RVA23, Multi-cluster, Hypervisor and Android

Overview

AndesCore™ AX66 64-bit multicore CPU IP is a high-performance quad decode 13-stage superscalar out-of-order processor based on AndeStar™ V5 architecture. It supports RISC-V standard “G (IMA-FD)”, “C” 16-bit compression, “B” bit manipulation, Vector and Vector Cryptography, Hypervisor and AIA, and Andes performance enhancements. It features MMU for Linux based applications, TAGE branch prediction for accurate branch execution, 4 wide instruction decode, 8 independent functional pipelines (4 integer, 2 Fp/Vector, 2 load/store), level-1 instruction/data caches and private level-2 cache for low-latency accesses. The AX66 symmetric multiprocessor supports up to eight cores and a level-3 cache controller with instruction and data prefetch. Coherence Manger ensures data coherence among CPU accesses and IO transactions from cacheless bus masters. Other features include ECC for cache memory soft error protection, CHI interface for cache coherent multi-cluster, StackSafe™, and PowerBrake and WFI for power management.

AndeStar™ V5 Architecture

Key Features Benefits
RISC-V RV64 GCB + Vector/Vector Crypto, Hypervisor/AIA
  • State-of-the art ISA from latest developments in computer architecture
  • Industry standard and open architecture
  • RVA23 profile compliant
Andes Extended Instructions Andes exclusive performance and functionality enhancements
MMU and Sv39/Sv48 virtual memory translation For Linux and advanced operating systems with protection between kernel and user program
64-bit CPU architecture Enabling software to utilize the memory spaces far beyond the 4G byte limit of 32-bit CPUs
  • Machine (M), optional User (U) and Supervisor (S) Privilege levels without hypervisor
  • Machine (M), Hypervisor-Extended Supervisor (HS), Virtualized Supervisor (VS) and Virtualized User (VU) with hyervisor
Full privilege protections with or without hypervisor

CPU Core

Key Features Benefits
>9.5 specint2006/Ghz (estimated) Superior performance
13-stage out-of-order 4 wide superscalar pipeline Superior performance-efficiency, while allowing for high speeds

Extensive branch predication features

  • TAgged GEometric length (TAGE) algorithm
  • 2 level Branch Target Buffer (BTB)
  • Return Address Stack (RAS)
  • Most accurate branch prediction algorithm
  • Branch Target Buffer to speed up control codes
  • Return Address Stack to speeds up procedure returns

MMU (Memory Management Unit):

  • Sv39, Sv48 virtual-memory systems
  • Sv39, Sv48X4 with Hypervisor
  • 16/32-entry fully associative level1 ITLB/DTLB
  • Up to 1024 entry 4-way set-associative L2 TLB
  • Hardware page table walker
  • Virtual memory support for full address space and easy code/data sharing
  • Support for full-featured OS such as Linux
  • Protection of supervisor and user privilege
  • Hardware for fast address translation
  • Hypervisor support
Physical Memory Protection (PMP), configurable up to 16 regions Basic read/write/execute memory protection with minimum cost
16 programmable physical memory attributes (PPMA) regions

Configurable memory attributes:

  • Memory, I/O, None
  • Cacheable/Non-cacheable
  • Read/write/read & write allocate, no allocate
  • Access fault for non-existent regions

RISC-V Vector support with 128bit VLEN

  • RVV 1.0
  • Dual Vector ALU
  • Vector Crypto support
  • SIMD support for Video, Audio and AI
  • High performance AES, SHA and SM4, SM4 crypto support
Performance monitors Program code performance tuning
StackSafe™ hardware stack protection Easy identification of stack size threshold during development Hardware error detection of stack overflow and underflow at runtime
PowerBrake technology Performance throttling to digitally reduce power consumption

Memory Subsystems

Key Features Benefits

Level-1 I-Cache & D-Cache

  • Size: 32KB or 64KB
  • Cache line size: 64 bytes
  • Set associativity: 2-way or 4-way

Private Level-2 I/D Unified Cache

  • Configurable from 128KB to 1MB
  • 64-byte cache line size
  • 8-way, pseudo random replacement

Shared Level-3 cache

  • Configurable from 256KB to 32MB
  • 64-byte cache line size
  • 16-way, pseudo random replacement
Accelerating accesses to slow memories Further accelerate performance with optional larger 2nd level private cache Larger 3rd level share cache
Optional ECC error protection with SRAM interface Code and data integrity protection
Bus maneger port: AXI with 128 or 256 bit data, I/D joint or separate bus High throughput with wide data path
Core/bus clock ratio of N:1 Simplified SoC integration

Multicore Cache Coherence

Key Features Benefits
  • Support up to 8 cores
  • MESI cache coherence protocol
  • 128/256-bit I/O coherence port for cacheless bus managers
  • Symmetric multicore and cache controller with cache coherence
  • Convenient and efficient interface for SoCs with rich I/O transactions

Advanced Interrupt Architecture (AIA)

Key Features Benefits

Implements RISC-V AIA specification v1.0

  • Up to 1023 APLIC interrupt sources
  • APLIC-Direct Mode
  • APLIC-IMSIC mode
  • IMSIC support
  • Virtualized interrupt support for Hypervisor
  • Interrupt handling for SoC with multiple processors
  • message signaled interrupt support
  • Hypervisor support

Debug Support

Key Features Benefits
Implements RISC-V debug specification v1.0 Supported by industry debug tool suppliers
JTAG Debug Port Industry-standard support
Embedded Debug Module with up to 8 triggers Flexible configurations to tradeoff between gate count and debugging capabilities
Exception redirection support Entering debugger upon selected exceptions without using breakpoints
RISC-V Trace 1.0 Instruction Trace interface Supported by Andes tools

Key Features

  • 64-bit out-of-order 4 wide decode 13-stage CPU core with 128 reordering buffers and 8 functional pipelines
  • Symmetric multiprocessing up to 8 cores
  • Private L2 cache support
  • Level-3 shared cache and coherence support
  • AndeStar™ V5 Instruction Set Architecture (ISA)
    • Compliant to RISC-V GCBV and Vector Cryptography
    • Hypervisor(RVH)
    • RVA23 profile compliant
  • 64-bit architecture for memory space over 4GB
  • TAGE Branch predication for highly accurate prediction
  • Linux-capable Memory Management Unit (MMU)
  • Physical Memory Protection (PMP) and latest architecture enhancement extension (ePMP) for access permission controls
  • AIA with APLIC and IMSIC support
  • ECC or Parity for SRAM error protection
  • StackSafe™ hardware to help measuring stack size, and detecting runtime overflow/underflow
  • PowerBrake and WFI (Wait for Interrupt) for different power saving occasions

Block Diagram

RVA23, Multi-cluster, Hypervisor and Android Block Diagram

Applications

  • Networking and Communications
  • Android Devices
  • Video and Multimedia Processing
  • SmartNIC or DPU
  • AI SoC
  • Edge Servers

Technical Specifications

Short description
RVA23, Multi-cluster, Hypervisor and Android
Vendor
Vendor Name
×
Semiconductor IP