Arm unveiled its Total Compute strategy, following on the company’s recently announced v9 chip architecture and including its first Cortex CPUs and Mali GPUs based on that architecture.
Paul Williamson, senior vice president and general manager, Client Line of Business, Arm, told gathered media during an online briefing that the v9 architecture “set the foundation for the next decade of purpose-built computing,” and that the Total Compute strategy and new CPUs and GPUs will usher in vast performance improvements for devices such as mobile smartphones, wearables, smart TVs, laptop PCs and IoT devices.
“Our Total Compute strategy is built on three pillars -- performance, accessibility to that performance and security,” he said.
The CPU portion of the launch includes the Arm Cortex-X2, which Williamson said will deliver 30% greater peak performance in v9 CPU clusters than Arm’s current X1 CPU enabled in Arm v8.2 CPU clusters. That translates to higher performance in devices such as smartphones that likely would include the X2 starting next year.
Also joining the Cortex family are new “big” and “LITTLE” core processors (referring to the company’s big.LITTLE heterogeneous computing concept.) The Cortex-A710 is the first Armv9 “big” CPU, and is capable of a 30% improvement in energy efficiency and 10% uplift in performance compared to the previous Cortex-A78. This means longer battery life for mobile devices, even when running the most performance-hungry applications, Williamson said.
In the “LITTLE” core processor segment, there is the new Cortex-A510, Arm’s first high-efficiency “LITTLE” core in four years, and which can deliver a 35% performance increase and more than three-times performance for machine learning applications.
Meanwhile, Arm also is focusing on boosting the visual experience of mobile device users with new GPUs, including the Arm Mali-G710, which Williamson said will provide a 20% performance improvement for high-fidelity gaming and other compute-intensive experiences in devices like smartphones and Chromebooks. The G710 also offers a 35% uplift for machine learning tasks.
Patrick Moorhead, principal analyst at Moor Insights & Strategy, told Fierce Electronics via email that Arm’s ability to balance machine learning improvements with across-the-board performance improvements stood out to him. “These new announcements give us one more click in how Arm is raising performance and securing its designs better. I think it's interesting how the company’s new design better leverages ML capabilities while increasing base CPU and GPU performance,” he said.
Also in the GPU segment, Arm unveiled the Mali-G610, which is a lower-priced but full-featured version of the G710, and the even more energy-efficient Mali-G510 for devices like mid-range smartphones. Finally, there’s the Mali-G310, aimed at delivering high-quality graphics technologies to lower-cost devices such as entry-level smartphones, augmented reality devices and wearables.
Between the all the performance and machine learning improvements in the new CPUs and GPUs, users of devices that incorporate these designs will see noticeable benefits, Moorhead said. “We will see the CPU, GPU and machine learning capabilities the second the first product launches with its designs. Device UI should be snappier, games will play better and collaboration should be smoother. Developers don’t need to specially code for anything different,” he said.