Pay-Per-Use Becomes Mantra for Altair Engineering

By Michael Feldman

January 18, 2008

To most people, developing the complex software that powers HPC systems is the “rocket science” that makes technical computing so hard. But perhaps an equally challenging aspect for software vendors has been devising a viable licensing model for the era of multicore processing and utility computing.

In general, as computing systems became more powerful, the software vendors charged a premium since less software per platform is required to do the same job. The traditional response to multicore hardware by ISVs has been to either charge by the socket or by the core. Usually though, a complex price bracketing scheme based on the total number of processor cores, processor type and even system memory size is imposed to try to reconcile the growing gap between hardware performance and software use. But since a lot of software doesn’t take complete advantage of additional cores and other hardware improvements, software licensing such as this confers a real cost penalty unless the hardware and software are superbly balanced. Hardware manufacturers see these licensing schemes holding back system sales because the cost of moving the software to the newer platforms has become a disincentive for users.

To address this, Altair Engineering has developed a unified pay-per-use licensing model across their entire product line — HyperWorks (CAE applications), PBS GridWorks (workload management) and HiQube (business intelligence analytics). The model is designed to circumvent some of the limitations of hardware-based licensing. Essentially what the company sells are license tokens that are only drawn when an application is running. Tokens are dispensed from a central license server when an application is executing and returned to the token pool when the application is finished. The model allows these application licenses to be shared across a system, a LAN, or even a wide area network.

The idea originated with the company’s original HyperWorks product line, where the token-based scheme was first introduced. Last summer, Altair converted their PBS GridWorks licensing model to the HyperWorks model. Prior to that, the GridWorks products were employing a more traditional licensing mechanism that treated each type of hardware platform differently. With the introduction of the PBS Professional 9.0, a single “license” uses three GridWorks tokens to run a single job on a single processor core. In the U.S., each GridWorks token costs $4.50, which works out to $13.50 per license per year, or about 1/4 the price of the previous licensing scheme.

By isolating the software licensing requirements from the hardware platform, users are able to use the hardware more flexibly. This is an especially important distinction when multicore processors and utility computing environments are involved, since it provides a more equitable model for sharing hardware resources with other software running concurrently on the same platforms.

While license tokens are the common currency across all Altair products, the denomination of tokens used for GridWorks products is different from that used for HyperWorks products: 1 HyperWorks token unit = 100 GridWorks token units. Also, different Altair applications may draw different numbers of tokens. For example, HyperMesh (a CAE HyperWorks application), draws 21 HyperWorks tokens independent of the core count, while PBS Professional uses three GridWorks tokens per job per core. If the job needs 4 cores, GridWorks would draw 12 tokens from the pool. Once execution completes, those tokens are available for other applications. So the 21 HyperWorks tokens used to run a HyperMesh application during the day could be used to run 175 simultaneous PBS GridWorks jobs (using four cores each) at night.

For codes that run 24/7, such as some solver applications (e.g., RADIOSS used at Ford for crash simulations), they would use dedicated licensing, where the tokens were statically locked to specific hardware. In this case, the tokens could be returned to the pool, at least temporarily, if for example, the hardware is taken down for maintenance.

According to Michael Humphrey, VP of the PBS GridWorks product line, the whole model is especially valuable if a company can maintain a continuous computing level. “The biggest companies — the Boeings, the Fords, the GMs of the world — get the most value out of their HyperWorks units because they use the most applications and they share them globally,” explains Humphrey.

One of the things Altair has going for it is an integrated product set of engineering software, business intelligence and workload management. The company provides an extra incentive to bundle their products through a patented licensing scheme called “leveling.” For example, if you have three different Altair apps running concurrently, the license manager will only draw the number of tokens used by the most expensive application. So if you buy a pre-processing app, a solver and a post-processing app, you really end up only paying for the more expensive solver if you run the three in tandem. This also encourages customers to try out additional Altair software without having to immediately purchase more licenses. Of course Altair is hoping the customer likes the product enough to use it more, requiring the purchase of additional license tokens.

Humphrey thinks the new PBS GridWorks licensing strategy will put a lot of pressure on its competitors, in a number of industry sectors. Even prior to the licensing changeover, he says PBS had become the dominant player in the manufacturing, government and academic sectors. In the automotive area, Altair went after the suppliers first, and then proceeded to the automotive OEMs. In aerospace, they managed to capture Platform Computing business from Lockheed, Boeing, Northrop Grumman and Raytheon.

The incorporation of the pay-per-use licensing strategy will provide additional leverage for PBS Professional to expand accounts in manufacturing, where HyperWorks is already entrenched. “For some of these big companies, their HyperWorks annual renewal is generally two, sometimes three, orders of magnitude (in dollars) bigger than what it would cost to put PBS [Professional] on the entire infrastructure,” says Humphrey.

Because of the price differential with Platform LSF, he contends that in the next two years, Altair will push Platform out of manufacturing almost entirely. However, in the financial services and EDA sectors, where DataSynapse and Platform Computing, respectively, are very strong, Altair has much less penetration.

In the oil and gas sector, Altair is making some headway. They recently won a PBS Professional deal from a top Houston-based oil and gas company, which is using a utility computing model for some of their seismic processing workloads. The Houston company has a large datacenter of their own, but rents an HP datacenter to offload peak computing demands. In this case, the license tokens are consumed and returned transparently to the application, just as if it were being run locally.

Intel has been a customer for some time, using PBS Professional for its benchmarking center. According to Humphrey, prior to switching to the pay-per-use model, Intel wasn’t all that pleased with the static licensing model, wanting a more dynamic arrangement to deal with the constant hardware reshuffling. The new pay-per-use model allows Intel to use and reconfigure the benchmarking systems more flexibly without having to deal with licensing issues.

In general, both chipmakers and system vendors are likely to be supportive of Altair’s pay-per-use model since it doesn’t penalize users for moving up to the next generation of hardware or for using a utility computing setup. But Altair will still have to adjust their token model when chip vendors add more cores on the processor in order to keep software costs from creeping up.

“What we’ve done historically when a new technology rolls out is to adjust our pricing in a negative direction,” says Humphrey. “But remember we’re not selling licenses, we’re selling GridWorks units. So the way we adjust our pricing in the future is not by changing the price of a GridWorks unit, it’s by changing the draw of how many tokens are required each time you run a job.”

That means at some point the company will likely be compelled to reduce the three-token ($13.50 U.S.) cost for a single core job to just two tokens ($9.00 U.S.). In the short term, Altair is planning to expand the functionality of PBS with additional modules and features, allowing them to hold the line on the tokens/job ratio through the quad-core era. But when 8-core processors are introduced, the company will probably have to reduce the tokens consumed per executing job. Beyond that, they may have to reduce the token price itself.

At a time when many ISVs are rather shy about publicizing their licensing model, Altair has positioned its pay-per-use model as an industry innovation that is a key differentiator for its software products. The company has received support from hardware vendors like Intel, AMD and SGI, and has garnered favorable analyst reviews. The real validation for Altair will be if other software vendors, competitors and otherwise, adopt a similar type of licensing model.

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