- Improve Aparapi native to enforce Kernel and Device max work group size limitations and provide query functions for clGetKernelWorkGroupInfo(...)
- Updated Docker and build configurations to use new image for cross-building Aparapi native for multiple platforms
- Allow Aparapi native to run on untested OpenCL Platform versions, just issuing a WARNING message
- Fixed Potential JVM crash when using multi-dimensional arrays (> 1D)
- Fixed exception handling when calling Java method OpenCLDevice.configure() to not hide exceptions
- Fixed local arrays handling 1D and ND, to cope with arrays resizing across kernel executions
- Significant speed-up on discrete GPUs with dedicated memory - OpenCLDevice.setSharedMemory(false)
- Now supports efficient execution on discrete GPU and other devices with dedicated memory
- Support for OpenCLDevice configurator/configure API
- Fixed JVM crash with multi-dimensional arrays in Local memory (2D and 3D local arrays are now supported)
- createProgram in Java had the wrong signature producing a unsatisfied link exception that is now fixed.
- The same scripts can now compile for Linux and OSX and potentially even windows.
- Now supports OpenCL 2.1.
- Fixed the context to a single device instead of all devices of a specific device type. creating a context per type breaks compilation on dopencl.
- Added jni method to write device name
- Fixed a bug whereby the kernel was compiled twice.
- Fixed some bugs where improper directory structure were referenced produced exceptions.
- Changed group id and package to com.aparapi