News Highlights:
Read More :- Western Digital Enables Industry 4.0 Transition With High-EnduranceStorage Solutions for Industrial-Grade AI ML And IOT Applications
- ADATA Launches SC680 External Solid State Drive
- Intel Xeon Scalable Processors Drive Advanced Research in World’sFastest Academic Supercomputer
- China Eastern Airlines Beijing Unicom and Huawei Beijing Launch 5GSmart Travel System
- HARMAN Teams Up with Leading OEMs to Bring Exceptional In-Car AudioExperiences to IAA 2019
- Intel launches oneAPI, a unified and scalable programming version to harness the energy of numerous computing architectures in the technology of HPC/AI convergence.
- Intel introduces a trendy-purpose GPU optimized for HPC/AI acceleration based at the Xe structure, code-named “Ponte Vecchio.”
- Intel unveils additional architectural details of the Aurora Supercomputer, turning in convergence at exascale at Argonne National Laboratory.
DENVER, Nov. 17, 2019 – At Supercomputing 2019, Intel unveiled its imaginative and prescient for extending its management in the convergence of high-overall performance computing (HPC) and artificial intelligence (AI) with new additions to its data-centric silicon portfolio and an bold new software program initiative that represents a paradigm shift from nowadays’s single-structure, single-vendor programming fashions.
Addressing the growing use of heterogeneous architectures in excessive-overall performance computing, Intel accelerated on its current technology portfolio to move, keep and method facts extra efficiently via pronouncing a brand new category of discrete popular-cause GPUs optimized for AI and HPC convergence. Intel also launched the oneAPI enterprise initiative to supply a unified and simplified programming model for software development across heterogenous processing architectures, along with CPUs, GPUs, FPGAs and other accelerators. The launch of oneAPI represents millions of Intel engineering hours in software improvement and marks a recreation-changing evolution from today’s restricting, proprietary programming processes to an open standards-based totally version for cross-structure oneAPI Fact Sheet demand numerous architectures, starting from CPUs, widespread-reason GPUs and FPGAs, to more specialised deep-studying NNPs, which Intel confirmed earlier this month,” said Raja Koduri, senior vice chairman, leader architect, and fashionable manager of structure, photos and software program at Intel. “Simplifying our clients’ ability to harness the energy of diverse computing environments is paramount, and Intel is devoted to taking a software-first approach that offers a unified and scalable abstraction for heterogeneous architectures.”
oneAPI: A Developer-Centric Approach to Heterogeneous Computing
The oneAPI initiative Intel released these days will outline programming for an more and more AI-infused, multi-architecture international. OneAPI gives you a unified and open programming revel in to developers on the architecture in their preference with out compromising performance and getting rid of the complexity of separate code bases, more than one-programming languages, and one of a kind equipment and workflows. OneAPI preserves existing software program investments with help for existing languages even as handing over flexibility for builders to create flexible packages.
oneAPI includes each an enterprise initiative based totally on open specifications and an Intel beta product. The oneAPI specification consists of an instantaneous programming language, powerful APIs and a low-stage hardware interface. Intel’s oneAPI beta software offers builders a comprehensive portfolio of developer tools that include compilers, libraries and analyzers, packaged into area-targeted toolkits. The initial oneAPI beta launch targets Intel® Xeon® Scalable processors, Intel® Core™ processors with incorporated pix, and Intel® FPGAs, with additional hardware help to follow in destiny releases. Developers can download the oneAPI gear, test drive them within the Intel oneAPI DevCloud, and examine extra approximately oneAPI at software.intel.com/oneAPI.
Intel’s Data-Centric Strategy Delivers the Foundation for AI/HPC Convergence
Intel’s silicon portfolio is made out of a numerous blend of architectures deployed in more than a few silicon systems. The foundation of Intel’s information centric strategy is the Intel® Xeon® Scalable processor, which today powers over 90 percent of the arena’s Top500 supercomputers. Intel Xeon Scalable processors are the only x86 CPUs with integrated AI acceleration which can be optimized to investigate the massive records units in HPC workloads.
At Supercomputing 2019, Intel unveiled a new category of popular-reason GPUs primarily based on Intel’s Xe architecture. Code-named “Ponte Vecchio,” this new high-performance, rather flexible discrete fashionable-reason GPU is architected for HPC modeling and simulation workloads and AI schooling. Ponte Vecchio will be manufactured on Intel’s 7nm era and might be Intel’s first Xe-primarily based GPU optimized for HPC and AI workloads. Ponte Vecchio will leverage Intel’s Foveros 3-D and EMIB packaging innovations and feature a couple of technology in-package deal, together with high-bandwidth reminiscence, Compute Express Link interconnect and other intellectual assets.
Building the Foundation for Exascale Computing
Intel’s facts-centric silicon portfolio and oneAPI initiative lays the inspiration for the convergence of HPC and AI workloads at exascale within the Aurora machine at Argonne National Laboratory. Aurora might be the first u.s. Exascale device to leverage the whole breadth of Intel’s information-centric technology portfolio, building upon the Intel Xeon Scalable platform and the use of Xe architecture-primarily based GPUs, in addition to Intel DC Persistent chronic reminiscence and connectivity technology. The compute node architecture of Aurora will feature 10nm-based Intel Xeon Scalable processors (code-named “Sapphire Rapids”) and six Ponte Vecchio GPUs. Aurora will assist over 10 petabytes of reminiscence and over 230 petabytes of storage. Aurora will leverage the Cray Slingshot material to attach nodes across extra than two hundred racks.
