Datacenter Trends

Intel, Lenovo Expand Datacenter Partnership to AI and HPC

• By John K. Waters
• 08/07/2019

Intel and Lenovo announced a multiyear collaboration agreement this week, building on their longstanding partnership in the datacenter.

The two companies will be focusing on "the rapidly growing opportunity in the convergence of high-performance computing (HPC) and artificial intelligence (AI)," they said in a statement. The plan is to accelerate the convergence of HPC and AI "to unlock new levels of customer insight."

"Intel is laser-focused on helping our customers spur innovation and discovery through the convergence of AI with HPC," said Navin Shenoy, executive vice president and general manager of Intel's Data Center Group. "Our extended collaboration with Lenovo combines the best of both companies' innovations to drive our customers' progress forward even faster."

The two companies plan to bring together a range of technologies, including:

• Lenovo's TruScale Infrastructure Services (its "pay-for-what-you-use" datacenter) and Intel's X^e compute architecture
• Intel Optane DC persistent memory
• Intel's One API programming framework
• Both current and future generations of Intel Xeon Scalable processors, which are CPUs with the built-in AI acceleration of Intel Deep Learning Boost (which provides enhanced AI inference)

This partnership will enable the two companies to "cascade" breakthrough HPC and AI technologies to users of any size -- what Lenovo calls "from exascale to everyscale." Exascale computing refers to computing systems capable of at least 1 quintillion (a billion billion) calculations per second.

"Our goal is to further accelerate innovation into the Exascale era, aggressively waterfalling these solutions to scientists and businesses of all sizes to speed discovery and outcomes," said Kirk Skaugen, executive vice president and president of Lenovo's Data Center Group, in a statement.

Lenovo will be optimizing Intel's full portfolio of HPC and AI hardware and software solutions. A key focus area will be building out Lenovo's smarter software offerings, including optimizing Lenovo's Intelligent Computing Orchestration (LiCO) HPC/AI software stack for Intel's next-generation technologies and alignment with the Intel One API programming framework.

Additionally, the collaboration will enable distributed asynchronous object storage (DAOS), which is the foundation of the Intel exascale storage stack. DAOS advanced storage frameworks and other exascale-class software optimizations will help HPC and AI users run their applications with greater ease, the companies said.

Both Intel and Lenovo are also aiming to add other partners to the mix to create a new ecosystem -- they're calling it "ecosystem enablement" -- for the convergence of HPC and AI. This includes building joint HPC and AI Centers of Excellence around the world to foster research and encourage university centers to develop solutions that address global issues like genomics, cancer, weather and climate, and space exploration.

Their earlier decision to combine Intel's 2nd Gen Xeon Scalable platform with Lenovo's Neptune liquid cooling technology -- a joint engineering effort utilizing a unique combination of HPC IP from the two companies -- has produced what can be fairly called remarkable results. According to Lenovo, 173 of the world's 500 fastest supercomputers in 19 markets run on its servers, and 17 of the world's top 25 research universities rely on Lenovo infrastructure.

"Lenovo's Neptune liquid cooling, in combination with the 2nd Gen Intel Xeon Scalable platform, helps customers unlock new insights and deliver unprecedented outcomes at new levels of energy efficiency," Skaugen said.

Lenovo has built its datacenter business on its 2014 purchase of IBM's x86-based server business. The datacenter accounts for about a 10th of the Chinese PC maker's overall business, but revenues have begun growing at a brisk pace. According to IDC, the company ranked fourth in the worldwide server market at the end of 2018 (behind Dell, HPE/New H3C Group and IBM, and tied with Inspur Power Systems), with a market share of 6.2 percent. But it generated $1.46 billion in server revenue in the fourth quarter of last year, which represents 34 percent growth year over year.