Back to Table of contents

Primeur weekly 2017-06-26

Focus

China's effort on HPC in the next 5 years - from exascale prototypes to exascale system ...

Crowd computing

BOINC Monitor 9.70 released ...

Focus on Europe

Eight Irish Supercomputer List: Irish HPC capacity more than doubles again ...

Memorandum of Understanding signed between CHPC and PRACE ...

GENCI to boost France scientific competitiveness and industrial innovation with new petascale supercomputer ...

Huawei inaugurates the HPC Global Center of Excellence ...

Hardware

Universal ultra high-dense and 'hot water' cooled RSC Tornado solution: ready to support Intel Xeon Processor Scalable Family, world's first Intel Omni-Path fabric based and 100% 'hot water' liquid cooled switches, improved RSC BasIS functionality ...

NVMe Revision 1.3 expands reach of fast storage for Enterprise, Client, and Cloud power users ...

DDN Storage named to elite $1 billion+ valuation "Storage Unicorn" list ...

Verne Global sets strategic roadmap to manage advanced computing requirements ...

Cisco and NetApp advance digital transformation with software-defined converged infrastructure solution for the next generation data centre ...

Cisco unveils network of the future that can learn, adapt and evolve ...

New Supermicro X11 SuperBlade boosts I/O performance featuring Intel Omni-Path fabric ...

Supermicro announces full portfolio of A+ server solutions optimized for new high-performance AMD EPYC processors ...

Mellanox interconnect solutions scale deep learning platforms to world-leading performance ...

Mellanox Ethernet and InfiniBand chosen by AMD as the preferred interconnect solutions to accelerate new EPYC data centre platforms ...

Applications

Shape and size of DNA lesions caused by toxic agents affects repair of DNA ...

SDSC's Comet is a key resource in new global dark matter experiment ...

New computing system takes its cues from human brain ...

Blue Brain team discovers a multi-dimensional universe in brain networks ...

Machine learning and high performance computing for industrial applications ...

Modelling the brain with Lego bricks ...

How pythons regenerate their organs and other secrets of the snake genome ...

The Cloud

European Commission to set up new High Level Expert Group 2017-18 for the European Open Science Cloud ...

Huawei releases HPC Cloud solution 2.0 ...

Mellanox interconnect solutions scale deep learning platforms to world-leading performance

20 Jun 2017 Sunnyvale, Yokneam - The leading deep learning frameworks such as TensorFlow, Caffe2, Microsoft Cognitive Toolkit, and Baidu PaddlePaddle now leverage Mellanox's smart offloading capabilities to provide world-leading performance and near-linear scaling across multiple AI servers. Mellanox RDMA and In-Network Computing offloads and NVIDIA GPUDirect are key technologies enabling users to maximize their application performance and system efficiencies.

Deep learning is used across industries and the research community to help solve many big data problems such as natural language processing, speech recognition, computer vision, healthcare, life-sciences, financial services and more. Mellanox is enabling these industries into a new era of performance and scalability with the powerful data-centric offload architecture that has been employed by the world's most advanced machine learning platforms.

TensorFlow is an open source software library originally developed by researchers and engineers within Google's Machine Intelligence research group. With the inclusion of RDMA technology in place of traditional TCP, TensorFlow data exchange performance between nodes was accelerated by 2x, enabling faster image processing.

Baidu's PaddlePaddle - Parallel Distributed Deep Learning - is a flexible and scalable deep learning platform. PaddlePaddle supports a wide range of neural network architectures and optimization algorithms, such that it is possible to leverage many CPUs and GPUs to accelerate training. PaddlePaddle leverages RDMA to achieve high throughput and performance, and takes advantage of the more advanced acceleration capabilities of the combined NVIDIA and Mellanox architectures to accelerate deep learning training time by 2x.

"Advanced deep neural networks depend upon the capabilities of smart interconnect to scale to multiple nodes, and move data as fast as possible, which speeds up algorithms and reduces training time", stated Gilad Shainer, vice president of marketing at Mellanox Technologies. "By leveraging Mellanox technology and solutions, clusters of machines are now able to learn at a speed, accuracy and scale that push the boundaries of the most demanding cognitive computing applications."

"Developers of deep learning applications can take advantage of optimized frameworks and NVIDIA's upcoming NCCL 2.0 library which implements native support for InfiniBand verbs and automatically selects GPUDirect RDMA for multi-node or NVIDIA NVLink when available for intra-node communications", stated Duncan Poole, Director of Platform Alliances at NVIDIA. "NVIDIA NVLink is available in Pascal-based Tesla P100 systems, including the NVIDIA DGX-1 AI supercomputer which has four Mellanox ConnectX-4 100 Gb/s adapters. This allows developers to focus on creating new algorithms and software capabilities, rather than performance tuning low-level communication collectives."

Source: Mellanox

Back to Table of contents

Primeur weekly 2017-06-26

Focus

China's effort on HPC in the next 5 years - from exascale prototypes to exascale system ...

Crowd computing

BOINC Monitor 9.70 released ...

Focus on Europe

Eight Irish Supercomputer List: Irish HPC capacity more than doubles again ...

Memorandum of Understanding signed between CHPC and PRACE ...

GENCI to boost France scientific competitiveness and industrial innovation with new petascale supercomputer ...

Huawei inaugurates the HPC Global Center of Excellence ...

Hardware

Universal ultra high-dense and 'hot water' cooled RSC Tornado solution: ready to support Intel Xeon Processor Scalable Family, world's first Intel Omni-Path fabric based and 100% 'hot water' liquid cooled switches, improved RSC BasIS functionality ...

NVMe Revision 1.3 expands reach of fast storage for Enterprise, Client, and Cloud power users ...

DDN Storage named to elite $1 billion+ valuation "Storage Unicorn" list ...

Verne Global sets strategic roadmap to manage advanced computing requirements ...

Cisco and NetApp advance digital transformation with software-defined converged infrastructure solution for the next generation data centre ...

Cisco unveils network of the future that can learn, adapt and evolve ...

New Supermicro X11 SuperBlade boosts I/O performance featuring Intel Omni-Path fabric ...

Supermicro announces full portfolio of A+ server solutions optimized for new high-performance AMD EPYC processors ...

Mellanox interconnect solutions scale deep learning platforms to world-leading performance ...

Mellanox Ethernet and InfiniBand chosen by AMD as the preferred interconnect solutions to accelerate new EPYC data centre platforms ...

Applications

Shape and size of DNA lesions caused by toxic agents affects repair of DNA ...

SDSC's Comet is a key resource in new global dark matter experiment ...

New computing system takes its cues from human brain ...

Blue Brain team discovers a multi-dimensional universe in brain networks ...

Machine learning and high performance computing for industrial applications ...

Modelling the brain with Lego bricks ...

How pythons regenerate their organs and other secrets of the snake genome ...

The Cloud

European Commission to set up new High Level Expert Group 2017-18 for the European Open Science Cloud ...

Huawei releases HPC Cloud solution 2.0 ...