When Will GPUs Go Mainstream in the Enterprise?
Leading companies are embracing GPUs to deliver performance for modern enterprise strategy and take their analytics to the next level.
- By William McKnight
- August 25, 2017
If you're looking for upcoming trends in computing resources, look west towards Silicon Valley and see what Google, Microsoft, Amazon, Tesla, eBay, and their ilk are doing.
Coming to light lately is their investment in graphics processing units (GPUs). GPUs have made the jump from controlling monsters, zombies, fast cars, and aliens to the supercomputers of the data leaders and it's only a matter of time until they'll be a trend for the enterprise.
Why GPUs?
When we think about enterprise computing challenges, the processing unit architecture (beyond cores and computing resources) should be added to the usual suspects we tweak, such as storage, memory, clustering, and sharding. None of these can be considered the "go to" for general tuning, and no improvement comes without cost and consequences. Multicore CPUs, for example, are not a completely free lunch because code must be parallelized to take advantage of them.
If you accept the value proposition of deep learning, GPUs are a natural part of the strategy to implement it. Acting more like a human brain, GPUs are a major improvement on the CPU idea of combining processing units.
GPU advantages are evident when performing the math and algorithms inherent in machine learning, but perhaps the entry point into the enterprise will be advanced analytics. A GPU does the processing of dozens of CPU servers with less overhead -- something clearly applicable to advanced analytics.
We're bringing more data to bear on our analytics, and with the increased use of the data lake, our data structures are becoming more unwieldy. If data leadership is our strategy, we need to process massive amounts of data continually analyzed together. Speed kills it when it comes to these queries. Great performance means more data utilization, which means higher achievement of data leadership. Advanced analytics in support of data leadership is where GPUs enter the enterprise, and soon.
One Example in Use
Consider the Nvidia DGX-1 system. Though Nvidia chips can be found in Lenovo, Dell, HP, and other systems, Nvidia also offers its own prebuilt supercomputer, the DGX-1.
The DGX-1 system has eight 16GB Tesla GPUs, a 7TB SSD, and a pair of Xeon processors. Each GPU provides 10.6 teraflops of single precision floating point performance. The DGX-1 is a way to get Nvidia Pascal-based Tesla compute into the enterprise and GPUs operating on enterprise advanced analytics.
GPUs deliver the performance that the modern enterprise needs. Leading companies are moving there already and taking their analytics to the next level.
About the Author
McKnight Consulting Group is led by William McKnight. He serves as strategist, lead enterprise information architect, and program manager for sites worldwide utilizing the disciplines of data warehousing, master data management, business intelligence, and big data. Many of his clients have gone public with their success stories. McKnight has published hundreds of articles and white papers and given hundreds of international keynotes and public seminars. His teams’ implementations from both IT and consultant positions have won awards for best practices. William is a former IT VP of a Fortune 50 company and a former engineer of DB2 at IBM, and holds an MBA. He is author of the book Information Management: Strategies for Gaining a Competitive Advantage with Data.