_INVALID_CHANNELS          = -19,
+    /*! \brief Invalid output tile length (Needs to adhere to function specification.) \ingroup group_rppdefs */
+    RPP_ERROR_INVALID_OUTPUT_TILE_LENGTH    = -20,
+    /*! \brief Shared memory size needed is beyond the bounds (Needs to adhere to function specification.) \ingroup group_rppdefs */
+    RPP_ERROR_OUT_OF_BOUND_SHARED_MEMORY_SIZE    = -21,
+    /*! \brief Scratch memory size needed is beyond the bounds (Needs to adhere to function specification.) \ingroup group_rppdefs */
+    RPP_ERROR_OUT_OF_BOUND_SCRATCH_MEMORY_SIZE    = -22,
+    /*! \brief Number of src dims is invalid. (Needs to adhere to function specification.) \ingroup group_rppdefs */
+    RPP_ERROR_INVALID_SRC_DIMS          = -23
 } RppStatus;
 
 /*! \brief RPP rppStatus_t type enums
@@ -446,6 +455,16 @@ typedef struct
 
 } RpptRoiLtrb;
 
+/*! \brief RPPT Tensor Channel Offsets struct
+ * \ingroup group_rppdefs
+ */
+typedef struct
+{
+    RppiPoint r;
+    RppiPoint g;
+    RppiPoint b;
+} RpptChannelOffsets;
+
 /*! \brief RPPT Tensor 3D ROI LTFRBB struct
  * \ingroup group_rppdefs
  */