GPA 3.4.1 release

chesik-amd · chesik-amd · commit 297d2a7bb54e · 2019-07-29T14:27:04.000-04:00
diff --git a/README.md b/README.md
@@ -51,13 +51,13 @@ Prebuilt binaries can be downloaded from the Releases page: https://github.com/G
 
 ## System Requirements
 * An AMD Radeon GPU or APU based on Graphics IP version 8 and newer.
-* Windows: Radeon Software Adrenaline 2019 Edition 19.7.1 or later (Driver Packaging Version 19.30 or later).
+* Windows: Radeon Software Adrenaline 2019 Edition 19.7.3 or later (Driver Packaging Version 19.30 or later).
 * Linux: Radeon Software for Linux Revision 19.20 or later (19.30 required for AMD Radeon 5700 Series GPUs).
 * HSA/ROCm version requires ROCm 2.0 or newer.
   * For use with ROCm, please make sure that the optional ROCm package "rocprofiler-dev" is installed.
     * Ubuntu: sudo apt install rocprofiler-dev
     * CentOS: sudo yum install rocprofiler-dev
-* Radeon GPUs or APUs based on Graphics IP version 6 and 7 are no longer supported by GPUPerfAPI. Please use an older version (([3.3](https://github.com/GPUOpen-Tools/GPA/releases/tag/v3.3))) with older hardware.
+* Radeon GPUs or APUs based on Graphics IP version 6 and 7 are no longer supported by GPUPerfAPI. Please use an older version ([3.3](https://github.com/GPUOpen-Tools/GPA/releases/tag/v3.3)) with older hardware.
 * Windows 7, 8.1, and 10.
 * Ubuntu (16.04 and later) and CentOS/RHEL (7 and later) distributions.
 
@@ -110,13 +110,11 @@ build of GPUPerfAPI that also exposes the raw hardware counters, but that is a d
    * sudo chmod ugo+w /sys/class/drm/card0/device/power_dpm_force_performance_level
  * The following performance counter values may not be accurate for DirectX 11 applications running on a Radeon 5700 Series GPU. This is expected to be addressed in a future
    driver release:
-   * VSVerticesIn, HSPatches, DSVerticesIn: These may incorrectly report zero values.
    * VALUInstCount, SALUInstCount, VALUBusy, SALUBusy for all shader stages: These values should be representative of performance, but may not be 100% accurate.
    * Most of the ComputeShader counters (all except the MemUnit and WriteUnit counters): These values should be representative of performance, but may not be 100% accurate.
  * The following performance counter values may not be accurate for OpenGL applications running on a Radeon 5700 Series GPU. This is expected to be addressed in a future
    driver release:
    * VALUInstCount, SALUInstCount, VALUBusy, SALUBusy for all shader stages: These values should be representative of performance, but may not be 100% accurate.
    * Most of the ComputeShader counters (all except the MemUnit and WriteUnit counters): These values should be representative of performance, but may not be 100% accurate.
-   * MemUnit counters (in ComputeShader and GlobalMemory groups), TextureUnit counters and L0Cache counters will be incorrect.
- * Setting the GPU clock mode is not working correctly for Radeon 5700 Series GPUs, potentially leading to some inconsistencies in counter values from one run to the next.
-   This is expected to be addressed in a future driver release.
+ * On Linux, setting the GPU clock mode is not working correctly for Radeon 5700 Series GPUs, potentially leading to some inconsistencies in counter values from one run to the
+   next. This is expected to be addressed in a future driver release.
diff --git a/Src/Examples/Vulkan/VkColorCube/CMakeLists.txt b/Src/Examples/Vulkan/VkColorCube/CMakeLists.txt
@@ -35,13 +35,13 @@ if(WIN32)
                        COMMAND CompileShaders.bat ${GPA_OUTPUT_DIR}/$<CONFIG>/
                        BYPRODUCTS vkcolorcubeshader.vert.spv vkcolorcubeshader.frag.spv
                        WORKING_DIRECTORY ${CMAKE_CURRENT_LIST_DIR}
-                       COMMENT "Compiling Vulkan Shaders vkcolorcubeshader.vert vkcolorcubeshader.frag")
+                       COMMENT "Compiling Vulkan Shaders vkcolorcubeshader.vert vkcolorcubeshader.frag vkcolorcubewireframeshader.frag")
 else()
     add_custom_command(TARGET ${GPA_PROJECT_NAME} PRE_BUILD
                        COMMAND ${CMAKE_CURRENT_LIST_DIR}/CompileShaders.sh ${GPA_OUTPUT_DIR}/$<CONFIG>/ ${VK_TOOLS_DIR}
                        BYPRODUCTS vkcolorcubeshader.vert.spv vkcolorcubeshader.frag.spv
                        WORKING_DIRECTORY ${CMAKE_CURRENT_LIST_DIR}
-                       COMMENT "Compiling Vulkan Shaders vkcolorcubeshader.vert vkcolorcubeshader.frag")
+                       COMMENT "Compiling Vulkan Shaders vkcolorcubeshader.vert vkcolorcubeshader.frag vkcolorcubewireframeshader.frag")
 endif()
 
 set_target_properties(${GPA_PROJECT_NAME} PROPERTIES FOLDER "Examples/Vulkan")
diff --git a/Src/Examples/Vulkan/VkColorCube/vkcolorcube.cpp b/Src/Examples/Vulkan/VkColorCube/vkcolorcube.cpp
@@ -158,7 +158,8 @@ AMDVulkanDemo::AMDVulkanDemo()
 #endif
     };
 
-    m_requiredDeviceExtensions = {VK_KHR_SWAPCHAIN_EXTENSION_NAME, AMD_GPA_REQUIRED_DEVICE_EXTENSION_NAME_LIST};
+    m_requiredDeviceExtensions = { VK_KHR_SWAPCHAIN_EXTENSION_NAME, AMD_GPA_REQUIRED_DEVICE_EXTENSION_NAME_LIST };
+    m_optionalDeviceExtensions = { VK_KHR_SWAPCHAIN_EXTENSION_NAME, AMD_GPA_REQUIRED_DEVICE_EXTENSION_NAME_LIST, AMD_GPA_OPTIONAL_DEVICE_EXTENSION_NAME_LIST };
 }
 
 AMDVulkanDemo::~AMDVulkanDemo()
@@ -773,12 +774,21 @@ bool AMDVulkanDemo::InitializeVulkan()
     deviceCreateInfo.pQueueCreateInfos       = &deviceQueueCreateInfo;
     deviceCreateInfo.enabledLayerCount       = 0;
     deviceCreateInfo.ppEnabledLayerNames     = nullptr;
-    deviceCreateInfo.enabledExtensionCount   = static_cast<uint32_t>(m_requiredDeviceExtensions.size());
-    deviceCreateInfo.ppEnabledExtensionNames = m_requiredDeviceExtensions.data();
+    deviceCreateInfo.enabledExtensionCount   = static_cast<uint32_t>(m_optionalDeviceExtensions.size());
+    deviceCreateInfo.ppEnabledExtensionNames = m_optionalDeviceExtensions.data();
     deviceCreateInfo.pEnabledFeatures        = &m_vkPhysicalDeviceFeatures;
 
     VkResult resultCreateDevice = vkCreateDevice(m_vkPhysicalDevice, &deviceCreateInfo, nullptr, &m_vkDevice);
 
+    if (resultCreateDevice == VK_ERROR_EXTENSION_NOT_PRESENT)
+    {
+        // try again with only required extensions
+        deviceCreateInfo.enabledExtensionCount = static_cast<uint32_t>(m_requiredDeviceExtensions.size());
+        deviceCreateInfo.ppEnabledExtensionNames = m_requiredDeviceExtensions.data();
+
+        resultCreateDevice = vkCreateDevice(m_vkPhysicalDevice, &deviceCreateInfo, nullptr, &m_vkDevice);
+    }
+
     if (resultCreateDevice != VK_SUCCESS)
     {
         std::cout << "ERROR: Failed to create device\n." << std::endl;
diff --git a/Src/Examples/Vulkan/VkColorCube/vkcolorcube.h b/Src/Examples/Vulkan/VkColorCube/vkcolorcube.h
@@ -159,6 +159,9 @@ class AMDVulkanDemo
     /// List of device extension names that are required by this demo.
     std::vector<const char*> m_requiredDeviceExtensions;
 
+    /// List of device extension names that are optional for this demo.
+    std::vector<const char*> m_optionalDeviceExtensions;
+
 #if defined(VK_USE_PLATFORM_WIN32_KHR)
     /// Application windowing instance
     HINSTANCE m_hInstance;
diff --git a/Src/GPUPerfAPI-Common/GPAVersion.h b/Src/GPUPerfAPI-Common/GPAVersion.h
@@ -11,7 +11,7 @@
 // Make sure to increment GPA_FUNCTION_TABLE_MAJOR_VERSION_NUMBER anytime GPA_MAJOR_VERSION is incremented
 #define GPA_MAJOR_VERSION 3   ///< Macro for GPA major version
 #define GPA_MINOR_VERSION 4   ///< Macro for GPA minor version
-#define GPA_UPDATE_VERSION 0  ///< Macro for GPA update version
+#define GPA_UPDATE_VERSION 1  ///< Macro for GPA update version
 
 /// Macro for GPA build number
 #define GPA_BUILD_NUMBER 0
diff --git a/Src/GPUPerfAPI-Common/GPUPerfAPI-VK.h b/Src/GPUPerfAPI-Common/GPUPerfAPI-VK.h
@@ -17,12 +17,18 @@
 /// Define the AMD shader core properties extension name
 #define VK_AMD_SHADER_CORE_PROPERTIES_EXTENSION_NAME "VK_AMD_shader_core_properties"
 
+/// Define the AMD shader core properties 2 extension name
+#define VK_AMD_SHADER_CORE_PROPERTIES_EXTENSION_NAME2 "VK_AMD_shader_core_properties2"
+
 /// Define a macro to help developers include all instance-level extensions required to support the AMD GPA Interface.
 #define AMD_GPA_REQUIRED_INSTANCE_EXTENSION_NAME_LIST VK_KHR_GET_PHYSICAL_DEVICE_PROPERTIES_2_EXTENSION_NAME
 
 /// Define a macro to help developers include all device-level extensions required to support the AMD GPA Interface.
 #define AMD_GPA_REQUIRED_DEVICE_EXTENSION_NAME_LIST VK_AMD_GPA_INTERFACE_EXTENSION_NAME, VK_AMD_SHADER_CORE_PROPERTIES_EXTENSION_NAME
 
+/// Define a macro to help developers include optional device-level extensions to support the AMD GPA Interface.
+#define AMD_GPA_OPTIONAL_DEVICE_EXTENSION_NAME_LIST VK_AMD_SHADER_CORE_PROPERTIES_EXTENSION_NAME2
+
 /// The struct that should be supplied to GPA_OpenContext().
 /// The instance, physicalDevice, and device should be set prior to
 /// calling OpenContext() to reflect the Vulkan objects on which profiling
diff --git a/Src/GPUPerfAPI-Common/GPUPerfAPITypes.h b/Src/GPUPerfAPI-Common/GPUPerfAPITypes.h
@@ -194,7 +194,8 @@ typedef enum
 {
     GPA_OPENCONTEXT_DEFAULT_BIT =
         0,  ///< Open contexts using all default options (all counters exposed, clocks are set to stable frequencies which are known to be power and thermal sustainable. The ratio between the engine and memory clock frequencies will be kept the same as much as possible)
-    GPA_OPENCONTEXT_HIDE_PUBLIC_COUNTERS_BIT   = 0x01,  ///< Prevent the public counters from being exposed
+    GPA_OPENCONTEXT_HIDE_DERIVED_COUNTERS_BIT  = 0x01,                                       ///< Prevent the derived counters from being exposed
+    GPA_OPENCONTEXT_HIDE_PUBLIC_COUNTERS_BIT   = GPA_OPENCONTEXT_HIDE_DERIVED_COUNTERS_BIT,  ///< For backwards compatibility
     GPA_OPENCONTEXT_HIDE_SOFTWARE_COUNTERS_BIT = 0x02,  ///< Prevent the software counters from being exposed
     GPA_OPENCONTEXT_HIDE_HARDWARE_COUNTERS_BIT = 0x04,  ///< Prevent the hardware counters from being exposed
     GPA_OPENCONTEXT_CLOCK_MODE_NONE_BIT =
diff --git a/Src/GPUPerfAPI-Common/IGPAContext.h b/Src/GPUPerfAPI-Common/IGPAContext.h
@@ -179,6 +179,11 @@ class IGPAContext : public IGPAInterfaceTrait
     /// Returns the active session for the context
     /// \return active session if any session is active otherwise nullptr
     virtual const IGPASession* GetActiveSession() const = 0;
+
+    /// Enable/disable the stable power state, using the stable clock mode specified when opening the context
+    /// \param[in] useProfilingClocks true to use GPU clocks for profiling, false to use default clock mode
+    /// \return GPA_STATUS_OK on success
+    virtual GPA_Status SetStableClocks(bool useProfilingClocks) = 0;
 };
 
 #endif  // _I_GPA_CONTEXT_H_
diff --git a/Src/GPUPerfAPICL/CLGPAContext.h b/Src/GPUPerfAPICL/CLGPAContext.h
@@ -52,10 +52,8 @@ class CLGPAContext : public GPAContext
     /// \return the CL command queue
     const cl_command_queue& GetCLCommandQueue() const;
 
-    /// Enable/disable the stable power state, using the stable clock mode specified when opening the context
-    /// \param[in] useProfilingClocks true to use GPU clocks for profiling, false to use default clock mode
-    /// \return GPA_STATUS_OK on success
-    GPA_Status SetStableClocks(bool useProfilingClocks);
+    /// \copydoc IGPAContext::SetStableClocks()
+    GPA_Status SetStableClocks(bool useProfilingClocks) override;
 
 private:
     /// Parses a version number returned from clGetDeviceInfo(CL_DRIVER_VERSION)
diff --git a/Src/GPUPerfAPIDX11/CMakeLists.txt b/Src/GPUPerfAPIDX11/CMakeLists.txt
@@ -26,7 +26,9 @@ set(AMD_DX_EXT_HEADERS ${AMD_DX_EXT_DIR}/AmdDxExt.h
                        ${AMD_DX_EXT_DIR}/AmdDxExtApi.h
                        ${AMD_DX_EXT_DIR}/AmdDxExtIface.h
                        ${AMD_DX_EXT_DIR}/AmdDxExtPerfProfile.h
-                       ${AMD_DX_EXT_DIR}/AmdDxExtPerfProfileApi.h)
+                       ${AMD_DX_EXT_DIR}/AmdDxExtPerfProfileApi.h
+                       ${AMD_DX_EXT_DIR}/AmdDxExtASICInfo.h
+                       ${AMD_DX_EXT_DIR}/AmdDxExtASICInfoApi.h)
 
 set(HEADER_FILES ${AMD_DX_EXT_HEADERS}
                  ${GPA_SRC_COMMON}/DXGetAMDDeviceInfo.h
@@ -63,7 +65,7 @@ target_link_libraries(${GPA_PROJECT_NAME}
                       GPUPerfAPICounterGenerator)
 
 if(${GPA_ALL_OPEN_SOURCE})
-    set(GPA_DX11_GET_DEVICE_INFO ${COMMON_DIR_LIB}/AMD/GPUPerfAPI/3_2/Bin/${AMDTPlatform}/GPUPerfAPIDXGetAMDDeviceInfo)
+    set(GPA_DX11_GET_DEVICE_INFO ${COMMON_DIR_LIB}/AMD/GPUPerfAPI/3_4/Bin/${AMDTPlatform}/GPUPerfAPIDXGetAMDDeviceInfo)
     add_custom_command(TARGET ${GPA_PROJECT_NAME} POST_BUILD
                        COMMAND ${CMAKE_COMMAND} -E copy ${GPA_DX11_GET_DEVICE_INFO}${AMDTPlatformSuffix}.dll ${CMAKE_CURRENT_LIST_DIR}/${DEPTH}/Output/$<CONFIG> VERBATIM
                        COMMENT "Copying GPUPerfAPIDXGetDeviceInfo to output folder...")
diff --git a/Src/GPUPerfAPIDX11/DX11GPAContext.h b/Src/GPUPerfAPIDX11/DX11GPAContext.h
@@ -80,10 +80,8 @@ class DX11GPAContext : public GPAContext
     /// \return the max event id of the specified block. Could be zero if block does not exist
     gpa_uint32 GetMaxEventIdCount(PE_BLOCK_ID block) const;
 
-    /// Enable/disable the stable power state, using the stable clock mode specified when opening the context
-    /// \param[in] useProfilingClocks true to use GPU clocks for profiling, false to use default clock mode
-    /// \return GPA_STATUS_OK on success
-    GPA_Status SetStableClocks(bool useProfilingClocks);
+    /// \copydoc IGPAContext::SetStableClocks()
+    GPA_Status SetStableClocks(bool useProfilingClocks) override;
 
 private:
     /// Initializes the AMD extensions
diff --git a/Src/GPUPerfAPIDX11/DX11GPAImplementor.cpp b/Src/GPUPerfAPIDX11/DX11GPAImplementor.cpp
@@ -16,8 +16,7 @@
 #include "DXGetAMDDeviceInfo.h"
 #include "GPACustomHWValidationManager.h"
 #include "GPACommonDefs.h"
-
-#include <AmdDxExtPerfProfileApi.h>
+#include "DX11Include.h"
 
 IGPAImplementor* s_pGpaImp = DX11GPAImplementor::Instance();
 
@@ -268,16 +267,6 @@ bool DX11GPAImplementor::GetAmdHwInfo(ID3D11Device* pD3D11Device,
 
                 hwInfo.SetGpuIndex(static_cast<unsigned int>(gpuIndex));
 
-                if (nullptr != pExtPerfProfile)
-                {
-                    pExtPerfProfile->Release();
-                }
-
-                if (nullptr != pExt)
-                {
-                    pExt->Release();
-                }
-
                 std::string strDLLName;
 
                 if (GPAUtil::GetCurrentModulePath(strDLLName))
@@ -361,6 +350,50 @@ bool DX11GPAImplementor::GetAmdHwInfo(ID3D11Device* pD3D11Device,
                                 hwInfo.SetTimeStampFrequency(deviceFrequency);
                             }
 
+                            unsigned int   majorVer = 0;
+                            unsigned int   minorVer = 0;
+                            unsigned int   subMinorVer = 0;
+                            ADLUtil_Result adlResult = AMDTADLUtils::Instance()->GetDriverVersion(majorVer, minorVer, subMinorVer);
+
+                            static const unsigned int MIN_MAJOR_VER = 19;
+                            static const unsigned int MIN_MINOR_VER_FOR_30 = 30;
+
+                            if ((ADL_SUCCESS == adlResult || ADL_WARNING == adlResult))
+                            {
+                                if (majorVer >= MIN_MAJOR_VER && minorVer >= MIN_MINOR_VER_FOR_30)
+                                {
+                                    if (nullptr != pExt)
+                                    {
+                                        IAmdDxExtASICInfo* pExtAsicInfo = reinterpret_cast<IAmdDxExtASICInfo*>(pExt->GetExtInterface(AmdDxExtASICInfoID));
+
+                                        if (nullptr != pExtAsicInfo)
+                                        {
+                                            AmdDxASICInfoParam infoParam = {};
+                                            AmdDxASICInfo*     pNewAsicInfo = new (std::nothrow) AmdDxASICInfo();
+
+                                            if (nullptr != pNewAsicInfo)
+                                            {
+                                                infoParam.pASICInfo = pNewAsicInfo;
+                                                pExtAsicInfo->GetInfoData(&infoParam);
+
+                                                if (nullptr != infoParam.pASICInfo && gpuIndex < infoParam.pASICInfo->gpuCount)
+                                                {
+                                                    AmdDxASICInfoHWInfo asicInfo = infoParam.pASICInfo->hwInfo[gpuIndex];
+                                                    hwInfo.SetNumberCUs(asicInfo.totalCU);
+                                                    hwInfo.SetNumberShaderEngines(asicInfo.numShaderEngines);
+                                                    hwInfo.SetNumberShaderArrays(asicInfo.numShaderArraysPerSE);
+                                                    hwInfo.SetNumberSIMDs(asicInfo.totalCU * asicInfo.numSimdsPerCU);
+                                                }
+
+                                                delete pNewAsicInfo;
+                                            }
+
+                                            pExtAsicInfo->Release();
+                                        }
+                                    }
+                                }
+                            }
+
                             success = true;
                             break;
                         }
@@ -370,6 +403,16 @@ bool DX11GPAImplementor::GetAmdHwInfo(ID3D11Device* pD3D11Device,
                 {
                     GPA_LogError("Unable to get the module path.");
                 }
+
+                if (nullptr != pExtPerfProfile)
+                {
+                    pExtPerfProfile->Release();
+                }
+
+                if (nullptr != pExt)
+                {
+                    pExt->Release();
+                }
             }
             else
             {
diff --git a/Src/GPUPerfAPIDX11/DX11Include.h b/Src/GPUPerfAPIDX11/DX11Include.h
@@ -11,6 +11,8 @@
 #include <d3d11.h>
 #include <d3dcommon.h>
 #include "GPACommonDefs.h"
+#include "AmdDxExtASICInfo.h"
+#include "AmdDxExtASICInfoApi.h"
 #include "AmdDxExtPerfProfileApi.h"
 
 #define PE_BLOCK_ID_MAX PE_BLOCK_MAX_GFX10  /// macro for count of DX11 block id, notably we need to update this with every next generation
diff --git a/Src/GPUPerfAPIDX12/DX12GPAContext.cpp b/Src/GPUPerfAPIDX12/DX12GPAContext.cpp
@@ -229,8 +229,6 @@ void DX12GPAContext::CleanUp()
 {
     std::lock_guard<std::mutex> lockContextResources(m_dx12GpaContextMutex);
 
-    SetStableClocks(false);
-
     // Release Device
     if (nullptr != m_pD3D12Device)
     {
@@ -267,7 +265,7 @@ void DX12GPAContext::CleanUp()
     }
 }
 
-void DX12GPAContext::SetStableClocks(bool useProfilingClocks)
+GPA_Status DX12GPAContext::SetStableClocks(bool useProfilingClocks)
 {
     if (nullptr != m_pGpaInterface)
     {
@@ -311,4 +309,6 @@ void DX12GPAContext::SetStableClocks(bool useProfilingClocks)
             m_pGpaInterface->SetClockMode(m_clockMode, nullptr);
         }
     }
+
+    return GPA_STATUS_OK;
 }
diff --git a/Src/GPUPerfAPIDX12/DX12GPAContext.h b/Src/GPUPerfAPIDX12/DX12GPAContext.h
@@ -67,17 +67,16 @@ class GPA_THREAD_SAFE_OBJECT DX12GPAContext : public GPAContext
     /// \return ID3D12Device pointer
     ID3D12Device* GetD3D12Device() const;
 
-    /// Destroys all the allocated resources for the context
-    void CleanUp();
+    /// \copydoc IGPAContext::SetStableClocks()
+    GPA_Status SetStableClocks(bool useProfilingClocks) override;
 
 private:
     /// Initializes the AMD Driver extension objects for the context
     /// \return true upon successful initialization otherwise false
     bool InitializeAMDExtension();
 
-    /// Enable/disable the stable power state, using the stable clock mode specified when opening the context
-    /// \param[in] useProfilingClocks true to use GPU clocks for profiling, false to use default clock mode
-    void SetStableClocks(bool useProfilingClocks);
+    /// Destroys all the allocated resources for the context
+    void CleanUp();
 
     mutable std::mutex             m_dx12GpaContextMutex;      ///< Mutex for DX12 GPA Context
     ID3D12Device*                  m_pD3D12Device;             ///< D3D12Device pointer
diff --git a/Src/GPUPerfAPIDX12/DX12GPAImplementor.cpp b/Src/GPUPerfAPIDX12/DX12GPAImplementor.cpp
@@ -166,7 +166,12 @@ bool DX12GPAImplementor::CloseAPIContext(GPADeviceIdentifier pDeviceIdentifier,
     if (nullptr != pContext)
     {
         DX12GPAContext* pDx12GpaContext = reinterpret_cast<DX12GPAContext*>(pContext);
-        pDx12GpaContext->CleanUp();
+        /*
+         * Deleting the context resources at this point is causing
+         * some issue in driver as some of the resources are still in use on factory object
+         * We will defer the release of driver extension resources at time of destroying the GPA
+         */
+        pDx12GpaContext->SetStableClocks(false);
         m_dx12GpaContextList.push_back(pDx12GpaContext);
     }
 
diff --git a/Src/GPUPerfAPIGL/GLGPAContext.h b/Src/GPUPerfAPIGL/GLGPAContext.h
diff --git a/Src/GPUPerfAPIROCm/ROCmGPAContext.cpp b/Src/GPUPerfAPIROCm/ROCmGPAContext.cpp
diff --git a/Src/GPUPerfAPIROCm/ROCmGPAContext.h b/Src/GPUPerfAPIROCm/ROCmGPAContext.h
diff --git a/Src/GPUPerfAPIVk/VkGPAContext.h b/Src/GPUPerfAPIVk/VkGPAContext.h
diff --git a/Src/GPUPerfAPIVk/VkGPAImplementor.cpp b/Src/GPUPerfAPIVk/VkGPAImplementor.cpp
diff --git a/Src/GPUPerfAPIVk/VkIncludes.h b/Src/GPUPerfAPIVk/VkIncludes.h
diff --git a/docs/source/conf.py b/docs/source/conf.py
diff --git a/docs/source/usage.rst b/docs/source/usage.rst
