At SQUADEX, we are doing our best to keep you up to speed with important updates. And this time we are going to share an overview from ISC High Performance 2018 event that took place in Frankfurt, Germany on June 24-28.

SQUADEX took its stage with the FlyElephant cloud platform to present their capabilities in high-performance computing and machine learning.

Dmitry Spodarets, ML & HPC Lead at SQUADEX, took on the role of Technology and Consulting Partner at the AWS conference booth, demonstrated a demo of the FlyElephant platform, and offered actionable advice about ML and HPC to the attendees.

“At ISC 2018, I demonstrated how data scientists could quickly create working environments using required software and powerful Amazon EC2 C5 or Amazon EC2 P3 instances for developing or automatic training of machine learning models. I exemplified that using a model for image classification.

As a special case for educational institutions, I shared some details on how the FlyElephant platform was deployed on AWS for simultaneous workloads of more than 400 students who did ML laboratory work.

On top of that, I presented to attendees and prospects how SQUADEX augments the FlyElephant platform with their services. For instance, FlyElephant would be supported by the SQUADEX consulting team to enhance the development and optimization of machine learning models. We are looking forward to delivering on the Data Science tasks of our clients. Having these capabilities joined, our target is to  help businesses get the most out of Data Science with the help of machine learning algorithms.”

Note that SQUADEX teamed up with FlyElephant just before the ISC 2018 event to empower their platform with top-notch consulting services, and to deliver more AI- and HPC-enabled solutions to the market.

But, without further ado, let’s get back to the conference.

About ISC High Performance 2018 Event

ISC High Performance is a celebrated HPC event. For over 33 years, it has been bringing together experts, engineers, and ecosystem-minded leaders who are specialized in a broad array of topics — from HPC and AI to climate change and cosmology.

In 2018, ISC High Performance was attended by more than 3,500 visitors from 60+ countries (70% of attendees and exhibitors were not from Germany). In total, 162 companies and research institutions demonstrated their latest technologies and services.

In an effort to advance high-performance computing in scientific and commerce fields, ISC 2018 pursued to strike the right balance between science and industry participants of the event and offered dedicated sessions to unveil specific aspects of HPC.

Individual sessions were also attributed to artificial intelligence, machine learning, and industrial technology.

This year ISC’s organizers encouraged conversations between industry leaders and researchers on such topics as global warming, climate change, and gender equality among others.

To celebrate the FIFA World Cup, organizers added some football flavor to the event.

ISC 2018: Announcements and Emerging Trends

Top500 Supercomputer Rankings Updated

International Supercomputing Conference 2018 would not have been complete without the obligatory update of the Top500 Supercomputer rankings, with a US-based, IBM-built supercomputer “Summit” taking the top spot.

Overall, the top 5 has seen a major shakeup, with four supercomputers new or considerably upgraded. As of June 2018, the top 5 is as follows:

Screenshot: https://www.top500.org/lists/2018/06/

Consisting of 4,356 nodes, “Summit” is powered by two 22-core Power9 CPUs and six NVIDIA Tesla V100 GPUs, and is able to deliver 122.3 petaflops per second.

The builds of other systems in the top 5 demonstrate a slow departure from heavy compute iron towards the use of CPUs and GPUs for augmentation.

AI & ML Development Within the HPC Domain

Though AI & ML are related to HPC, up to 2018 there were only a few cases of their successful use to improve performance, scalability, and configurability of real-world solutions.

However, as it was demonstrated by some apps at ISC 2018, in the near future AI, ML, and Big Data (HDPA) would be utilized along HPC to improve a product’s architecture by optimizing the use of computer resources, and to significantly increase a system’s overall performance.

Potentially, this “convergence” will lead to the development of broad capability systems with next-level configurability and cross pollination. Systems as such will require a new set of performance metrics, other than Linpack.

HPC Moves to the Cloud

According to the industry report presented at ISC 2018 by Addison Snell, CEO of Intersect360 Research, in comparison to 2017 HPC in the cloud grew by 44%.

Estimated to have reached $1.1bn USD by mid-2018, the sector is expected to further gain momentum in 2019, and to reach $3bn USD by 2022.

According to Snell, the growth of HPC cloud computing can be attributed to two factors:

  1. Extensive use of machine learning in the cloud, since ML-based applications are predisposed to be used in the cloud by default;
  2. Ease of license migration to the cloud, since licensing models have matured, and are now providing user-friendly options to developers.

The total amount of cloud HPC jobs is to go way up by 15-20% by 2020.

HPC Space Will Continue to Grow

According to Earl Joseph, CEO of Hyperion Research, high-performance computing grew by 10% to reach $24.3bn USD in 2017.

The sector is projected to keep growing by 10% year-on-year to eventually reach $38.4bn USD in 2022.

The growth will be driven by AI & ML, high-performance data analytics, HPC in the cloud, and government spending on the next-gen exascale supercomputer systems.

Below is Addison Snell’s forecast for HPC Product & Services (up to 2022):

Conclusion

International Supercomputing Conference High Performance 2018 was an amazing event, and we are proud to have participated in it.

Not only did we showcase SQUADEX and FlyElephant as highly proficient service providers in 

ML, HPC, and Big Data spaces, but also analyzed major trends that would impact an HPC landscape for years to come.

Specifically, artificial intelligence, machine learning, and HDPA will continue to converge and reinforce each other; HPC in the cloud will keep 

growing, both service- and job-wise; HPC spending and revenues will gain mom

entum, with a steady growth of 10% yearly up to 2022.

Additionally, the advance of AI, ML, and Big Data in the HPC space may result in a considerable update of Linpack metrics, with a focus on more flexible, real-world indicators of performance.

SQUADEX will keep you updated about upcoming conferences and events.

Stay tuned!