AMD revealed today a set goal to bring thirty times more energy efficiency for their EPYC CPUs and Instinct accelerators for AI training and HPC, or High-Performance Computing, programs that are processed by accelerated computational nodes. The expected start date would be no sooner than the year 2025. This includes AMD’s high-processing CPUs, their efficient and powerful GPU accelerators which they utilize for AI training, and HPC accelerated CPU configurations.
“Achieving gains in processor energy efficiency is a long-term design priority for AMD and we are now setting a new goal for modern compute nodes using our high-performance CPUs and accelerators when applied to AI training and high-performance computing deployments. Focused on these very important segments and the value proposition for leading companies to enhance their environmental stewardship, AMD’s 30x goal outpaces industry energy efficiency performance in these areas by 150% compared to the previous five-year time period.”
— Mark Papermaster, Executive Vice President and CTO, AMD
To achieve this goal, AMD would need to increase the power efficiency of their computational nodes at such a rate that it would have to be more than 2.5 times faster than the standard set by the industry over the last five years.
“With computing becoming ubiquitous from edge to core to cloud, AMD has taken a bold position on the energy efficiency of its processors, this time for the accelerated compute for AI and High Performance Computing applications. Future gains are more difficult now as the historical advantages that come with Moore’s Law have greatly diminished. A 30-times improvement in energy efficiency in five years will be an impressive technical achievement that will demonstrate the strength of AMD technology and their emphasis on environmental sustainability.”
— Addison Snell, CEO of Intersect360 Research
Accelerated compute nodes are extremely powerful, as well as also being extremely advanced. In fact, they are the most advanced systems in the world. Accelerated compute nodes are used for research and supercomputer tests that most standard systems would not be able to process. Scientists utilize accelerated computational nodes to create discoveries and breakthroughs in several fields, such as climate estimations and alternative energy solutions. When talking about AI, accelerated compute nodes allow for studies of neural networks studying “speech recognition, language translation and expert recommendation systems, with similar promising uses over the coming decade.” AMD’s plan would save several billions of kilowatt-hours of electricity by the year 2025. In fact, the reduction of utilized power would be able “to complete a single calculation by 97% over five years.”
“The energy efficiency goal set by AMD for accelerated compute nodes used for AI training and High Performance Computing fully reflects modern workloads, representative operating behaviors and accurate benchmarking methodology.”
— Dr. Jonathan Koomey, President of Koomey Analytics
AMD has always studied ways to simplify energy output. Fortune Magazine recently added AMD to their “Change the World” list last for 2020. The magazine’s list showcases efforts made by companies that are striving to meet and go beyond the needs of society. AMD’s transparency about its environmental achievements has been in practice for over 25 years. These new goals that AMD is setting for themselves are part of the Environmental, Social, Governance plan, which is utilized in every aspect of the company.
AMD also plans to use their “segment-specific datacenter power utilization effectiveness (PUE) with equipment utilization taken into account.” Their power consumptions for both CPUs and GPUs are set on specific segment utilization percentages—both active and idle—”then multiplied by PUE to determine actual energy use for calculation of the performance per Watt.” AMD’s baseline for power consumption uses the industry energy rates calculated between 2015 to 2020, and are then “extrapolated to 2025.” The measure of energy per operation improvement for the next five years is then measured by estimated global volumes, and multiplied by the TEC, or Typical Energy Consumption, of each segment to find the actual energy utilization worldwide.