v0.9.0

v0.9.0

Full Changelog

Merged pull requests:

• Readded missing codecov cmake targets #345 (axsaucedo)
• Fix example build #340 (ToKiNoBug)
• Fix find_package, enable CPack and support clang-cl #338 (ToKiNoBug)
• Fix cmake install2 (with sign off) #337 (ToKiNoBug)
• [performance bug] Fix Sequence::clear() #328 (Crydsch)
• Fix vulkan version parsing #325 (Crydsch)
• Adding storage tensor functionality with tests #316 (axsaucedo)
• Logging fixes for spdlog #315 (axsaucedo)
• Adding CI setup to ensure examples are tested automatically #314 (axsaucedo)
• Updating default headers to Vulkan v1.3.231 #312 (axsaucedo)
• Fixed build error on gcc 12 #311 (hungrymonkey)
• Upgrading to Vulkan 1.3 #309 (axsaucedo)
• Updated the fetch_content git repo and commit hash #307 (COM8)
• Fixes two mistakes in comments and removes one unused vk::MappedMemoryRange #301 (MiroPalmu)
• Added an option to get the current vk instance #292 (COM8)
• fix(CMakeLists.txt): typo #291 (tpoisonooo)
• Fixed compiler warnings #289 (COM8)
• Refactor build system #287 (COM8)
• Added .clang-format file and formatted everything #286 (COM8)
• Enum class fix for fmt 8.x #285 (COM8)
• Fixed Vulkan deprecation warning #284 (COM8)
• Removed '#pragma once' from cpp files #283 (COM8)

What's Changed

• Fixed Vulkan deprecation warning by @COM8 in #284
• Removed '#pragma once' from cpp files by @COM8 in #283
• Enum class fix for fmt 8.x by @COM8 in #285
• Added .clang-format file and formatted everything by @COM8 in #286
• Fixed compiler warnings by @COM8 in #289
• fix(CMakeLists.txt): typo by @tpoisonooo in #291
• Added an option to get the current vk instance by @COM8 in #292
• Fixes two mistakes in comments and removes one unused vk::MappedMemoryRange by @MiroPalmu in #301
• Refactor build system by @COM8 in #287
• Updated the fetch_content git repo and commit hash by @COM8 in #307
• Upgrading to Vulkan 1.3 by @axsaucedo in #309
• Fixed build error on gcc 12 by @hungrymonkey in #311
• Updating default headers to Vulkan v1.3.231 by @axsaucedo in #312
• Adding CI setup to ensure examples are tested automatically by @axsaucedo in #314
• Logging fixes for spdlog by @axsaucedo in #315
• Adding storage tensor functionality with tests by @axsaucedo in #316
• Fix vulkan version parsing by @Crydsch in #325
• [performance bug] Fix Sequence::clear() by @Crydsch in #328
• Fix cmake install2 (with sign off) by @ToKiNoBug in #337
• Fix find_package, enable CPack and support clang-cl by @ToKiNoBug in #338
• Fix example build by @ToKiNoBug in #340
• Readded missing codecov cmake targets by @axsaucedo in #345

New Contributors

• @COM8 made their first contribution in #284
• @MiroPalmu made their first contribution in #301
• @hungrymonkey made their first contribution in #311
• @Crydsch made their first contribution in #325
• @ToKiNoBug made their first contribution in #337

Closed issues:

• State of the Project #344
• Does it support Mali GPU? Like rk35xx series board? #330
• After compilation, only the library file has no header file #324
• windows building error #321
• How can I know to which extent a particular GPU is supported? #319
• windows building error #318
• Array multiplication example code needs to be updated everytime kompute have a build issue. #313
• gcc12 build fails because std:shared_ptr requires explicit declation of <memory> #310
• Android sample is not compiling #305
• read/write operations from disk from GPU/device arrays ? #299
• memory sanitizer reports errors for simple example #298
• Build failure: static assertion failed: Cannot format an argument. To make type T formattable provide a formatter<T> specialization: https://fmt.dev/latest/api.html\#udt #282
• Build cleanly without warnings #278
• Upgrade to version 1.3 of Vulkan #265

Full Changelog: v0.8.1...v0.9.0

v0.8.1

v0.8.1

Full Changelog

Closed issues:

• Discord link in README and docs is broken #276
• Website examples typo's and 6500 XT unknown GPU #275
• [Question] How to disable all log ? #274
• full diagram 404 #271
• Error when enabling KOMPUTE\_ENABLE\_SPDLOG #268
• Add KOMPUTE_LOG_ACTIVE_LEVEL instead of current SPDLOG_ACTIVE_LEVEL #267
• Update/Fix Android project #264
• Update compileSource function in examples/docs to correct one #261
• Technically can Kompute be modified to support data visualization? #260
• Data-transfer for Integrated GPU #258
• Python "getting started" example fails #252
• Python example in README doesn't work #248
• Running Android app #234

Merged pull requests:

• Added active log level definitions for kompute #280 (axsaucedo)
• Fix TestDestroy.TestDestroyTensorSingle #279 (ScheissSchiesser)
• Updated discord link #277 (axsaucedo)
• style(src/Algorithm): fix typo #273 (tpoisonooo)
• Fix Android Example confirmed with blog post steps #266 (axsaucedo)
• Adding Governance with TSC charter #263 (axsaucedo)
• Updating array_mutiplication example to work correctly #262 (axsaucedo)
• Updated formatting #257 (axsaucedo)
• Fix first two python examples in the docs #256 (lopuhin)
• Remove nonexisting "single_include" from INSTALL_INTERFACE #254 (ItsBasi)
• Added community page #253 (axsaucedo)
• Updated readme to reflect shader utils #249 (axsaucedo)
• Avoid using pointers to temporary copies of desired extensions. #247 (ItsBasi)

v0.8.0

v0.8.0

Full Changelog

Closed issues:

• An unset KOMPUTE_ENV_DEBUG_LAYERS leads KP_LOG_DEBUG to pass envLayerNamesVal==nullptr #245
• Extend utils shader helpers in test for windows #240
• Python segfaults after import kp #230
• Simple and extended python examples do not work (v 0.7.0) #228
• Python macOS issue (ImportError: dlopen(...): no suitable image found. Did find: ...: mach-o, but wrong architecture) #223
• Python macOS issue (Symbol not found: __PyThreadState_Current ... Expected in: flat namespace) #221
• Finalise Migration of Kompute into Linux Foundation #216
• CMake Error: Imported target "kompute::kompute" includes non-existent path "/usr/local/single_include" #212
• Incompatibality inroduced with #168 on Vulkan 1.1.x #209
• external libraries #201
• Starting slack group or discord for alternative / faster version of asking questions #198
• Test SingleSequenceRecord is not thread safe and fails in AMD card #196
• Update Kompute headers to reference the glslang headers for install vs build interfaces #193
• Integrate with GLSLang find_package file when issue is resolved in the glslang repo #191
• Release 0.7.0 #187
• Get number of available devices #185
• Deep Learning Convolutional Neural Network (CNN) example implementation #162
• Create example compiling and running in raspberry pi with Mesa Vulkan drivers #131
• Add support for VK_EXT_debug_utils labels #110

Merged pull requests:

• Fix for null debug log causing exception in fmt lib #246 (axsaucedo)
• 0.8.0 Release #244 (axsaucedo)
• Adding support for different types for spec and push consts #242 (axsaucedo)
• Extend shader helper functions in tests to support windows #241 (axsaucedo)
• Increase test cov across codebase #239 (axsaucedo)
• Updated collab link for C++ notebook #237 (axsaucedo)
• Updating repo licenses and links #236 (axsaucedo)
• Removing GLSLang as core dependency #235 (axsaucedo)
• Naive matrice multiplication example #233 (Corentin-pro)
• Fixed typo in CMakeLists.txt (ANDOID => ANDROID) #232 (Corentin-pro)
• Set kp_debug, kp_info, kp_warning and kp_error to py::none() when the program terminates. #231 (thinking-tower)
• VGG7 Python example #227 (20kdc)
• Add documentation for CMake flags #224 (thinking-tower)
• Set PYTHON_INCLUDE_DIR and PYTHON_LIBRARY during installation #222 (thinking-tower)
• Removing xxd.exe binary and add instructions to build #220 (axsaucedo)
• [PYTHON] Ensure numpy array increments refcount of tensor to keep valid #219 (axsaucedo)
• Added destroy for manager #218 (axsaucedo)
• Revert "Fixed the issue that caused CMake to look for non-existent path after being installed" #217 (axsaucedo)
• Fixed the issue that caused CMake to look for non-existent path after being installed #213 (unexploredtest)
• omitted .data() because it is incompatible with vulkan 1.1.x #211 (unexploredtest)
• vkEnumeratePhysicalDevices(*(this->mInstance) ... doesn't work on Linux i386 #208 (unexploredtest)
• Raises an error when having no/exceeding vulkan device's limit #207 (unexploredtest)
• Updated README and fixed a syntax error on C++'s example #206 (unexploredtest)
• removed the extra comma after KOMPUTE_OPT_REPO_SUBMODULE_BUILD #205 (unexploredtest)
• Extending list_devices test for multiple devices #204 (axsaucedo)
• Fix #include <SPIRV/GlslangToSpv.h> #200 (unexploredtest)
• Added memory barrier on test #199 (axsaucedo)
• Add function to list physical devices #195 (axsaucedo)
• v0.7.0 release #189 (axsaucedo)
• Add instructions for running on Pi4 #180 (hpgmiskin)

v0.7.0

Release v0.7.0

The 0.7.0 release introduces a very extensive list of features - a high level overview includes:

• Support for push constants
• Support for specailisation constants
• Support for tensor data types bool, float, double, int32 and uint32
• Ability to define Operations outside manager
• Ability to create Algorithm outside manager
• New OpMemoryBarrier to add custom barriers
• New OpAlgoDispatch to dispatch algorithm with push constants
• New interface for sequences
• New memory relationships all managed by top level manager with weak references allowing for smart pointers to terminate objects
• Code coverage metrics using gcov + lcov

Implemented enhancements:

• Extend non-spdlog print functions to use std::format #158
• Add code coverage reports with codecov #145
• Explore removing std::vector mData; completely from Tensor in favour of always storing data in hostVisible buffer memory (TBC) #144
• Update all examples to match breaking changes in 0.7.0 #141
• Avoid copy when returning python numpy / array #139
• Cover all Python & C++ tests in CI #121
• Add C++ Test for Simple Work Groups Example #117
• Expose push constants in OpAlgo #54
• Expose ability to create barriers in OpTensor operations #45
• Create delete function in manager to free / destroy sequence #36
• Make specialisation data extensible #12
• Support multiple types for Kompute Tensors #2
• Added re-record sequence functionality and updated docs #171 (axsaucedo)
• Extend non-spdlog print functions to use fmt::format / fmt::print #159 (axsaucedo)
• Added support for custom SpecializedConstants and removed KomputeWorkgroup class #151 (axsaucedo)
• Added destroy functions for tensors and sequences (named and object) #146 (axsaucedo)

Fixed bugs:

• push_constant not working in my case? #168
• DescriptorPool set is not being freed #155
• Updated memory barriers to include staging buffers #182 (axsaucedo)
• Adds push const ranges in pipelinelayout to fix #168 #174 (axsaucedo)
• Added destructor for staging tensors #134 (axsaucedo)

Closed issues:

• Update memory barriers to align with tensor staging/primary memory revamp #181
• Move shader defaultResource inside kp::Shader class #175
• Reach at least 90% code coverage on tests #170
• Add functionality to re-record sequence as now it's possible to update the underlying algorithm #169
• Use numpy arrays as default return value #166
• Update all shared_ptr value passes to be by ref or const ref #161
• Amend memory hierarchy for kp::Operations so they can be created separately #160
• Customise theme of documentation #156
• Remove KomputeWorkgroup class in favour of std::array<uint32_t, 3> #152
• Passing raw GLSL string to Shader Module depricated so remove this method from supported approach #150
• Add python backwards compatibility for eval_tensor_create_def #147
• Document breaking changes for 0.7.0 #140
• Tensor memory management and memory hierarchy redesign #136
• Staging tensor GPU memory is not freed as part of OpCreateTensor removal #133
• eStorage Tensors are currently unusable as OpTensorCreate calls mapDataIntoHostMemory #132
• 0.6.0 Release #126
• java.lang.UnsatisfiedLinkError: dlopen failed: library "libkompute-jni.so" not found #125
• Initial exploration: Include explicit GLSL to SPIRV compilation #107
• Add support for push constants #106

Merged pull requests:

• Resolve moving all functions from tensor HPP to CPP #186 (axsaucedo)
• Device Properties #184 (alexander-g)
• Too many warnings #183 (alexander-g)
• Add support for bool, double, int32, uint32 and float32 on Tensors via TensorT #177 (axsaucedo)
• Support for Timestamping #176 (alexander-g)
• Test for ShaderResources #165 (aliPMPAINT)
• Amend memory hierarchy to enable for push constants and functional interface for more flexible operations #164 (axsaucedo)
• made changes for include paths for complete installation #163 (aliPMPAINT)
• Added dark mode on docs #157 (axsaucedo)
• Glslang implementation for online shader compilation #154 (axsaucedo)
• Adding test code coverage using gcov and lcov #149 (axsaucedo)
• Added temporary backwards compatibility for eval_tensor_create_def function #148 (axsaucedo)
• Amend memory ownership hierarchy to have Tensor owned by Manager instead of OpCreateTensor / OpBase #138 (axsaucedo)
• Removed Staging Tensors in favour of having two buffer & memory in a Tensor to minimise data transfer #137 (axsaucedo)

v0.6.0

v0.6.0

Full Changelog

Implemented enhancements:

• Add simple test for Python log\_level function #120
• Add further numpy support #104
• SWIG syntax error - change order of keywords. #94
• Create mocks to isolate unit tests for components #8
• Disallowing zero sized tensors #129 (alexander-g)
• Added further tests to CI and provide Dockerimage with builds to swiftshader #119 (axsaucedo)
• Workgroups for Python #116 (alexander-g)
• Ubuntu CI #115 (alexander-g)
• Faster set_data() #109 (alexander-g)
• String parameter for eval_algo_str methods in Python #105 (alexander-g)
• Added numpy() method #103 (alexander-g)

Fixed bugs:

• [PYTHON] Support string parameter instead of list for eval_algo_data when passing raw shader as string #93
• [PYTHON] Fix log_level on the python implementation (using pybind's logging functions) #92

Closed issues:

• Add documentation for custom operations #128
• Numpy Array Support and Work Group Configuration in Python Kompute #124
• Remove references to spdlog in python module #122
• Setup automated CI testing for PRs using GitHub actions #114
• Python example type error (pyshader). #111
• Update all references to operations to not use template #101
• Getting a undefined reference error while creating a Kompute Manager #100

Merged pull requests:

• 122 remove spdlog references in python #123 (axsaucedo)
• Native logging for Python #118 (alexander-g)
• Fixes for the c++ Simple and Extended examples in readme #108 (aliPMPAINT)
• Fix building shaders on native linux #102 (aliPMPAINT)

Release v0.5.1

• Fixed error with Kompute-go SWIG integration #94
• Added python list operations for Tensor (setitem, len, etc)
• Added new python blog post

Release v0.5.0

• Migrated all OpAlgoBase components to use dynamic layouts instead of templates #26, #57
• Updated all examples to use spir-v bytes by default #86
• Added compatibility for Vulkan Headers HPP v1.2.154+ #84
• Added Python Pypi package for Kompute #87
• Added python interface functions to process python spirv bytes directly* Added implementation of Logistic Regression implementation in Python
• Extended examples to showcase pyshader to use more pythonic GPU development
• Enabled spdlog builds by default on python package
• Added multi-platform python package installs via pypi https://pypi.org/project/kp/
• Added log level config functionality in python
• Added Python Bindings for Kompute library
• Added Python tests showcasing core functionality using Manager and Sequences
• Added documentation integrating pyhton class references (to be automated)
• Changed sequences to be returned as shader_ptr instead of weak_ptr
• Added sequence memory management via init member function
• Added explicit definition on VulkanDestroy funtions for VulkanHPP 1.2.155-1.2.158 compat
• Removed template parameters from OpAlgoBase functions (added Op*.cpp files)
• Added python build to main cmake file
• Added pybind11 submodule
• Added Sequence tests to verify memory management via init member function
• Update e2e examples
• Add Python documentation and further examples

Release v0.4.1

• Support for Vulkan HPP version 1.2.155 and above (tested with every version until 1.2.158)
• Updated linux docker image to test multiple versions of Vulkan SDK
• CCLS support for submodule builds (besides vcpkg builds)
• Submodules for core dependencies for flexibility when testing dependencies with particular versions
• Build flags to configure submodule builds vs vcpkg (toolchain based) builds
• Removed range prints which removes explicit dependency on fmt
• Syntax correction of private/protected member variables
• AUR package added by contributor https://aur.archlinux.org/packages/vulkan-kompute-git/ via #81

Release v0.4.0

• Added async/await capabilities for manager and sequence
• Added fence resource as member object for Sequence
• Ensure sequence begin() function clears previous operations
• Ensure begin does not get called if sequence in running state
• Ensure manager creates new sequence when default functions called
• Added capabilities for multi queue support in manager with explicit allocation on sequences
• Fixed compile warnings on Linux (ubuntu)
• Added createManagedSequence on manager with ability for default sequence name to be created
• Added tests for asynchronous and parallel processing
• Added LogisticRegression shader as cpp header
• Added documentation on advanced examples
• Added documentation on shader-to-cpp-header scripts
• Added CNAME for kompute.cc domain

Release v0.3.2

• Enabled dynamic loading of Vulkan library for Vulkan support
• Added Android NDK Wrapper with Vulkan HPP support
• Updated CMAKE min standard to C++14
• Downgraded min cmake requirement to 3.4.1
• Added new shader file as shaderlogisticregression as part of AggregateHeader
• Added #pragma once guard to Kompte.hpp single header
• Updated createComputePipeline to return Result instead of ResultValue for backwards compatibility
• Added new compute flags for android including:
  • KOMPUTE_OPT_INSTALL
  • KOMPUTE_OPT_ANDROID_BUILD
  • KOMPUTE_OPT_DISABLE_VK_DEBUG_LAYERS
  • KOMPUTE_VK_API_VERSION
  • KOMPUTE_VK_API_MINOR_VERSION
  • KOMPUTE_VK_API_MAJOR_VERSION