In an interview with CRN, AMD’s Matt Unangst says channel partners have a ‘really important role’ to play in helping business customers understand the value of AI PCs and how its latest Ryzen Pro processors are enabling ‘best-in-class’ performance and battery life.

The head of AMD’s commercial PC business said the company’s latest Ryzen processors offer “best-in-class” performance and efficiency for the emerging category of AI PCs—a message he’s hoping channel partners will take the streets in its continued fight against rivals like Intel.

Matt Unangst, senior director of AMD’s commercial client and workstation business, made the comments in a recent CRN interview shortly before the company launched its Ryzen Pro 8000 series processors for business PCs earlier this month.

[Related: Why Intel Thinks 2024 Is A Big Commercial PC Refresh Year]

With the new processors expected in systems from HP Inc., Lenovo and other OEMs starting this quarter, Unangst said channel partners have a “really important role” to play in helping business customers understand the value of AI PCs and how AMD stands out in the space.

“We’ve actually had a lot of engagement and discussion and collaboration with channel partners across the board to make sure that they understand our value [proposition] and our capabilities and our [product] roadmap so that they can then go articulate to their customers where the industry is going, where the solutions are going,” he said.

The Ryzen Pro 8000 processors take the advances in general purpose and AI computing that came with the Ryzen 8000 chips that launched a few months ago and add enterprise-grade remote management and security features, such as AMD Memory Guard, for business customers.

The Santa Clara, Calif.-based chip designer is making a major commercial push for the chips as Intel attempts to defend its market share with a competing lineup for AI PCs, the Core Ultra series. At the same time, mobile chip designer Qualcomm is plotting a revitalized push into the PC market this year with its AI-focused Snapdragon X processors.

AMD’s recently launched processors include the Ryzen Pro 8040 series for AI-enabled commercial notebooks and Ryzen Pro 8000 series for AI-powered commercial desktops. The chips in both lineups feature a neural processing unit (NPU) in addition to a CPU and GPU, a key requirement from Microsoft for AI PCs running Windows 11, according to Intel.

How AMD’s Ryzen Pro 8040 Compare To Intel’s Core Ultra

Compared to Intel’s Core Ultra processors for laptops, AMD is claiming that the Ryzen Pro 8040 chips have advantages in several areas: multitasking, productivity applications, content creation applications, efficiency, and AI inference.

The company said it has similar advantages with its Ryzen Pro 8000 desktop chips when compared to Intel’s 13^th-Gen Core processors, which launched last year and don’t have an NPU. (Intel plans to launch its first line of desktop processors with an NPU later this year.)

“We can deliver best-in-class performance and best-in-class battery life, or energy efficiency, which is something that nobody else in the market can say,” said Unangst (pictured above).

AMD said it has proven this across several benchmark tests measuring general performance.

For instance, when compared to Intel’s 15-watt Core Ultra 7 165U for laptops, AMD said its 15-watt Ryzen 7 Pro 8840U is 18 percent faster on the Geekbench v6 single-core test for responsiveness, 19 percent faster on the Geekbench v6 multi-core test for multitasking and 43 percent faster on the Blender classroom for measuring CPU rendering performance, according to the company.

AMD said the Ryzen 7 Pro 8840U, which is designed for thin-and-light laptops, can even beat Intel’s Core Ultra 7 165H, which uses nearly double the power at 28 watts.

Across productivity benchmarks, AMD said the Ryzen 7 Pro 8840U has an even greater performance lead over the Core Ultra 7 165U, ranging from 67 percent advantages for Procyon Office’s Excel and PowerPoint tests to a 76 percent upper hand for Procyon’s Outlook test.

On the efficiency front, AMD claimed that the Ryzen 7 Pro 8840U can run a Microsoft Teams video conference call and Microsoft Office applications simultaneously 69 percent faster using 55 percent power compared to Intel’s Core Ultra 7 165H.

When NPU-enabled AI effects are turned on for Microsoft Teams in this workflow, AMD said the Ryzen processor becomes 72 percent faster and uses 84 percent less power.

In terms of battery life, AMD said a Ryzen 7 Pro 8840U-powered laptop can last for 7 hours and 21 minutes while running a Microsoft Teams video conference call. By contrast, AMD said, an Intel Core Ultra 7 165U-powered laptop can only last for 5 hours and 9 minutes doing the same thing.

When it comes to AI performance, AMD said the 15-watt Ryzen 7 Pro 8840U outperforms the 28-watt Core Ultra 7 155H anywhere from 4 percent for the MobileNetV3 computer vision model and 17 percent for the Deeplabv3 image segmentation model, to 44 percent for the Procyon AI Float32 benchmark and 79 percent for the YoloV8 computer vision model.

As for large language models, AMD didn’t have any performance data for its Ryzen Pro 8040 series but it did say that a processor from the previous lineup, the 15-watt Ryzen 7 Pro 7840U, enabled 14 percent more tokens per second and a 79 percent faster time to first token with a 7-billion-parameter Llama 2 model compared to Intel’s 28-core Core Ultra 7 155H. It

The older Ryzen chip also enabled 17 percent more tokens per second and a 41 percent faster time to first token for a 7-billion-parameter Mistal AI model, which AMD said is similar to the GPT 3.5 and GPT 4 models that power OpenAI’s ChatGPT.

The AI computing advantages of the Ryzen Pro 8040 series are made possible in part by the processors’ ability to achieve up to 39 tera operations per second (TOPS) for the entire chip and up to 16 TOPS for the NPU alone, according to AMD. By contrast, Intel’s Core Ultra processors top out at 34 TOPS total and 11 TOPS for the NPU, the company said.

“When you take that Ryzen AI technology and layer that on top of it, you’re taking what is already a leadership class product for either a notebook or desktop, and you’re applying that best-in-class AI technology, which, increasingly, we’re seeing more and more applications that can take advantage of that,” Unangst said.

While AMD has largely focused on Intel for competitive comparisons, the company knows Qualcomm could shake things up with its upcoming Snapdragon X processors,

“We anticipate that they’re going to have a strong set of capabilities. We take that very seriously,” the senior director said. “We’re constantly evaluating our roadmap and our technology. We feel very good about where we’re at, from the technology that we have and the ability to compete and win in both performance and battery life there.”

Unangst said AMD also feels “great” about its AI technology, “both what we have in market today as well as where that’s going to advance over the next couple of years.”

Engineering Work With ISVs A ‘Foundational Piece’ Of AMD’s Strategy

With software enablement being key to the success of the AI PC category, AMD said it’s working with more than 150 independent software vendors (ISVs) on AI-powered applications that can take advantage of its latest Ryzen processors.

These ISVs range from major players like Adobe, Cisco, Microsoft and Zoom to smaller players like BufferZone, Bitdefender, Splashtop and Nero. By contrast, Intel said it’s working with more than 100 ISVs on more than 300 AI-powered features for PCs powered by Core Ultra processors.

“The engagements, both in terms of the experiences and the number of specific ISVs that we’re working with, it’s growing and expanding on a consistent basis,” Unangst said.

Unangst said AMD is working with these ISVs at an engineering level, which he considers a “foundational piece” of the company’s work with developers.

“We’ve provided a lot of demo hardware for them to use as they prototype and optimize their applications starting with Ryzen 7040 and now more so with our latest [Ryzen] 8040,” he said.

For AMD, ISV enablement also involves close engagement with OEMs.

“We’ve also worked closely with our OEM partners to make sure that our strategy around who to engage with and what to focus on is aligned with them and makes sense given where they see the market going,” Unangst said.

A key aspect of AMD’s work with ISVs is the company’s development software, namely the AMD Ryzen AI Software, which allows developers to take pre-trained AI models and quantize them and run them in live applications using the open-source ONNX Runtime.

While the inclusion of the NPU in its latest Ryzen processors represents a new paradigm for AMD’s personal computing strategy, Unangst said the company’s software aims to help developers determine whether they should map AI features to the NPU or a traditional engine like the GPU.

“Because there actually are a number of AI workloads that might make more sense to run on a GPU. And then there’s others that make more sense to run on the NPU,” he said.

“And so we’re really not focused on how do you take every single AI experience and move it all the way to the NPU. We’re more focused on how do you take these experiences and move it to the right IP to deliver the best experience,” he added.