|
IntelŽ oneAPI Toolkits
IntelŽ oneAPI products deliver the freedom to develop with a unified toolset and to deploy applications and solutions across CPU, GPU, and FPGA architectures. Native code toolkits implement oneAPI industry specifications and primarily focus on Data Parallel C++ (DPC++), C++, C and Fortran code development. Data science and AI toolkits support machine learning and deep learning developers who primarily use Python* and AI frameworks.
Native Code Toolkits
IntelŽ oneAPI Base Toolkit - This foundational base toolkit enables the building, testing, and optimizing of data-centric applications across XPUs.
IntelŽ oneAPI HPC Toolkit IntelŽ oneAPI IoT Toolkit IntelŽ oneAPI Rendering Toolkit - Toolkits include the new DPC++ programming language, as well as familiar C, C++, and Fortran languages.
- Domain-specific toolkits support specialized workloads.
IntelŽ oneAPI Base Toolkit
Heterogeneous Development Made Easier The IntelŽ oneAPI Base Toolkit is a core set of tools and libraries for building and deploying high-performance, data-centric applications across diverse architectures. It features the Data Parallel C++ (DPC++) language, an evolution of C++ that: - Allows code reuse across hardware targetsCPUs, GPUs, and FPGAs1
- Permits custom tuning for individual accelerators
Domain-specific libraries and the IntelŽ Distribution for Python* provide drop-in acceleration across relevant architectures. Enhanced profiling, design assistance, and debug tools complete the kit.
Specialized Workloads You can complement the Intel oneAPI Base toolkit with additional toolkits.
What's Included
Implement optimized communication patterns to distribute deep learning model training across multiple nodes.
Boost machine learning and data analytics performance.
Develop fast neural networks on IntelŽ CPUs and GPUs with performance-optimized building blocks.
Compile and optimize DPC++ code for CPU, GPU, and FPGA target architectures.
Speed up data parallel workloads with these key productivity algorithms and functions.
Accelerate math processing routines, including matrix algebra, fast Fourier transforms (FFT), and vector math.
Simplify parallelism with this advanced threading and memory-management template library.
Deliver fast, high-quality, real-time video decoding, encoding, transcoding, and processing for broadcasting, live streaming and VOD, cloud gaming, and more.
Design code for efficient vectorization, threading, and offloading to accelerators.
IntelŽ Distribution for GDB* Enable deep, system-wide debug of DPC++, C, C++, and Fortran code.
Achieve fast math-intensive workload performance without code changes for data science and machine learning problems.
Migrate legacy CUDA code to a multi-platform program in DPC++ code with this assistant.
Program these reconfigurable hardware accelerators to speed specialized, data-centric workloads. Requires installation of the Intel oneAPI Base Toolkit.
Speed up performance of imaging, signal processing, data compression, and more.
IntelŽ Integrated Performance Primitives Cryptography A secure, fast, and lightweight library of building blocks for cryptography, highly optimized for various IntelŽ CPUs.
Find and optimize performance bottlenecks across CPU, GPU, and FPGA systems.
Key Specifications
Processors: - IntelŽ and compatible processors
GPUs: - IntelŽ Processor Graphics Gen9
- Xe architecture with integrated graphics
FPGAs: - IntelŽ ArriaŽ 10 FPGAs
- IntelŽ StratixŽ 10 FPGAs
Operating systems: - Linux
- Windows
- macOS (limited support)
Languages: - Data Parallel C++ (DPC++)
- C++
- Python*
Development environments: - Windows: Microsoft Visual Studio*
- Linux: Eclipse*
IntelŽ oneAPI Toolkits Comparison
|
