Arm Announces Cortex-A78AE CPU, Mali-G78AE GPU & Mali-C71AE ISP for autonomous automotive & industrial applications

Arm has announced new CPU, GPU, and ISP specifically designed for autonomous automotive and industrial applications with respectively Cortex-A78AE CPU, Arm Mali-G78AE GPU, and Arm Mali-C71AE ISP.

Arm Cortex-A78AE CPU

Cortex-A78AE

Key features and specifications:

  • Architecture – Armv8.2-A (AArch32 at ELO only)
  • Extensions – Armv8.1, Armv8.2, and Armv8.3 extensions (LDAPR instructions only), RAS extensions, Armv8.4 Dot Product, Cryptography extensions, RAS extensions
  • Microarchitecture
    • Up to 4x CPU cores per cluster
    • Out of order pipeline
    • Neon / Floating Point Unit included with INT8 Dot Product and IEEE FP16
    • Optional Cryptography Unit
    • 48-bit Physical Addressing (PA)
  • Memory system and external interfaces
    • 32kB to 64kB L1 I-Cache / D-Cache
    • 256kB to 512kB L2 Cache
    • Optional 512kB to 4MB L3 Cache
    • ECC Support
    • LPAE
    • Bus interfaces – AMBA ACE or CHI
    • Optional ACP, peripheral port
    • Functional Safety Support – ASIL D Systematic1 and ASIL D Diagnostic2
    • Security – TrustZone
    • Interrupts – External GICv4
    • Generic timer – Armv8-A
    • Non-intrusive debug (trace)

Cortex-A78AE succeeds Cortex-A76AE processors and delivers 30% performance improvement at the same power efficiency, 25% increased performance at the same total power, or 60% reduction in power at the same performance level.

Cortex-A78AE vs Cortex-A76AE

Cortex-A78AE supports DynamIQ, and the high-performance cores would typically be paired with Cortex-A65AE low-power cores in an automotive SoC. You’ll find more details on the developer page, and Arm community’s blog post.

Arm Mali-G78AE GPU

Arm Mali-G78AE Automotive GPU

Mali-G78EU is the very first autonomous/automotive GPU with from the company, is designed for functional safety, and offers the following key features:

  • Configurable from 1 to 24 cores
  • Anti-aliasing – 4x/8x/16x MSAA (Multi-Sampling Anti-Aliasing)
  • API support
    • OpenGL ES 1.1, 2.0, 3.1, 3.2
    • Vulkan 1.1, 1.2
    • OpenCL 1.1, 1.2, 2.0 Full profile
    • OpenGL SC 1.0, 2.0
  • Bus interface – AMBA 5 ACE, ACE-LITE, and AXI
  • L2 cache – Configurable 256KB – 4MB
  • Adaptive Scalable Texture Compression (ASTC) – Low Dynamic Range (LDR) and High Dynamic Range (HDR); supports both 2D and 3D images.
  • Arm Frame Buffer Compression (AFBC) – Version 1.3

The GPU supports automotive and industrial safety standards to help manufacturers meet ASIL B / SIL 2 requirements, and adds features like Flexible Partitioning and virtualization to isolate various graphics & compute tasks and increase reliability.

Arm GPU Flexible Partitioning
12-core Mali-G78AE MC12 GPU with different workloads

Again you’ll find more details both on Arm’s developer page and corresponding blog post.

Arm Mali-C71AE ISP

Mali-C71AE ISP

Autonomous vehicles and robots requires penty of camera inputs, so it makes perfect sense for Arm to introduce the Mali-C71AE Image Signal Processor (ISP) with support for up to four real-time cameras or 16 buffered cameras with a total throughput of 1.2 GP/s (Giga pixel per second).

The ISP also supports advanced spatial (2D) noise reduction (“Sinter”), per-exposure noise profiling and chromatic aberration correction, integrates 400 fault-detection circuits to make sure data is reliable, and provides supports for independent dynamic range management both for optimum human vision (display) and for computer vision (CV) applications. More in-depth information about Mali-C71AE ISP can be found in the announcement’s post.

The new Arm Mali-78AE processor, Mali-G78AE GPU and Mali-C71AE ISP should soon be found in SoCs used in applications when functional safety is critcal from smart manufacturing to enhancing ADAS and digital cockpit applications in automotive.

Share this:

Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress

ROCK Pi 4C Plus
Subscribe
Notify of
guest
The comment form collects your name, email and content to allow us keep track of the comments placed on the website. Please read and accept our website Terms and Privacy Policy to post a comment.
0 Comments
Khadas VIM4 SBC