Ampere Partners with Qualcomm to Launch Arm-based AI Server | TechCrunch - Latest Global News

Ampere Partners with Qualcomm to Launch Arm-based AI Server | TechCrunch

Ampere and Qualcomm aren’t the most obvious partners. Finally, both offer ARM-based chips to power data center servers (although Qualcomm’s biggest market is still the mobile market). But as the two companies announced today, they are now joining forces to offer an AI-focused server that uses Ampere’s CPUs and Qualcomm’s Cloud AI100 Ultra AI inference chips to run – not train – models.

Like every other chip manufacturer, Ampere wants to benefit from the AI ​​boom. However, the company’s focus has always been on fast and energy-efficient server chips. While it can leverage Arm IP to add some of these features to its chips, this isn’t necessarily a core competency. “That’s why Ampere decided to work with Qualcomm (and SuperMicro to integrate the two solutions),” Arm CTO Jeff Wittich tells me.

“The idea here is that while I’m showing you great performance for Ampere CPUs, where AI inference is just running on the CPUs, you want to scale up to even larger models – for example, models with tens of billions of parameters – exactly Like all the other workloads, there is no one-size-fits-all solution for AI,” Wittich told TechCrunch. “We worked with Qualcomm on this solution, combining our super-efficient Ampere CPUs to do a lot of the common tasks you do related to inference, and then using their really efficient cards, we have one.” Server level solution.”

Regarding the partnership with Qualcomm, Wittich said Ampere wants to put together best-in-class solutions.

“[R]“We’ve had a really good collaboration with Qualcomm here,” he said. “That’s one of the things we’ve been working on. I think we have a lot of really similar interests, which is why I find that really compelling. They really develop very efficient solutions for many different market areas. We’re developing really, really efficient solutions on the server CPU side.”

The Qualcomm partnership is part of Ampere’s annual roadmap update. Part of this roadmap is the new AmpereOne chip with 256 cores, which is manufactured using a modern 3nm process. These new chips aren’t quite widely available yet, but Wittich says they’re ready at the factory and should hit the market later this year.

In addition to the additional cores, the defining feature of this new generation of AmpereOne chips is the 12-channel DDR5 RAM, which enables Ampere’s data center customers to better tailor their users’ memory access to their needs.

However, the selling point here is not just performance, but also power consumption and the cost of running these chips in the data center. This is especially true for AI inference, where Ampere likes to compare its performance to Nvidia’s A10 GPU.

It’s worth noting that Ampere isn’t abandoning any of its existing chips in favor of these new ones. Wittich emphasized that even these older chips still have numerous use cases.

Ampere also announced another partnership today. The company is working with NETINT to develop a joint solution that couples Ampere’s CPUs with NETINT’s video processing chips. This new server will be able to transcode 360 ​​live video channels in parallel while using OpenAI’s Whisper speech-to-text model to caption 40 streams.

“We started this journey six years ago because it was clear it was the right path,” Ampere CEO Renee James said in today’s announcement. “In the past, low power consumption meant low performance. Ampere has proven that this is not true. We have pushed the efficiency frontier of computing, delivering performance beyond traditional CPUs in an efficient computing environment.”

Sharing Is Caring:

Leave a Comment