Avispado

Avispado
In-Order
Efficient RISC-V Core

A highly efficient and power-conscious core, ideal for AI edge and embedded applications.

In-Order,
Efficient RISC-V Core
A highly efficient and power-conscious core, ideal for AI edge and embedded applications.

Key Features

A highly efficient and power-conscious core,
ideal for AI edge and embedded applications

64-bit In-Order Execution

Optimized for energy efficiency.

Vector-Ready

Supports RISC-V Vector Specification 1.0 for AI acceleration.

Ideal for

Machine Learning, Edge AI, IoT, and embedded computing.

Multiprocessing Support

Scalable to multi-core implementations, Linux Ready.

High Bandwidth Access

Gazzillion Technology™ ensures efficient data movement with up to 64 simultaneous memory requests.

特征

64位内核
(RISCV64GCV)
2宽度 顺序执行
支持多处理器
(AXI/CHI)
对非对齐访问的直接硬件支持
支持MMU Linux

可用扩展

比特操作
CMO’s
数据长度支持Half/bf16/Single/Double
Zifencei
数据加密
开放向量接口(OVI)
更多信息
向量单元
更多信息
Tensor Unit
更多信息

可定制选项

分支预测器
向量规范1.0(支持向量)
I$ 从 8KB 到 32KB
D$ 从 8KB 到 32KB
Gazzillion Misses™
更多信息

支持多处理器

Avispado支持cache相干的多处理器环境。其内在的CHI接口可被裁剪成AXI,取决于您的需求。

无论是2,4甚至上百个内核,Avispado都可为您的下一代SOC服务。

Avispado测试芯片

Avispado Test Chip