Developer Tools for Intel® Xeon® Processors

With Intel® Software Development Tools and optimized AI frameworks powered by oneAPI, developers can maximize application performance by activating the advanced capabilities of the latest generation Intel® Xeon® processors for AI and accelerated computing, including Intel® Advanced Matrix Extensions (Intel® AMX), Intel® Advanced Vector Extensions 512 (Intel® AVX-512), and Intel Advanced Vector Extensions 2 (Intel® AVX2).

 

 

 

Accelerate AI Frameworks and Applications

Accelerate generative AI (GenAI), LLMs, and other deep learning and data science pipelines using the Intel® oneAPI Base Toolkit and AI Tools from Intel.

Developers can achieve significant performance improvements out of the box on Intel hardware now that Intel® oneAPI library optimizations are being regularly upstreamed to the latest versions of PyTorch,* TensorFlow,* and other leading deep learning frameworks, enabling developers to achieve orders of magnitude performance improvements on Intel hardware using their existing AI workflows.

Intel® oneAPI Deep Neural Network Library (oneDNN) accelerates deep learning and generative AI models on Intel® Xeon® 6 processors with P-cores, the first Intel CPU platform supporting AI acceleration with Intel AMX through FP16 and complex FP16 instruction sets, building on already existing int8 and BF16 extensions.

  • Up to 3x better Llama 2 performance compared to the prior generation for large language models (LLMs)1
  • Up to 1.86x gen-to-gen performance improvement in AI inferencing2

Get the Quick Start Guide to Accelerate AI with Intel AMX and Intel software optimizations.

 

 

Accelerate AI and General Compute Workloads

Build, analyze, optimize, and scale applications with the latest techniques in vectorization, multithreading, multi-node parallelization, and memory using the Intel oneAPI Base Toolkit, Intel® Distribution for Python*, and Intel® oneAPI HPC Toolkit.

  • Accelerate math functions across multiple domains such as BLAS, LAPACK, and FFT on Intel Xeon 6 processors with Intel® oneAPI Math Kernel Library (oneMKL) performance tuning for up to 2.5x better HPCG performance when compared to the prior generation with MRDIMM.3
  • Intel® oneAPI DPC++/C++ Compiler improves data access through preloading cache, reducing latency and Intel AMX FP16 instruction support used by oneDNN. Get up to 1.6x C++ application performance advantage on Linux relative to other compilers on Intel® Xeon® 6 Processors.4
  • Intel® Fortran Compiler supports back end code generation and can achieve up to 1.46x application performance advantage on Linux relative to other compilers on Intel® Xeon® 6 Processors.5
  • Intel® MPI Library now supports 128-core tuning and optimizations for scale out and scale up.
  • Intel® VTune™ Profiler has new features, such as hotspots, microarchitecture and memory access, I/O, and platform diagram, that make identifying performance bottlenecks and memory issues easier.
  • Intel® oneAPI Threading Building Blocks (oneTBB) is enhanced to scale parallel execution performance on Intel Xeon 6 processor's higher CPU core count to accelerate multi-threaded applications.

1) See [9A2] at intel.com/processorclaims: Intel Xeon 6 processors. Results may vary.

2) See [9A3] at intel.com/processorclaims: Intel Xeon 6 processors. Results may vary.

3) See [9H10] at intel.com/processorclaims: Intel Xeon 6 processors. Results may vary. More details on oneMKL benchmarks for Intel Xeon 6 processors.

(4) See https://www.intel.com/content/www/us/en/developer/tools/oneapi/dpc-compiler.html

(5) See https://www.intel.com/content/www/us/en/developer/tools/oneapi/fortran-compiler.html

Multiarchitecture Acceleration

Intel Software Development Tools offers a complete set of advanced compilers, libraries, analysis tools, and optimized frameworks that help developers optimize multiarchitecture applications on Intel CPUs, GPUs, and FPGAs.  

In multiarchitecture systems with Intel Xeon processors and Intel GPUs, using a single code base through oneAPI programming delivers productivity and performance. Accelerate time-to-value when developing high-performance applications. Ensure your software investments continue to add value with future-ready programming.  

Learn how Intel Software Development Tools enable multiarchitecture acceleration.

Migrate from CUDA* to C++ with SYCL* to unlock your code from the constraints of vendor-specific tools and accelerators.

Get Started

Develop and scale your code with confidence with Intel Software Development Tools, the highly productive development stack for AI and open accelerated computing.

Developers can obtain cloud-based access to the latest Intel hardware, including Intel Xeon 6 processors, and Intel Software Development Tools on the Intel® Tiber™ AI Cloud.