Hacker News new | past | comments | ask | show | jobs | submit login
Open-V: The World's First Open Source RISC-V-based 32-bit Microcontroller (crowdsupply.com)
38 points by msl on Nov 22, 2016 | hide | past | web | favorite | 19 comments

This is great! The source code is written in Verilog HDL and is located here[0]. I hope they use english instead of spanish as a language to attract more contributors.

[0] - https://github.com/onchipuis/mriscvcore

Some pretty concerning red flags: They said "RISC-V", but that's not saying much. Is it RV32I, RV32IM, RV32G, does it have supervisor support, etc. Yes, I know I can look it up on github, but this kind of information should be at the very top of the front page.

RV32IM with interruptions capability. Sure, we will add this at the CS site.

Great initiative but worrysome execution. Can the team share why we can't buy chips in lots of 100, 1k or even 1M chips? $500k isn't enough for TSMC, so who'll pay them? The ESP32 seems to have a much, much stronger story.

I think at some point the whole purpose of the project was completely lost:

1. What is offered in the ESP32 is completely different to what Open-V can bring to the table. ESP32 is not a microcontroller, is a platform for Wi-Fi and bluetooth applications, aim to be used in IoT apps. Does it have a microcontroller? Yes. Can you see what is inside? No. Do you have access to schematics and RTL code? No. Is it cheaper? Yes; and of course it is! Right now it is impossible to compete with big vendors at price levels. They have the triumph card when mass production is taken into account.

2. But wait, maybe you are thinking: what can I do with all this open information?. Do I really care to know how things work? Think about open-source software at its early beginnings; now think how open-source software has ushered in the last two decades of software innovation; the people that believe in this kind of projects, believes that open source silicon will unleash, as its software counterpart, a flood of innovation, hardware-wise. Imagine what could we do in a future, where small IC design startups could have access to any kind of RTL code or open silicon analog circuitry to build their own projects, to being able to use the Open-V code as a core, and add accelerators, peripherals, or whatever is needed. ENDLESS possibilities, but we have to start somewhere, and that's Open-V.

I know I may have been too aggressive, but I'm just too passionate about this :)

In the campain site they commented about getting an MPW and 70k chips. It seems like they are prototyping and getting some few number of wafers to get TSMC into it. The coolest thing is that they mentioned in http://hackaday.com/2016/11/22/mrisc-v-the-first-open-source... comments that they will release chip schematics.

What your saying about the quantities makes sense but your comparison to ESP32 is strange as that isn't an open core at all.

It is all about economies of scale, price will come down eventually. They have a reward with 15 chips per 199 + dev board. That goes $100/15 per chip.

I think they are produced on wafers with other things so, presumably a third party is collecting together a load of stuff and having it made together.

Analog IPs are one of the most expensive part of the design. I hope these IPs will be free. Is there any voltage regulator?

I read in hackaday that Analogs IP will be open, But I couldn't find anything about voltage regulators.

Most of dev boards for hobby-end microcontrollers in the market are using high level language to easy the micro programming, is any high level language such as C supported?

C is supported. We are using the riscv.org toolchain considering that the micro has a RV32IM set. An arduino bootloader will make the micro Arduino-compatible .

Getting 160MHz makes things more expensive, technology wise. Why not starting in a cheaper technology to be open source?

Good point. We are trying to cover some mid-end apps too. Basically the reason is we are planning to implement some functionality on-the-chip, such as a USB 3.1 PHY which requires a decent tech.

Hopefully they have something like the modern micros which includes USB interfaces onchip...

The current measured chip v.1 of the Open-V has not USB PHY. We are using a USB-to-SPI cable to program the micro. However, we are developing a low-speed USV PHY: PCS and PMA; and also a high speed PHY USB3.1 https://twitter.com/onchipUIS/status/787124721955508224. We are hoping to tapeout soon the USB PHYs to see feasibility inclusion for Open-V v.2.

when you talk about mass production, how many tape outs are expecting before sending the chips to your sponsors? how can we be sure the chips have a proper ESD strategy or are not weak to typical electrical over stresses?

We have now a first prototype, fully tested, so, with this campaign we expect to improve our first version but both chips will be available. We use ESD protections structures provided by the foundry.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact