TornadoVM 1.0.5

[jjfumero]

Improvements

• Support to construct TornadoNativeArrays from ByteBuffers through memory segment direct copies #402: Support for TornadoNativeArrays from FFI buffers.
• [Refactor] Cleanup and refactor of the code for analysis of parallel loops and loop interchange #403: Clean-up and refactoring for the code analysis of the loop-interchange.
• [jit][cpu] Disable loop-interchange for CPU offloading #405: Disable Loop-Interchange for CPU offloading..
• [jit][ocl] Debug and build kernel improved #407: Debugging OpenCL Kernels builds improved.
• [OCL] CPU Block Scheduler disabled by default and option to switch between thread schedulers #410: CPU block scheduler disabled by default and option to switch between different thread-schedulers added.
• TornadoOptions and Logger improved #418: TornadoOptions and TornadoLogger improved.
• MxM example measured in ns to obtain a better reference on fast GPUs #423: MxM using ns instead of ms to report performance.
• Vector types for Float<Width> with Int<Width> supported at the API-level #425: Vector types for Float<Width> and Int<Width> supported.
• Documentation Updated: Installation, landing page and resources #429: Documentation of the installation process updated and improved.
• SPIR-V Support for the OpenCL Runtime in TornadoVM #432: Support for SPIR-V code generation and dispatcher using the TornadoVM OpenCL runtime.

Compatibility

• [docs] Guidelines to build the documentation #409: Guidelines to build the documentation.
• [windows] Windows installer improved #411: Windows installer improved.
• [installer] install python dependent python modules with pip #412: Python installer improved to check download all Python dependencies before the main installer.
• [docs] Installation for Windows docs improved #413: Improved documentation for installing all configurations of backends and OS.
• [rt][ocl] Select Generic Scheduler if Driver is < 550 #424: Use Generic GPU Scheduler for some older NVIDIA Drivers for the OpenCL runtime.
• [installer] check TORNADO_SDK upfront  #430: Improved the installer by checking that the TornadoVM environment is loaded upfront.

Bug Fixes

• Fix batch computation when the global index is written in the output buffer #400: Fix batch computation when the global thread indexes are used to compute the outputs.
• Refactor TestFields unittest to use Panama data types #414: Recover Test-Field unit-tests using Panama types.
• [chore] Checkstyle errors fixed #415: Check style errors fixed.
• [fix] Problem with FPGA execution for multiple tasks and the default scheduler #416: FPGA execution with multiple tasks in a task-graph fixed.
• Extend the TestFields unittest with lazy and partial copy-out tests and fix reading offsets #417: Lazy-copy out fixed for Java fields.
• [fix] mandelbrot example temp directory fixed to accommodate Windows #420: Fix Mandelbrot example.
• NVIDIA 2D Thread Scheduler Fixed #421: OpenCL 2D thread-scheduler fixed for NVIDIA GPUs.
• [fix] Exclude the nvml library in cmake if platform is jetson nano as… #422: Compilation for NVIDIA Jetson Nano fixed.
• Fix logger for all backends #426: Fix Logger for all backends.
• Math operations for cos/sin with vector types supported #428: Math cos/sin operations supported for vector types.
• Update Jenkinsfile & Fix of JDK link #431: Jenkins files fixed.

---

This discussion was created from the release TornadoVM 1.0.5.