Now Available: 2024 Release of Intel Development Tools
November 20, 2023 | Intel® Software Development Tools
Expanding Multiarchitecture Performance, Porting & Productivity for AI & HPC
The 2024 Intel® Software Development Tools are available, bringing to developers even more multiarchitecture capabilities to accelerate and optimize AI, HPC, and rendering workloads across Intel CPUs, GPUs, and AI accelerators. Powered by oneAPI (now driven by the Unified Acceleration Foundation), the tools are based on open standards and broad coverage for C++, OpenMP, SYCL, Fortran, MPI and Python.
5 Key Benefits
- Future-Ready Programming – Accelerates performance on the latest Intel GPUs including added support for Python, Modin, XGBoost, and rendering; supports upcoming 5th Gen Intel® Xeon® Scalable and Intel® Core™ Ultra CPUs; and expands AI and HPC capabilities via broadened standards coverage across multiple tools.
- AI Acceleration – Speeds up AI and machine learning on Intel CPUs and GPUs with native support through Intel-optimized PyTorch and TensorFlow frameworks and improvements to data-parallel extensions in Python.
- Vector Math Optimizations – oneMKL integrates RNG offload on target devices for HPC simulations, statistical sampling, and more on x86 CPUs and Intel GPUs, and supports FP16 datatype on Intel GPUs.
- Expanded CUDA-to-SYCL Migration – Intel® DPC++ Compatibility Tool (based on open source SYCLomatic) adds CUDA library APIs and 20 popular applications in AI, deep learning, cryptography, scientific simulation, and imaging.
- Advanced Preview Features – These evaluation previews include C++ parallel STL for easy GPU offload, dynamic device selection to optimize compute node resource usage, SYCL graph for reduced GPU offload overhead thread composability to prevent thread oversubscription in OpenMP, and profile offloaded code to NPUs.
Discover the Power of Intel CPUs & GPUs + oneAPI
- The ATLAS Experiment achieves performance gains by implementing heterogeneous particle reconstruction on Intel GPUs optimized by Intel software tools, including benchmarking of SYCL and CUDA code on Intel and NVIDIA GPUs.
- STAC-A2 Benchmark implementation for oneAPI sets records on Intel GPUs versus NVIDIA.
- VMware and Intel deliver jointly validated AI stack to unlock private AI everywhere for model development and deployment.
