Specifications

The first generation Armv9 "big” Cortex CPU. The Cortex-A710 CPU delivers the best balance of performance and power efficiency for industry-leading multiple form factors.

Cortex-A710 provides Armv9 features in performance, security and machine learning at much improved power efficiency for multiple markets. Paired with Cortex-X2 CPU and the Cortex-A510 CPU in a scalable DynamIQ big.LITTLE configuration, Cortex-A710 brings an uplift of up to 10% in performance and up to 30% in energy efficiency over Cortex-A78.

Arm is focused on delivering a Total Compute solution for our partners, including enhanced Cortex-X and Cortex-A CPUs, Mali GPUs, Ethos NPUs, CoreLink and CoreSight System IP. Cortex-A710 provides the end user with a performant device with multi-day battery life.

Arm Cortex-A710 CPU diagram

Arm Cortex-A710 CPU 

Architecture Armv9-A (Harvard)  
Extensions

Up to Armv8.5 extensions
SVE2 extensions
Memory tagging extensions (MTE)
Cryptography extensions
RAS extensions

 
ISA support

A64
A32 and T32 (at EL0 only) 

 
Microarchitecture
Pipeline Out of order
Superscalar
Yes
Neon / Floating Point Unit / SVE2 Included
Cryptography Unit
Optional
Max number of CPUs in cluster
8
Physical Addressing (PA) 40-bit
Memory system and external interfaces
L1 I-Cache / D-Cache 32KB or 64KB
L2 Cache 256KB or 512KB
L3 Cache Optional, 256KB to 16MB
ECC Support Yes
Bus interfaces            AMBA AXI5 or CHI.E
ACP Optional
Peripheral Port Optional
Other  Security TrustZone, Secure-EL2
Interrupts GIC interface, GICv4.1
Generic timer Armv9-A
PMU PMUv3
Debug Armv9-A
CoreSight CoreSightv3
Embedded Trace Extension ETEv1.0
Trace Buffer Extension  Yes

Key benefits of Cortex-A710

  • Supports Armv9 features across performance, security and machine learning. Key features are SVE2, MTE, PAC, BTI, Secure EL2, support for BFloat16 format and Matrix Multiply instructions for the Int8 and BFloat16
  • CPU performance uplift of up to 10% in the same power envelope over Cortex-A78
  • Improvement of up to 30% in CPU power consumption at the same performance footprint
  • Machine learning workloads benefit from up to a 200% increase over Cortex-A78