Print Save PDF

About 7 minutes

In our last article, we talked about CIO priorities for 2018.

For the coming year, innovation claims the top spot for CIO priorities. Companies are now aware that if they don’t innovate, they will face possible extinction.

IBM Power9: The Relationship Between Application Performance, Storage, and Innovation


Bonus read: An Introduction to IBM POWER9

The key to innovation and survival in business can be found in an organization’s data.

As we discussed in our CIO priorities article, high-performing companies use their data to help them innovate and improve their products and services.

Data helps companies see market trends and understand customer behavior.

The insights provided by data analytics can lead to innovative new products and services. 

Modern data analytics are driven by machine learning systems that study data, look for patterns, and provide valuable insights.

According to the Harvard Business Review, machine learning systems (AI) will have a transformational impact on how companies do business, and most of the big opportunities have yet to be exploited.

In the next decade, the effects of cognitive computing will be magnified as organizations transform their business models and core processes.

Businesses have been slow to implement machine learning, which leaves attractive openings for startups, or new opportunities for established companies.

To take the best advantage of these opportunities, companies need competitive business intelligence capabilities.

What Makes for Innovation Success?

Because innovation is now closely tied with big data analytics, the companies that have the best technology are most likely to win.

The faster, more accurately a company can access and derive insights from its data, the more quickly it can develop products and services in response to shifts in the market or customer preferences.

Organizations need high-performance storage to truly use data-driven personalization, or real-time analytics.

Storage performance has been stagnant for the past 15 years, while networking and compute have experienced significant performance gains. 

However, recent advancements in software-defined and flash storage have put high-performance analytics within reach for organizations across the size spectrum.

IBM Unveils New Family of Processors Designed for Modern Workloads

While many workloads are moving to the cloud, high-performance, analytics-intensive applications should stay in the data center near the data they analyze.

As more enterprises become dependent on business intelligence, the demand for performance increases.

Companies are struggling to get the performance they need because CPUs are reaching the limits of Moore’s law.

To compensate, data centers have begun to use FPGAs and GPUs for key workloads. 

IBM has recognized the need for accelerated computing and developed the POWER9 solution.

This year, IBM is debuting the POWER9 with a new CPU created to handle the high-performance demands and workloads of the future.

Here are some of the performance features and capabilities of the IBM POWER9 family of processors:

Enterprise Scale

  • — Leading readability, availability, and serviceability (RAS)
  • — Flat scale-up systems
  • — Improved caching
  • — Buffered memory for increased capacity

High Performance Computing

  • — Advanced CPU/GPU memory sharing and interaction
  • — Direct attach memory with high bandwidth
  • — Highest bandwidth GPU attach

Emerging AI, Cognitive, and Analytics

Cloud and Hyperscale Data Centers

  • — State of the art IO technology for storage and network performance
  • — Virtualization features, including QoS, interrupt, power management, and storage performance
  • — Cost, power, and packaging options for a range of platforms

IBM offers two types of POWER9: Scale-out (SO) and Scale-up (SU).


Image source: See Jeff Stuecheli's full IBM POWER9 presentation here. Read all about the highlights via IBM Systems Magazine

The scale-out variations are designed to integrate with traditional datacenter clusters that have single and dual socket setups. They feature direct attach memory and 8 direct DDR4 ports.

The scale-up variations are designed for NUMA servers that have four or more sockets supporting large throughput and memory.

There are two types of scale-up dual memory subsystems: the 12-core SMT8 model and 24-core SMT4 model. The SMT4 is designed and optimized for the Linux ecosystem, and the SMT8 is optimized for the PowerVM ecosystem.

Accelerate Machine Learning

The POWER9 family of processors is ready to handle the high-performance demands of present-day and future machine learning systems.

IBM and its partners recently laid out a roadmap for power chips through 2020. Rated at 120 teraflops, these new chips are powerful enough to drive the world’s most powerful supercomputers.

Oak Ridge Leadership Computing Facility (ORNL) is in the process of developing a new supercomputer, SUMMIT, that will be built on a hybrid architecture and IBM POWER9 CPUs.

When it is complete, ORNL says it will be the smartest supercomputer in the world.


Modern innovation is data-driven.

Companies that want to stay competitive need real-time data analysis capabilities. With business intelligence applications, organizations can keep track of internal performance metrics, current market conditions, and customer preferences.

With real-time insights, business leaders can discover new product or service ideas that will appeal directly to their customer base.

IBM’s new breed of CPUs provides the performance and speed needed for real-time data analysis. Most companies have both structured and unstructured data.

One of the biggest BI challenges has been finding a way to integrate and analyze both data types and turn them into meaningful insight.

Cloud Services Solutions

Most, if not all, companies have data in the cloud and in on-location datacenters.

IBM is making it easy to integrate all your data, no matter where it lives.

Through IBM’s combination of high-performance POWER9 processors and seamless cloud services, you can infuse your organization with cognitive systems that learn your insights and improve them over time.

Many companies have already recognized the potential of the new POWER9 processors and are developing products and solutions built on POWER9 CPUs.

For example, Raptor Computing Systems announced its Talos II line of mainboard bundles, development kits, and secure workstations. The Talos II product line is one of the first to come to market with the IBM 14nm POWER9 processor.

Data analytics, machine learning, and cloud computing are making business innovation possible on a whole new level.

Companies that invest in high-performance storage and applications stand the best chance of gaining the kinds of insights that lead to market-changing, disruptive, and profitable new innovations.

Written by IBM BP Network