Solving Simulation Complexity

By David Lourerio

July 15, 2014

Although HPC has helped solve important problems in nuclear physics, earth sciences, meteorology, etc. since the 1960s, HPC has traditionally existed only for specialists. But, with the democratization of x86 servers, large storage systems, and high-end networking technologies, companies and research facilities can now buy clusters at a reasonable ratio (power/cost) with very powerful, complex technologies or get access to remote HPC cloud platforms.

However, even though the resources are available, widespread use has been limited due to the difficulties in accessing and using the resources and/or adapting existing HPC software and hardware that still exist. Back in the nineties, researchers could spend time getting used to those hurdles, but today business ROI drives engineers and researchers to produce results in much less time. As a result, overcoming the HPC access and usability hurdles has become crucial and this has been further exacerbated by the fact that many more users of large-scale platforms are not specialists of HPC use and programming.

This article addresses the need for HPC to become a tool that’s as simple and transparent as possible so that small and medium businesses (SMBs) can easily access the resources of mid-sized clusters and use complex applications on hybrid HPC resources. To achieve this without extensive financial outlay, end users need to execute and control the simulation of applications and data directly from their desktops.

The HPC Challenge for the Smaller Business

While R&D labs and large companies may be accustomed to accessing and exploiting large computing infrastructures, small-to-medium-businesses and smaller business units within large companies that have no dedicated computing resources or HPC experts must now deal with increasing amounts and complexity of data. These smaller groups cannot afford to abandon the advantages HPC provides, but need a way to access the processing power and capabilities that mid-sized clusters offer. This need for readily available “super applications” is first referred to by Dr. Erik Deumens, Director of Research Computing at the University of Florida in his paper “What Drives Investment in the Middle of HPC?” to describe simulations that run on mid-sized computing clusters that are fully accessible on the end-user desktop.

At an international scale, some countries have tried to address this through various resources. In France, INRIA, BPI (the French public investment bank), and GENCI launched “Initiative HPC-PME” (http://www.initiative-hpc-pme.org/), an initiative aiming at providing knowledge, funds, and computing resources to enable simulation software to be powered on HPC resources for SMBs. Free Field Technologies, a French SME, participated in this program and in this testimonial, they present how improved simulation times of HPC helped them win Airbus as a customer.

In Europe, “Fortissimo” was (http://www.fortissimo-project.eu/index.html) launched in October 2013. In this initiative, SMBs needing HPC computation gain access to the resources of a research lab and a computing facility. SMEs like Pipistrel, for example, are using these HPC facilities to improve their simulations of lightweight aircraft designs.

In Silicon Valley, ÜberCloud (http://www.theubercloud.com/hpc-experiment/), launched by Burak Yenier and Wolfgang Gentzsch, was established to show how the SaaS (Software-as-a-Service) model can be used for HPC simulation. A number of large companies, software providers, and computing facilities have now implemented these technologies:

  • “High-resolution Computer Simulations of Blow-off in Combustion Systems” involving CSE, a US-based company
  • “Drifting Snow Around Arbitrary Structures” where end-users were SMEs like Binkz. Inc, a Canadian Based company
  • “Performance Analysis of GROMACS Molecular Dynamics for Simulating Enzyme Substrate” with the Indian SME called Polyclone Bioservices, and the second one has been released last week with use-cases about, or in the field of biology with a use case around.

Moving the Needle Forward

Although these advances have been implemented by some of these initiatives, there’s a need to take the solution further, to provide broad-scale access to many users. To understand what’s needed, let’s first take a look at what these users have in common.

All these different types of users need to upload their data, launch (large-scale) non-interactive simulations, and post-process them. This workflow might be business- or research-centered, but for both communities, there’s a very real need to get better results faster. This need far outweighs knowing if the computation involves using GPUs, fat nodes, or an ultra-fancy middleware designed specially for their purpose.

A preamble of all these tasks is access to the applications. The SaaS (Software-as-a-Service) model does fit with this goal. Through the browser we use every day to check our e-mail or visit websites, we can also work on complex and resource-demanding applications.

What barriers do SMBs need removed from HPC?

The ability to create structures that allow non-technical users to directly access hybrid resources is essential. Here we explore the characteristics of a web portal that would enable SMBs to easily gain access to them.

Simplified universal access—To accommodate bandwidth, the interface must be as light as possible: no browser plugins, no java applets, no flash support, or port opening required. By stripping out all of these dependencies, the interface becomes “vanilla”—able to interface with the researcher’s smart phone, desktop, or tablet regardless of its operating system.

This simplified interface creates the necessary foundation for a SaaS model to access HPC simulation software running remotely on computing clusters. Figure 1 offers an example of interface that is able to adapt to various terminals thanks to the latest HTML/CSS/JavaScript frameworks. This framework provides access to the basic info a user needs to start working: his current projects, past jobs, etc., from anywhere.


Screen Shot 2014-07-15 at 12.10.21 PM

Fig. 1 Example of interface providing access to user’s projects and his previous works

Secured access—Small businesses also need secure access. There’s been a lot of discussion about what should be used—X.509 certificates, login and passwords, SSH public and private keys. In fact, there is no standard way of achieving security. It depends heavily on the policies you want to set-up.

Inside the clusters, you could use any of these methods. But from the user’s point of view, the processes must be unified with at least a login and password that then manages every security system underneath. By doing this, the interface reduces the barrier for the end-user.

And, transfers must be, of course, encrypted to ensure no disseminations to competitors for example.

Familiar file structure—When uploading data or dealing with the HPC computing facilities, users need something that’s familiar, that doesn’t change their habits, such as classical files managers of Windows or FTP clients. This ensures that users can process vast amounts of data without having to take an excessive amount of time to prepare the files and to post process the results afterwards.

Remote Visualization—Today, computing facilities provide both computing and graphical resources. To be able to run non-interactive computations and to then do graphical post-processing on the same site is a great improvement for end-users. If the workflow interface runs fluently from computation into post-processing, then users no longer have to go onsite and disturb administrators to transfer video streams. Instead they can use remote desktop technologies from the same interface to bridge from simulation to graphical post-processing without any additional requirements.

Link with the previous capability, e.g. simplified data transfers, the traditional workflow consisting of input simulation data staging, processind and post-processing can be run in a drastically reduced timeframe.

Screen Shot 2014-07-15 at 12.09.55 PM

Fig. 2 File management interface allowing users to transfer to and from the computing resources

Simplified graphical interface—While there are many ways to access applications—home-made scripts, proprietary code, web portals—the interface needs to offer users a way to define specific parameters or modify scripts depending on the outputs they seek. This management of applications has to be as simple as possible and take the least amount of time.

The idea of using XML files to define any kind of interfaces was vogue for a while, but such interfaces were quickly recognized as just hype since they were not productive. Instead, what’s needed is a graphical tool that provides a way to design the interface that the end-users will use to interact with the scripts that execute their applications.

In designing graphical interface access, it’s easy to go overboard and provide full access to scripts and the underlying plumbing, which adds to system complexity. Well-defined and flexible right management must be provided. Only users developing scripts should be allowed to add, update, and modify applications, and only managers should be able to access usage reporting, user management, etc.

Streamlined—As shown in Figure 3, integrating an application should be no more complex than defining a script to execute, some documentation, and the input parameters the user must provide. A specific form directing the application use can then launch for the end-user. This efficiency helps the user concentrate on the most important part of their task—running the simulation with good input data.

Screen Shot 2014-07-15 at 12.10.49 PM

Fig. 3 Web interface providing an easy way to design submission forms for non-interactive applications

Billing the user—Once a framework provides access to data and the applications and links these to a global, simple-to-use workflow, the resources need to be monitored and time spent needs to be easily allocated to cost centers. Whether computing hours are divided between certain computing facilities or allocated to business units in companies, clear and precise accounting must detail who had access to the applications and resources.

Usage reporting also needs to be available in a digital format that interfaces to accounting packages for invoice creation or be integrated into existing quota systems. Figure 4 shows an interface presenting accounting and monitoring of resources and applications usage that are the basis of billing and invoicing.

Screen Shot 2014-07-15 at 12.11.05 PM

Fig. 4 Usage statistics showing the consumption of a project on different clusters

Democratized HPC

It is through these types of improvements and simplifications that SMB employees will gain access to HPC simulation. With such a simplified interface to powerful complex resources, businesses will boost productivity and be able to efficiently do their high-end computational work in much less time.

Removing HPC complexity from the game, streamlines data analysis and simulation for the SMBs, ensuring maximum productivity and increasing competitiveness.

About the Author

David Loureiro is the CEO and co-founder of SysFera, a software innovator that simplifies management and accessibility of high performance computing (HPC) environments. David holds a Master of Science in Applied Mathematics and started his career at INRIA working on distributed grid and cloud middleware in the INRIA research team called Avalon. His research interests are focused on cloud technologies, HPC resource management, distributed computing and scientific visualization web portals.

david-loureiro-chapo_vignette

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