Intelligent Application of SSDs to Accelerate HPC Workloads

By Nicole Hemsoth

October 1, 2012

Introduction

In most industries today, (whether it is financial services, manufacturing, academic research, healthcare and life sciences, or energy exploration) data analysis, modeling, and visualization efforts are critical to success.

To gain a competitive edge, most organizations are incorporating ever-large data sets and more variable data formats into these computational workflows to help derive better information upon which to make smarter decisions.

These big data applications are placing new attention on the high performance computing (HPC) solutions used to run the algorithms and process the raw data. Due to the larger volumes and greater variety of data types, as well as the desire to use more robust analysis, modeling, and visualization routines, HPC solutions can be used to provide high sustained I/O and throughput, while being optimized to cost-effectively handle highly variable workflows.

The essential element in all of this work is a need for speed. Organizations need fast time-to-results so that they can make the right decisions (which well to drill, which new drug candidate to develop, which product design to produce, which customer to award a lower rate loan to) before their competitors.

Complications and challenges that can impede HPC workflows

When looking to accelerate HPC workloads, there are several factors that can play a major role in overall performance.

To start, today’s analysis, modeling, and visualization efforts are carried out using much more sophisticated algorithms in order to derive more detailed and realistic results. The output from these routines offers finer spatial or temporal resolution and consequently results in much larger size output data sets. In a typical workflow, those output files might be used as input to another analysis, modeling, or visualization application.

These operations can impact HPC workflows since the great volumes of data produced by the initial run must be written to disk and saved and then the data must be ingested by yet another routine. Both operations can generate high I/O and throughput demands on an infrastructure. And if the infrastructure is not capable of sustaining these data transfers, the computational workflows can slow significantly.

Another factor has to do with the data that is being used in today’s analysis, modeling, and visualization efforts. Nearly every industry is now making use of much larger data sets, richer sets (such as that produced from newer seismic imaging tools or next-generation sequencers), and many more types of data. However, most users, even those who primarily have large data sets, also have large numbers of small files – even if they consume a relatively small percentage of the total capacity.

Big data and HPC solutions must therefore not only be capable of quickly accessing the large volumes of data required for the computations, they also must intelligently stage the different types of data, which comes in varying file formats and sizes, on suitably high performance storage.

Required storage solution characteristics

Organizations continually deploy new servers with more powerful CPUs to improve and speed up their analysis, modeling, and visualization efforts. To make the best use of such computing resources, an HPC solution must have a suitable storage solution to sustain HPC workflows.

A storage solution for today’s big data and HPC environments must be able to easily scale. Some solutions offer help meeting the growing data volume demands, but fall short when trying to keep CPUs satiated. To help accelerate HPC workflows, a storage solution must also scale in performance so that as the data volumes grow, the system supports the higher I/O and throughput required to get faster results.

Finally, a storage solution must be optimized to handle today’s HPC big data workflows consisting of data sets of files of all sizes. If all data used were in the same format – a structured database, for example – or of the same relative file size, a solution could be highly optimized to handle the specific data. Working with the mixed data sets used today requires a storage solution that optimizes workflow performance for each data type.

Panasas introduces an integrated SSD/SATA approach

Panasas ActiveStor storage systems have a modular blade architecture integrated with its PanFS parallel file system. The design eliminates the bottleneck of a single RAID controller to deliver high-performance, scalable storage. Prior generations of ActiveStor have been based solely on SATA drives and were well-tuned for high throughput.

With the fifth-generation ActiveStor 14, Panasas has taken a unique approach, leveraging lightning fast SSDs integrated with high capacity SATA disk to improve storage performance while keeping costs down. Rather than use SSD for caching or for “most recent” file access as many other vendors have done, ActiveStor 14 stores all metadata and small files (less than 60KB) on the SSDs and larger files on SATA drives.

Metadata is accessed frequently so fast metadata access benefits all types of workloads. All file operations, including reads and writes, require access to metadata. In many cases, such as directory listings, access to the metadata is all that is required to satisfy an I/O request. Storing metadata on SSD boosts performance for all storage operations, especially for directory functions (listing, searches, etc.) and RAID rebuilds in the event of a drive error. Rebuild performance has been improved so that the new 4TB drives can be rebuilt in the same amount of time as the 3TB drives in the prior generation ActiveStor 12, maintaining a high level of data integrity and system reliability.

Small file access can be disproportionately slow when read from, or written to, standard hard disk drives. Accesses of less than a full sector are inefficient, particularly for random I/O. Furthermore, reads and writes of small files can conflict with streaming reads or writes of large files to the same disk. By maintaining small files on SSD, such conflicts are eliminated. In addition, ActiveStor 14 stores the first 12KB of all files inside the file system metadata, improving SSD efficiency while increasing small file performance. This efficient storage of small files on SSD, dramatically improves response time and IOPS, as evidenced by very impressive SPEC sfs2008 NFS IOPS results that Panasas has published.

ActiveStor 14 is available in three configurations with varying sizes of SSD, SATA and cache. The amount of SSD for acceleration ranges from 1.5 percent up to 10.7 percent of total storage capacity. The bulk of the storage capacity, however, is on cost-effective SATA drives, keeping the overall cost per terabyte lower than the prior generation, and very competitive in the market today.

The Importance of Ease of Use and Management

Equally important to the performance and reliability of any storage system is the ease of use and management of the product. With ActiveStor, organizations can simply add blade enclosures to non-disruptively increase capacity and performance of the global file system as storage requirements grow. Parallel access to data and automated load balancing ensure that performance is optimized. This makes it easy to linearly scale capacity to over eight petabytes and performance to 150GB/s or 1.4M IOPS.

Conclusion

The end result is a high-performance storage system that delivers high throughput and IOPS, ideal for the most demanding HPC and big data workloads and accelerates time-to-results. ActiveStor delivers unmatched scale-out NAS performance in addition to the manageability, reliability, and value required by demanding computing organizations in the biosciences, energy, finance, government, manufacturing, media, and other research sectors.

To learn more about how the Panasas ActiveStor 14 can help your organization, register for the live webinar: http://www.panasas.com/news/webinars

 

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!

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion XL — were added to the benchmark suite as MLPerf continues Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

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

March 18, 2024

Nvidia's latest and fastest GPU, codenamed 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…

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…

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…

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…

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…

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…

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

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…

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…

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…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire