當前版本:2021
特色:
- 轉換Intel Parallel Studio XE 與 System Studio 工具套件至對應的 oneAPI產品
- 開發者可以用於 CPU、GPU、FPGA 或是其它加速器開發人工智慧等應用
- 有效率地建立、分析,與最佳化跨架構應用程式
- Fortran/C++ 、Python、OpenMP編譯軟體、MKL 函式庫
- 應用於全球 IoT 邊緣運算
- 可用於 NVIDIA GPU的Data Parallel C++ (DPC++) 編譯器
- oneAPI 可運用在 NVIDIA的GPU產品
- Microsoft、TesorFlow支援 oneAPI
- 操作平台:Linux / Unix/MacOS
Intel® oneAPI 2021更新重點 :
Modernize Code for Performance, Scalability, and Portability
- Write applications that scale with improved parallel performance on the latest Intel® Xeon® and Intel Core™ processors using Intel® Advanced Vector Extensions 512 (Intel® AVX-512) instructions
- Gain greater scalability and reduce latency with next generation Intel® MPI Library
- Vectorize and thread your code (using OpenMP*) to take advantage of the latest SIMD-enabled hardware, including Intel AVX-512
- Speed up machine learning by enabling new high performance Python* optimizations with Intel® Distribution for Python* and Intel® Data Analytics Acceleration Library (Intel® DAAL)
- Stay up to date with the latest standards and integrated development environments (IDE):
- Full C++14 and expanded C++ 2017
- Full Fortran 2008 and partial Fortran 2018
- Python 2.7 and 3.6
- Full OpenMP 4.5 and initial OpenMP 5.0 draft
- Microsoft Visual Studio* 2017 integration
各子版本敘述:
Intel(r) oneAPI Base & HPC Toolkit Single-Node(單機版) | INTEL® ONEAPI HPC 具備核心工具集,包含同類最佳Fortran/C++編譯器、強效函式庫、相容性工具以及分析與除錯工具, 幫助開發者更快速推出 Fortran、OpenMP、MPI 應用程式. For use on laptop, notebook, desktop, PC, or workstation |
Intel(r) oneAPI Base & HPC Toolkit Multi-Node(叢集計算) | INTEL® ONEAPI HPC 具備核心工具集,包含同類最佳Fortran/C++編譯器、強效函式庫、相容性工具以及分析與除錯工具.幫助開發者更快速推出 Fortran、OpenMP、MPI 應用程式. For use on notebook, PC, workstation and HPC clusters |
Intel ONEAPI 版本功能比較:
INTEL® ONEAPI BASE TOOLKIT | INTEL® ONEAPI BASE & HPC TOOLKIT | INTEL® ONEAPI BASE & IOT TOOLKIT | |
Intel® oneAPI Collective Communicaitons Library | √ | √ | √ |
Intel® oneAPI Data Analytics Library | √ | √ | √ |
Intel® oneAPI DPC++/C++ Compiler | √ | √ | √ |
Intel® oneAPI DPC++ Library | √ | √ | √ |
Intel® oneAPI Deep Neural Network Library | √ | √ | √ |
Intel® oneAPI Math Kernel Library | √ | √ | √ |
Intel® oneAPI Threading Building Blocks | √ | √ | √ |
Intel® oneAPI Video Processing Library | √ | √ | √ |
Intel® Advisor | √ | √ | √ |
Intel® Distribution for Python * | √ | √ | √ |
Intel® Integrated Performance Primitives | √ | √ | √ |
Intel® VTuneTM Profiler | √ | √ | √ |
Intel® Distribution for GDB* | √ | √ | √ |
Intel® C++ Compiler Classic | √ | √ | |
Intel® Cluster Checker | √ | ||
Intel® Fortran Compiler | √ | ||
Intel® Fortran Compiler Classic | √ | ||
Intel® Inspector | √ | ||
Intel® MPI Library | √ | ||
Intel® Trace Analyzer and Collector | √ |
函式庫及平行模組資訊:
- Math library: Intel® Math Kernel Library (Intel® MKL)
- *NEW - Data analytics library: Intel® Data Analytics Acceleration Library (Intel® DAAL)
- Threading library: Intel® Threading Building Blocks (Intel® TBB)
- Media/data library: Intel® Integrated Performance Primitives (Intel® IPP)
- Numerical analysis functions for Fortran applications: IMSL* numeric library
- Threading and vectorization parallel model: Intel Cilk Plus
Intel DAAL
利用 Intel® 資料分析加速函式庫 (Intel® DAAL) C++和Java* API在同一節點上處理更多大量資料。 Intel DAAL 提供高度優化的演算法建構模組,加快週邊設備到伺服器等平台上的大量資料分析的性能。涵蓋所有資料分析階段(預先處 理、轉換、分析、建模和決策),可用於離線、串流媒體和分布式分析模式。與常用資料平台(包括Hadoop 和Spark)緊密集成,實現高效的資料存取。