The initial release of Intel® oneAPI 2021.1 is now live!
All tools have now achieved production level performance and quality across all supported CPU, GPU, and FPGA platforms with the exception of the Intel® Fortran Compiler (Beta).
Intel® VTune™ Profiler now provides power analysis for both CPU and discrete GPUs to debug throttling issues and tune flops/watt.
Introduced support for Nuget, Conda, and PIP distribution repositories for improved developer convenience.
Added support for 11th generation Intel® Core™ processors.
Intel® AI Analytics Toolkit enables performance on End-to-end Machine Learning workloads with each stage of the pipeline giving drop-in accelerations against stock versions.
For deep learning workflows, the toolkit adds auto mixed precision optimizations with support for FP32, BF16 and Int8 data types to boost performance of training and inference operations on CPUs with Intel® Optimization for PyTorch and Intel® Low Precision Optimization Tool.
Intel® AI Analytics Toolkit provides GPU support for scikit-learn algorithms such as DBSCAN, K-means, Linear, Logistic regression and more. With Data Parallel Python, the toolkit provides automatic GPU offload with device, queue and USM Data management capabilities.
Download Intel® oneAPI Toolkits TODAY!