قالب وردپرس درنا توس
Home / Technology / Arm announces its new premium CPU and GPU designs – TechCrunch

Arm announces its new premium CPU and GPU designs – TechCrunch



Arm, the company that designs the basic chip architecture for most smartphones in the world, today announced the launch of its next design suite for premium phones. It will be a while before you see the first phones that use chips based on this design. As a rule, the first actual chips are displayed before the end of the year. With this introduction, the company announces the Cortex-A77 CPU, the Mali G77 GPU and a more energy efficient and powerful machine learning processor.

In light of recent trends, it is no surprise that the new Cortex-A77 is not just focused on overall performance improvements, although the company's promise to improve its IPC performance by 20% in the last generation is no cause for concern , Thanks to a combination of hardware and software optimizations, the Cortex-A77 now promises significantly better machine learning performance.

Why focus on if the company also offers a machine learning processor? Arm argues that most smartphones today do not use a dedicated neural processor. In fact, the company argues that today, 85 percent of smartphones run machine learning with just one CPU or CPU + GPU combination. And even if an accelerator is available, the CPU must pass it to the accelerator, whether it's a GPU or a dedicated machine learning chip.

As with any new generation of arm CPUs, the Cortex A77 also promises to be more energy efficient and provide better raw processing performance. According to Arm, the performance has increased fourfold since 2013.

Another area that is on the alert is mobile games – and in a broader sense mobile VR and AR experiences. The new Mail G77 GPU architecture is the first to be based on the company's Valhall GPU design and promises a 1.4x performance improvement over the G76. It's also 30 percent more energy-efficient and – and I suppose you can spot a theme here – 60 percent faster in executing machine learning and neural network utilization inferences Trillium, its heterogeneous machine learning computing platform, based in Combination with the company's CPUs. Since Trillium's announcement last year, the company has been able to double its energy efficiency and scale performance to 8 cores and 32 TOPs, adding further software innovation, "notes the company in today's announcement. "For developers, the CPU is more critical than ever, because it does not just do general computing, but also does much of the device's ML calculations that need to be scaled beyond today's limits. The same applies to more immersive, unconnected AR / VR applications and HD games on the go. "


Source link