New Mellanox Performance Benchmarks Released

June 25, 2015

BOSTON, Mass., June 25 — Mellanox Technologies, Ltd., a leading supplier of high-end cloud and storage networking solutions, today announced it has produced a set of storage performance benchmark results working with partners and customers—that demonstrate an increase in performance for Red Hat Ceph Storage and Red Hat Gluster Storage with network connections at speeds up to 100Gb/s. Mellanox partnered with Red Hat and its partners, including SanDisk, Scalable Informatics, Supermicro, and Storage Foundry, to show the benefits of using fast, low-latency 10, 40, 56, and 100Gb Ethernet networks for Red Hat Ceph Storage clusters. These solutions allow larger and faster Red Hat Ceph Storage daemon (OSD) servers when using all-flash configurations or large numbers of hard drives in each server to perform beyond standard 10Gb networking, making Red Hat Ceph Storage deployments high performing and cost effective.

Red Hat Ceph Storage is well suited for archival and rich media and cloud infrastructure workloads, such as OpenStack. As enterprise and cloud customers evaluate and deploy Red Hat Ceph Storage on faster servers, with more flash, and for more demanding users, they are looking to faster networking solutions to increase performance.

“Mellanox is investing heavily in delivering the benefit of high speed networks to Red Hat Ceph Storage and Red Hat Gluster Storage customers,” said Kevin Deierling, vice president of marketing at Mellanox Technologies. “We see a groundswell of interest in using software-defined storage and in deploying Red Hat Ceph Storage and Red Hat Gluster Storage with 10, 40, 56 and 100Gb/s Ethernet network speeds in order to maximize performance and support flash storage and new workloads such as SQL databases and Big Data.”

Mellanox, Red Hat, Seagate, and Supermicro are also in an ongoing Red Hat Ceph Storage benchmarking project to demonstrate the performance with various combinations of flash and hard drives. The first phase proved the following benefits of using 40GbE instead of 10GbE networks:

  • Increases large read performance up to 100 percent for Red Hat Ceph Storage servers with 72 drives and up to 20 percent for Red Hat Ceph Storage servers with 36 disks
  • Reduces read and write latency up to 50 percent
  • Provides optimal price performance for write-intensive workloads on large-capacity Red Hat Ceph Storage servers

Scalable Informatics has been collaborating with Mellanox and Red Hat using 100GbE networking on an all-flash based Unison Red Hat Ceph Storage appliance. They have demonstrated 8GB/s+ reads from disk between a single appliance and client using Red Hat Ceph Storage. This combination provides an incredibly dense, high performance Red Hat Ceph Storage SSD solution with simplified networking.

In addition, Remote Direct Memory Access (RDMA) – supporting both RoCE and InfiniBand – is now available as a technology preview in the Ceph Hammer community release and has recently been enhanced in Red Hat Gluster Storage 3.1, having been first made available in Red Hat Gluster Storage since release 3.0.3 in January 2015. Mellanox testing shows RDMA accelerates Red Hat Ceph Storage large block random read throughput up to 23 percent and Red Hat Ceph Storage small block random read IOPs up to 45 percent. In addition, RDMA increases Red Hat Gluster Storage large-file read throughput up to 18 percent.

“SanDisk has significantly improved Red Hat Ceph Storage performance on flash with our InfiniFlash all-flash storage system, demonstrating a fourfold to tenfold increase in Red Hat Ceph Storage IOPS,” said Gary Lyng, sr. director of solutions and product marketing, enterprise solutions at SanDisk. “By working with Mellanox and leveraging its 40GbE networking, we can help enable Red Hat Ceph Storage for new workloads and offer a high performance solution at an unprecedented cost structure. We look forward to continuing our work with Red Hat, Mellanox and the open source community to help increase Red Hat Ceph Storage performance and expanding workloads even further.”

“Testing of an all-flash configuration with Mellanox 100GbE networking and our SSDs based Unison Ceph Appliance let us achieve a sustained 8+ gigabytes per second from a single server to a single client,” said Russell Nordquist, systems engineering manager at Scalable Informatics, a leading Hyper-converged appliance vendor headquartered in Plymouth, Michigan, that sells storage and analytics appliances to the financial services, pharmaceutical, and research customers. “Our appliances are designed to support incredibly high IO rates, and leveraging Mellanox’s 100GbE networking allows us to serve the data at twice the speed as before over a single link. We have customers now deploying Unison Ceph appliances using 100Gb Ethernet and flash drives to handle bigger and faster workloads beyond the traditional archive and object storage.”

“Storage is a big challenge for everyone, and we’re all hunting for a simple, scalable, and cost-effective answer,” said Craig Forbes, Product Manager at the storageFOUNDRY. “Those requirements created the Nautilus line of software-centric storage systems, a hybrid of hardware and software technologies that include Red Hat Enterprise Linux and Red Hat Ceph Storage. For the connectivity of Nautilus, Mellanoxcreates an amazing high-speed network using their 10GbE-56GbE ConnectX-based adapters and SwitchX-2 switches, which have allowed us to support over 3700MB/sec for a single client connecting to Nautilus via NFS.”

“Supermicro’s longstanding relationship with Mellanox and Red Hat enables innovation across our server and storage portfolio including 2U TwinPro and 36, 72, and 90x hot-swap Double-Sided Storage solutions,” said Don Clegg, vice president of marketing and business development at Supermicro. “Our collaborative efforts on high performance networking are helping drive the next generation of software-defined storage. With the Supermicro advantage, customers have the performance, flexibility and manageability to deliver cloud scale applications.”

RDMA networking also complements the low latency of PCIe flash media. Red Hat Gluster Storage test results using 4-node Gluster system with 4 clients, using 2 PCIe flash drives per node with RDMA vs. TCP using all-flash media show the following performance improvements:

  • 1MB block random IO (Gluster):
    • Reads: RDMA does 13.03M IOPS vs. 8.25M with TCP – 57 percent performance improvement
    • Writes: RDMA does 2.08M IOPS vs. 1.39M with TCP – 49 percent performance improvement
  • 16KB block random IO (Gluster):
    • Reads: RDMA does 2.08M IOPS vs. 1.39M with TCP – 49 performance improvement
    • Writes: RDMA does 0.97M RDMA vs. 0.82M with TCP – 18 percent improvement

Red Hat Gluster Storage is purpose-built as a scale-out file store with a straightforward architecture suitable for public, private, and hybrid cloud environments. Red Hat Gluster Storage offers mature NFS, SMB and HDFS file interfaces and is well suited for enterprise virtualization, analytics, enterprise sync and share, and rich media workloads.

“To scale to the needs of enterprise IT departments and web-scale service providers, software-defined storage can be deployed with advanced networking capabilities that provide the requisite high throughput, IOPS and low latency,” said Brent Compton, director, ecosystem, Storage and Big Data, Red Hat. “Our benchmark testing shows that certain Red Hat Ceph Storage configurations are faster or more cost effective using 40Gb Ethernet.”

About Mellanox

Mellanox Technologies is a leading supplier of end-to-end InfiniBand and Ethernet interconnect solutions and services for servers and storage. Mellanox interconnect solutions increase data center efficiency by providing the highest throughput and lowest latency, delivering data faster to applications and unlocking system performance capability. Mellanox offers a choice of fast interconnect products: adapters, switches, software, cables and silicon that accelerate application runtime and maximize business results for a wide range of markets including high-performance computing, enterprise data centers, Web 2.0, cloud, storage and financial services. More information is available at www.mellanox.com.

Source: Mellanox Technologies

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, code-named Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from its predecessors, including the red-hot H100 and A100 GPUs. Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. While Nvidia may not spring to mind when thinking of the quant Read more…

2024 Winter Classic: Meet the HPE Mentors

March 18, 2024

The latest installment of the 2024 Winter Classic Studio Update Show features our interview with the HPE mentor team who introduced our student teams to the joys (and potential sorrows) of the HPL (LINPACK) and accompany Read more…

Houston We Have a Solution: Addressing the HPC and Tech Talent Gap

March 15, 2024

Generations of Houstonian teachers, counselors, and parents have either worked in the aerospace industry or know people who do - the prospect of entering the field was normalized for boys in 1969 when the Apollo 11 missi Read more…

Apple Buys DarwinAI Deepening its AI Push According to Report

March 14, 2024

Apple has purchased Canadian AI startup DarwinAI according to a Bloomberg report today. Apparently the deal was done early this year but still hasn’t been publicly announced according to the report. Apple is preparing Read more…

Survey of Rapid Training Methods for Neural Networks

March 14, 2024

Artificial neural networks are computing systems with interconnected layers that process and learn from data. During training, neural networks utilize optimization algorithms to iteratively refine their parameters until Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, code-named Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Houston We Have a Solution: Addressing the HPC and Tech Talent Gap

March 15, 2024

Generations of Houstonian teachers, counselors, and parents have either worked in the aerospace industry or know people who do - the prospect of entering the fi Read more…

Survey of Rapid Training Methods for Neural Networks

March 14, 2024

Artificial neural networks are computing systems with interconnected layers that process and learn from data. During training, neural networks utilize optimizat Read more…

PASQAL Issues Roadmap to 10,000 Qubits in 2026 and Fault Tolerance in 2028

March 13, 2024

Paris-based PASQAL, a developer of neutral atom-based quantum computers, yesterday issued a roadmap for delivering systems with 10,000 physical qubits in 2026 a Read more…

India Is an AI Powerhouse Waiting to Happen, but Challenges Await

March 12, 2024

The Indian government is pushing full speed ahead to make the country an attractive technology base, especially in the hot fields of AI and semiconductors, but Read more…

Charles Tahan Exits National Quantum Coordination Office

March 12, 2024

(March 1, 2024) My first official day at the White House Office of Science and Technology Policy (OSTP) was June 15, 2020, during the depths of the COVID-19 loc Read more…

AI Bias In the Spotlight On International Women’s Day

March 11, 2024

What impact does AI bias have on women and girls? What can people do to increase female participation in the AI field? These are some of the questions the tech Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Analyst Panel Says Take the Quantum Computing Plunge Now…

November 27, 2023

Should you start exploring quantum computing? Yes, said a panel of analysts convened at Tabor Communications HPC and AI on Wall Street conference earlier this y Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Training of 1-Trillion Parameter Scientific AI Begins

November 13, 2023

A US national lab has started training a massive AI brain that could ultimately become the must-have computing resource for scientific researchers. Argonne N Read more…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire