Highlights include * Improved performance and stability across all compilers, libraries, and tools in preparation for the gold product release coming up later this year.
* Find performance degrading memory transfers with offload cost profiling for both DPC++ and OpenMP using Intel® VTune™ Profiler.
* Debug throttling issues and tune flops/watt using power analysis for CPU using Intel® VTune™ Profiler. GPU power analysis coming soon.
* Intel Distribution of Modin achieves parallelization for 95% of Pandas APIs, while providing 100% functional compatibility. Omnisci backend adds Optane support with 6TB per node for efficient scaling of large data. Also provides the capability to extend to cloud seamlessly from local notebook without manual cluster spawning.
* Accelerate Python math code with the initial release of Data Parallel NumPy (dpnp) a native library and NumPy-like API accelerated with SYCL and Intel GPU support.
* Adds optimizations for scikit-learn algorithms such as Support Vector Classification (SVC), Random Forest and KNeighbors classifiers to speed up model fitting and prediction on Intel CPUs.
* Trained models from XGBoost and LightGBM can now be converted using daal4py library to accelerate model prediction on Intel CPUs.
* Mix ninja level CPU assembly and GPU virtual instruction set code inline with the Intel® oneAPI DPC+/C++ Compiler.
* Perform high-fidelity, ray traced, interactive, and real-time rendering through the new Intel® OSPRay Studio, a scene graph application with graphical user interface, an add-on to Intel® OSPRay.