资源与支持

SiFive 博客

来自 RISC-V 专家的最新洞察与深度技术解析

November 29, 2016

The Best “Three-Month Project” Ever

A couple weeks ago, Jack and I traveled to Taiwan. Not that unusual for those of us in the semiconductor industry used to making a pilgrimage to visit customers and partners. This trip, however, was big for us, and we brought back the best “souvenirs” ever – the first ever commercially available RISC-V SoCs!

This morning, we announced the availability of the open-sourced Freedom Everywhere 310 (FE310) SoC and HiFive1, a companion development kit.

The FE310 is the first member of the Freedom Everywhere family of customizable SoCs. Designed for microcontroller, embedded, IoT, and wearable applications, the FE310 features SiFive’s E31 RISC-V Core, a high-performance, 32-bit RV32IMAC core. Running at 320+ MHz, the FE310 is among the fastest microcontrollers in the market.

Additional features include a 16KB L1 Instruction Cache, a 16KB Data SRAM Scratchpad, hardware multiply/divide, a debug module, one-time programmable non-volatile memory (OTP), flexible clock generation with on-chip oscillators and PLLs, and a wide variety of peripherals including UARTs, QSPI, PWMs, and timers. Multiple power domains and a low-power standby mode ensure a wide variety of applications can benefit from the FE310.

What’s even more exciting is that the FE310 is the first open-source RISC-V SoC available in the industry. SiFive has contributed the FE310 RTL code to the open source community. That means you can see what’s inside the chip and completely understand how the hardware works.

Take a look: SiFive GitHub

By releasing the RTL code, we want to encourage open source development of both software support for RISC-V as well as promote open hardware development.

Personally, I couldn’t be more thrilled or proud of the team’s hard work to hit this milestone. As part of the team who’s worked on RISC-V from its earliest days (back in 2010 when we embarked on a “three-month project” to identify alternate architectures for academic research), I must say it’s a dream come true to see a commercial RISC-V chip.

As the industry’s first commercial silicon based on the free and open RISC-V ISA, FE310 represents a milestone in the growth of the RISC-V ecosystem. As Krste puts it, “The FE310 is a major step forward in the movement toward open-source and mass customization, and SiFive is excited to bring the opportunity for innovation back into the hands of system architects.”

If launching our newest product and a development kit wasn’t enough, we’re also busy this week presenting four sessions at the upcoming RISC-V Workshop:

  • “SiFive FE310 and Low-Cost HiFive1 Development Board,” Jack Kang on Tuesday at 9:30 a.m.
  • “RISC-V Technical Committee Update,” Yunsup Lee (me!) on Tuesday at 2:30 p.m.
  • “Free Chips Project: A nonprofit foundation for hosting open-source RISC-V implementations, tools, code,” Yunsup Lee (also me!) on Tuesday at 2:45 p.m.
  • “’V” Vector Extension Proposal,” Krste Asanovic on Wednesday at 9:30 a.m.

If you’re attending the Workshop, please come say “hi!” and see a demo of the HiFive1 board. The event’s sold out so if you can’t attend, you’ll have to wait until the proceedings are posted after the event. Or come see us at Booth No. 1522 at ESC Silicon Valley next week.

Yunsup Lee
Yunsup Lee
Co-Founder and CTO at SiFive

Read more Insights from the RISC-V Experts

全力投入:开启增长新篇章
最新文章
全力投入:开启增长新篇章
我们自信地宣布公司发展历程中最重要的里程碑之一:完成 **4 亿美元** 的融资。本轮融资由 Atreides Management 领投,其他顶级投资机构\*包括 Apollo Global Management、NVIDIA(英伟达)、Point72 Turion 和 T. Rowe Price Investment Management, Inc.,以及现有投资者 Prosperity7 Ventures 和 Sutter Hill Ventures 参投。此次融资使公司估值达到 **36.5 亿美元**,并将加速 SiFive 的 RISC-V CPU 及 AI IP 解决方案推向数据中心和 AI 基础设施市场的核心地带。
RISC-V 代码模型(2026 版)
最新文章
RISC-V 代码模型(2026 版)
RISC-V 指令集架构 (ISA) 在设计上兼顾简洁与模块化。为了实现上述设计目标,RISC-V 有意识地减少了寻址方式的种类,从而降低了实现复杂 ISA 时的一项重大成本。寻址方式成本高昂:在小型设计中,会增加解码开销;在大型设计中,则会引入隐式依赖成本。
模块化是 AI 的未来:为何 SiFive-NVIDIA 的里程碑意义重大
最新文章
模块化是 AI 的未来:为何 SiFive-NVIDIA 的里程碑意义重大
AI 的巨大潜力目前正受限于一个主要瓶颈:数据传输。在当今系统中,GPU 的处理速度往往受到互联技术以及 CPU、加速器与系统其余部分间数据流动效率的限制。