Advanced Computing in the Age of AI | Friday, March 29, 2024

New Enterprise Users Pushing HPC Server Growth 

The rising chorus heard at SC15 this week, particularly from server vendors, is that the enterprise sector presents the strongest area of market growth. This trend contributed to the overall HPC server market’s growth of nearly 12 percent to $11.4 by the end of 2015, according to industry watcher IDC, presented at its annual breakfast here at SC15. This is an upward revision to a forecast IDC made earlier this year due in part to the rapid adoption of advanced scale computing servers in the financial services sector.

“Big data combined with HPC is creating new solutions,” said IDC’s Earl Joseph, Vice President, High-Performance Computing, citing particularly strong growth in data-intensive simulation and analytics, “adding many new users/buyers to the HPC space.”

Perhaps the most eye-popping numbers in the report were ROI figures. On average, according to the latest IDC data, $514.7 in revenue is returned for every dollar of HPC invested.

There is broad strength throughout most of the HPC market, according to IDC, with storage again the fastest growing segment. The total 2015 HPC market (see figures below) is forecast to be $22.1B.

The ongoing collision of big data with HPC continues to force changes in the way IDC defines and monitors this emerging market. A while back IDC coined the term High Performance Data Analysis (HPDA). Joseph noted the convergence is ‘creating new solutions and adding many new users/buyers to the HPC space. The finance sector, for example, grew faster than “what we reporting over the last two years (by ~50% higher).”

Identifying the most appropriate buckets within the HPDA category has been an ongoing exercise. Currently, IDC singles out four verticals in HPDA:

* Fraud and anomaly detection – This “horizontal” workload segment centers on identifying harmful or potentially harmful patterns and causes using graph analysis, semantic analysis, or other high performance analytics techniques.

* Marketing – This segment covers the use of HPDA to promote products or services, typically using complex algorithms to discern potential customers’ demographics, buying preferences and habits.

* Business intelligence – The workload segment uses HPDA to identify opportunities to advance the market position and competitiveness of businesses, by better understanding themselves, their competitors, and the evolving dynamics of the markets they participate in.

* Other Commercial HPDA – This catchall segment includes all commercial HPDA workloads other than the three just described. Over time, IDC expects some of these workloads to become significant enough to split out, i.e. the use of HPDA to manage large IT infrastructures, and Internet-of-Things (IoT) infrastructures.

The next new HPDA segment, said Joseph, will be precision medicine. One example is what’s called outcomes-based medical diagnosis and treatment planning. In this paradigm, a patient’s history and symptomology are in a database. While a patient is still in the office, health physicians could sift through millions of archived patient records for relevant outcomes. The care provider considers the efficacies of various treatments for “similar” patients but is not bound by the findings. In effect this is potentially a powerful decision-support tool.

Public and private payer organizations have long promoted the development of similar evidence-based medicine approaches in which whole populations of patients’ records could be examined and evaluated to determine which drugs and therapies are worthwhile and should be approved. In theory, the result in both instances would be better outcomes and reduction of costly outlier practices.

Cloud-based HPC computing is on the rise, both in sheer volume – from 13.8% of sites in 2011, to 23.5% in 2013, to 34.1% in 2015 – and number of workloads being run.

Other key points from the IDC report:

* Growing recognition of HPC’s strategic value is helping to drive high-end sales. Low-end buyers are back into a growth mode.

* HPC vendor market share positions will shift greatly in 2015.

* Recognition of HPC’s strategic/economic value will drive the exascale race, with 100PF systems in 2015 and more in 2016. 20/30MW exascale systems will wait till 2022-2024.

* Non-x86 processors and non-CPUs could alter the landscape – Power, ARM, others; coprocessors, GPUs, FPGAs.

* China looms large(r). Lenovo, growing domestic market, export intentions. Other Chinese vendors are planning to extend to Europe.

* Growing influence of the data center in IT food chain will impact HPC technology options, perhaps providing new approaches.

* HPC in the Cloud Gaining Traction and the big questions are how much and how soon.

Pain points, of course, remain. Software is the number one roadblock – better management software is needed; parallel software is lacking for most users; and many applications will need a major redesign to run in HPC environments. Clusters also remain hard to use and manage. Power, cooling and floor space are major issues. There’s still a lack of support for heterogeneous environment and accelerators. Storage and data management are becoming new bottlenecks.

About the author: George Leopold

George Leopold has written about science and technology for more than 30 years, focusing on electronics and aerospace technology. He previously served as executive editor of Electronic Engineering Times. Leopold is the author of "Calculated Risk: The Supersonic Life and Times of Gus Grissom" (Purdue University Press, 2016).

EnterpriseAI