The ARM v9 processor architecture was introduced in the first quarter of 2021. It is based on ARM v8 with blocks and instructions for artificial intelligence, security, and specialized computing. ARM has now announced the next generation of ARM v9-based computing cores used in a wide range of devices, from smartphones and home appliances to wearable electronics. Additionally, the developer spoke about the updated line of Mali graphics systems.

Cores Cortex-X2, Cortex-A710, Cortex-A510

Cortex-X2, Cortex-A710, and Cortex-A510 are direct extensions of Cortex-X1, Cortex-A78, and Cortex-A55, respectively. ARM categorizes the new cores into categories: Cortex-X2 is called “efficient,” Cortex-A710 is called “large,” Cortex-A510 is called “small.”

Cortex-X2 is the second generation of flagship high-performance ARM cores. Compared to its predecessor, the Cortex-X2 will provide 16% better performance and 2x faster machine learning processing. The core is intended for use in premium smartphones and laptops.

The new “big” core has moved to a more mid-range position with the introduction of the Cortex-X. The performance improvement of the Cortex-A710 against the background of the Cortex-A78 will be 10% at the same clock speed. Promises 30% more energy efficiency.

YOU CAN ALSO READ:   Fake Site was distributing Fake Afterburner with a Virus

For the first time in the past four years, ARM has developed a completely new “small” core. The upcoming Cortex-A510 will offer 35% faster performance than the Cortex-A55, 20% better energy efficiency, and 3x faster machine learning processing. ARM says the Cortex-A510 is on par with the Cortex-A73, the “big” core found in its 2017 flagship smartphones.

As before, “large” and “small” cores within one processor can work independently of each other using the DynamIQ Shared Unit technology. Moreover, with the ARM v9 architecture, there will be a DSU-110 layout that supports up to eight Cortex-X2 cores in a cluster. It targets laptops, desktops, and even servers. The L2 cache will double to 8MB, reducing data access latency.

Write A Comment