HPE Expands HPC Server Portfolio

Hewlett Packard Enterprise (HPE) announced workload-optimized compute platforms and solutions to help customers accelerate innovation and time-to-value with deep learning systems, HPC and financial services industry (FSI) applications.

Apollo 6500

Organizations running complex HPC and big data workloads, such as modeling, simulation, high frequency trading and deep learning are now able to modernize their data centers with infrastructure solutions that are purpose-built and optimized to analyze, massive volumes of data with speed, scale and efficiency.

IDC forecasts robust growth of the worldwide HPC market from $21 billion in 2015 to $31 billion in 2019. ⁽¹⁾

“As high performance and webscale applications become mainstream, HPE’s continued focus on this market is yielding positive results for our customers,” said Bill Mannel, VP and GM, HPC, big data and IoT servers, HPE. “Already, more than a third of the HPC market is using HPE compute platforms⁽²⁾ to enhance scientific and business innovation and gain a competitive edge. Today’s announcement reinforces our commitment to delivering new infrastructure solutions that satisfy our customers’ insatiable need for massive compute power to fuel new applications and unlock the value of their data.“

Apollo 6500 rear

Accelerate time to value for deep learning applications
Deep learning algorithms are exploding in the industry as organizations are under competitive pressure to support the increasing sophistication of simulation and machine learning models. With up to eight high performance NVIDIA Corp.‘ GPU cards designed for maximum transfer bandwidth, the Apollo 6500 system is purpose-built for deep learning applications. Its high ratio of GPUs to CPUs, dense 4U form factor and efficient design enable organizations to run deep learning recommendation algorithms faster and more efficiently, reducing model training time and accelerating the delivery of real-time results, all while controlling costs.

When used with comprehensive GPU computing platforms like the NVIDIA Tesla Accelerated Computing Platform, the Apollo 6500 provides maximum GPU processing capacity across a broad ecosystem of tools. It is designed to support deep learning computing platforms and application programming interface models, such as Caffe, CUDA, Torch, Theano, Tensorflow, the NVIDIA Deep Learning SDK, and the newly announced Cognitive Computing Toolkit from the company.

“To handle the unprecedented demands of today’s large, complex workloads, the HPC industry has turned to GPU acceleration to deliver the necessary computational performance. The high-density GPU design of the HPE Apollo 6500 maximizes the power of accelerated computing to drive the most computationally intensive deep learning, big data analytics and other HPC applications,” Ian Buck, VP, accelerated computing, NVIDIA.

Apollo 4520

Lower TCO with HPC storage server solution
Most HPC-specific storage solutions on the market today are closed appliances, which limits flexibility for customers, drawing IT organizations into vendor lock-in. The Apollo 4520 system is designed to give customers the opportunity to implement reliable open and supported parallel file system architectures to address their HPC storage needs.

The Apollo 4520, a dual-node system with high performance fabrics and drive failover, is specifically designed to support Lustre implementations. In turn, customers who are looking to accelerate innovation have the choice to implement either an firm’ supported Lustre solution, based on the Intel Corp. Enterprise Edition for Lustre software or Open Source Lustre with community support. Through company’s contributions to the Lustre community as well as tight integration with Open ZFS and Intel Enterprise Edition for Lustre software, the Apollo 4520 solution delivers resiliency and flexibility to meet customers’ most demanding requirements.

“Workload-optimized storage is essential for organizations running high-performance systems to deliver smarter, data-driven decisions. With Intel Enterprise Edition of Lustre on the Apollo 4520, customers are able to realize the flexibility of open-source software with the extreme density of HPE storage hardware. The innovative Apollo 4520 solution was made possible by the HPC alliance with HPE. We look forward to further collaborations to address customers’ most complex HPC needs through the Intel Scalable System Framework.” Charles Wuischpard, VP, datacenter group, GM, HPC platform group, Intel Corp.

“From the validation of Lustre on Open ZFS for enhanced performance tuning, to the hardware support for our development of a S3 HSM copy tool for Lustre, HPE has constantly been at the forefront of innovations for the Lustre community,” said Frederick Lefebvre, HPC system architect, Calcul Quebec, University Laval, Quebec City. “The new HPE Apollo 4520 is a highly flexible foundation for HPC storage innovation and enables Calcul Quebec to provide our users with an optimized solution that is highly scalable with a very compelling TCO.“

Increasing trader productivity, trading performance
and ensuring compliance for the financial services industry
To help financial services customers boost business productivity, gain competitive differentiation in high-frequency trading performance, and manage increasing regulatory compliance needs, the company is introducing a number of industry solutions:

Moonshot Trader workstation solution: Enhances financial trader experience and productivity using Moonshot and Citrix Systems, Inc.‘ solutions. This enables trading companies to deploy physical hosted desktop environments, providing a secure workstation experience with accelerated graphics and high compute performance to run online trading applications.
Trade and Match Server solution: Leverages the density-optimized Apollo 2000 server with frequency optimized single socket processors to consistently deliver up to 28% performance improvement⁽³⁾ on trade processing and minimizing system latency for increased competitive differentiation with lower TCO .
Risk Compliant Archive solution: Enables organizations to keep up with the increasing number of complex, regulatory archiving compliance requirements efficiently and securely, without the need for disruptive and time-consuming data migrations. This solution, composed of iTernity iCAS software verified for data archiving standards and Scality RING file and object storage with Apollo 4000 servers, provides a long-term, robust, and cost-effective platform for implementing compliant data archives combined with an ability to support multiple media formats, applications, and locations.
In addition to these industry solutions, the firm is releasing an innovative edition of Vertica for Hadoop – configured Apollo and Proliant platforms. The company will also be issuing reference architectures for optimizing company’s Vertica for SQL on Hadoop to deliver fast, advanced analytics. This combined offering helps organizations tackle their largest analytics challenges to support any Hadoop distribution.

“Enterprises across diverse industries are challenged by the demanding performance, scale and efficiency requirements of complex simulation, modeling and analytics problems,” said Steve Conway, research VP, HPC and data analysis, IDC. “HPE’s new HPC solutions feature innovations in systems design, workload optimization, density optimization and open source software that are designed to accelerate time-to-value in areas such as deep learning, energy exploration and mechanical design, as well as financial trading and regulatory compliance.”

To help customers with their HPC environments, the company offers skilled professionals to deliver a range of design, deploy and management services.

Pricing and availability:

Apollo 6500 will be available in Q3 2016.
Apollo 4520 will be available on April 18. The Apollo 4520 chassis will start at $8,500.
FSI solutions are available and pricing will vary based on customer requirements.

Additional Information:
High Performance Computing: 5 Trends That Could Define This Year (and Beyond) by Bill Mannel
HPC Market Leader Takes Aim at Big Compute-Big Data Convergence

⁽¹⁾ IDC, Worldwide HPC Server 2015-2019 Forecast, Mar 2015, Doc #255147, Market Analysis
⁽²⁾ IDC, HPC WW IDC QView 4Q 2015, Earl Joseph and Bob Sorensen; March 20, 2016
⁽³⁾ Based on HPE internal benchmarks