Design cycles are shortened from

weeks to hours with in-house AI tools

 

 

Meteor Lake processor design is accelerated by augmented AI, which will be used in upcoming client chip families.

For many years, it took equal parts science and art to figure out where on Intel’s client CPUs to precisely place heat-sensitive sensors.
When determining the optimal location for thermal sensors on the central processor units (CPUs) used in contemporary laptops, circuit designers would consult historical data. To pinpoint the specific location of hotspots where outbreaks typically occur, they would also rely on experience. Up to six weeks of testing, simulating workloads, and perfecting sensor placement could be needed for this laborious dance, after which the procedure would be repeated.

Thanks to an internal tool created by Intel engineers, system-on-a-chip (SoC) designers no longer have to wait six weeks to find out if they have struck the sweet spot for a sensor. The responses they receive come in minutes.
One Tiny Step for AI, One Massive Advancement for Silicon Design
The technology assists Intel’s system architects in accounting for thousands of variables in upcoming silicon designs. It was created by the Augmented Intelligence team, which is directed by Dr. Olena Zhu, senior principal engineer and AI solution architect in Intel’s Client Computing Group (CCG). This is just one instance of how Intel teams are using their expertise in AI to optimize a range of workloads.

A subclass of artificial intelligence called augmented intelligence focuses on human-machine collaboration.
Laptops and other client products mostly depend on peak and turbo frequencies. The goal is for the SoC to explode at higher frequencies, which produces thermal heat, according to Mark Gallina, senior system thermal and mechanical architect and lead engineer at CCG.
He describes how, in order to locate thermal hotspots correctly, engineers must carefully examine intricate, concurrent workloads that engage the CPU core, input/output (I/O), and other system operations. Determining the optimal location for the minuscule temperature sensors each no bigger than the tip of a pin complicates the procedure.

Gallina states, “We are only able to look into one or two workloads at a time, and that process takes a few weeks.”
That guesswork is eliminated by Intel’s new augmented intelligence capability. When engineers enter their boundary conditions, the program analyzes thousands of variables and provides optimal design recommendations in a matter of minutes.
The tool was utilized by engineers working on the SoC designs for the Intel® CoreTM Ultra mobile processor family (Meteor Lake) the Intel Core Ultra family was introduced on December 14 and it will be used for next client products like Lunar Lake and its offspring, which will expand the range of laptops available in the AI PC class.

More AI: Using Augmented AI to Identify Thermal Workloads and Improve Silicon Design
A companion tool that quickly identifies important thermal workloads was also developed by Olena Zhu and team member Ivy Zhu, principal engineer and AI solution architect.
According to Olena, the process is as follows: her group uses simulations or measurements of a limited set of workloads to train AI models. Then, using workloads that Intel is neither simulating nor measuring, these AI models forecast other workloads.

When used in tandem, these two augmented intelligence techniques improve how engineers optimize silicon architecture for Intel’s upcoming chip families, which include the client processors that will drive the upcoming AI PC generation.
Without a certain, augmented intelligence is not going to replace real engineers anytime soon, even though both technologies are useful.
“With augmented intelligence, we’re identifying the best areas to invest our limited resources by combining computational machine learning with human engineering expertise,” explains Gallina.
The way we do thermals now has been totally transformed by this new technology. Before we switch on the SoC, it provides us with much greater visibility into thermal hazards and is far more efficient. Through the use of augmented intelligence, we have been provided with a flashlight to help us navigate in the dark.

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.