More AI: IDing Thermal Workloads with Augmented AI to Optimize Silicon Design

Members of the Intel Augmented Intelligence team — from left: Mark Gallina, Olena Zhu and Michael Frederick – pose in a Client Computing Group lab in Hillsboro, Oregon. (Credit: Intel Corporation)

Olena Zhu and team member Ivy Zhu, principal engineer and AI solution architect, also developed a companion tool that rapidly identifies critical thermal workloads.

Olena says it works like this: Her team trains AI models based on simulations or measurements of a small number of workloads. These AI models then predict other workloads that are not being simulated or measured by Intel.

Together, both augmented intelligence tools boost the way engineers optimize silicon design for Intel’s forthcoming chip families, including client processors that will power the next generation of AI PCs.

While both tools are helpful, make no mistake, augmented intelligence isn’t replacing actual engineers anytime soon.

“With augmented intelligence, we’re using computational machine learning combined with human engineering expertise to identify the best areas to invest our limited resources,” Gallina says.

“This new tool has completely revolutionized the way we do thermals today. It’s so much more efficient and gives us so much more visibility to thermal risks before we turn the SoC on. We’ve been feeling our way in the dark, but with augmented intelligence, we’ve been given a flashlight to guide the way forward.”

Augmented Intelligence Helps ‘Find the Needle in the Haystack’

Olena’s aha moment came a few years ago when she realized the rapid advancement of AI investments opened new doors to how we design.

“Augmented intelligence leads to a new breed of tools that allows us to manipulate data much more efficiently than ever before,” Olena says. “When we combine AI with our existing engineering bench strength, we can find the needle in the haystack much more efficiently.”

Thanks to Olena and her team, engineers across Intel are embracing AI. CCG’s Augmented Intelligence team continues to find ways AI can speed up hardware and software design. Consider these recent examples:

• A fast and accurate signal integrity analysis tool for high-speed I/O reduces design time from months to an hour. The bonus? Intel is the first in the industry to embrace this technique, which has supported generations of chip designs.
• An AI-based automatic failure analysis tool for high-speed I/O design, deployed since 2020, led to 60% efficiency gains.
• An augmented intelligence tool called “AI-Assist” uses an AI model to automatically determine customized overclocking values for different platforms. This reduces overclocking time from days to just one minute. AI-Assist is available on Raptor Lake Refresh machines. (Video: How AI Assist Uses Machine Learning to Make Overclocking Easy)
• An AI-based automatic silicon floor plan optimizer is incorporated into Intel’s SoC design flow.
• A smart-sampling tool to help power and performance engineers crunch smart design experiments reduced the number of testing cases by as much as 40%.
• A user-interactive tool builds AI models to predict the performance of architectural proposals and help answer CPU design trade-off questions.
• A new way to automatically place tiny board components drives down cycle time from days to hours.

Across Intel, other engineering teams are finding clever uses for AI across Intel’s wide product suite: An AI Intel® Thread Director algorithm that debuted in 13th Gen Intel® Core™ CPUs contributed to workload improvements of over 20%.

In another example, engineering teams shaved time taken to test individual processors by 50% thanks to a smart AI algorithm developed in-house.

“There’s a fast growing movement in the industry to infuse AI into similar engineering usages, and Intel is definitely taking advantage of it and embracing it,” Olena says.