RISC-V architecture development

maddog's Doghouse

Article from Issue 219/2019
Author(s):

As open source software development for the RISC-V architecture moves ahead, maddog says stop complaining and start contributing to the project.

Recently a colleague sent me a link to an online article about the RISC-V Foundation and the Linux Foundation agreeing to work together to advance Linux and other open source software on top of the RISC-V architecture.

RISC-V, for those of you who do not know about it, is a relatively new architecture that originated at the University of California, Berkeley (the same people who developed the Berkeley Software Distribution, also known as BSD).

Having years of history and experience behind computer science, these architects "started from scratch" and were able to develop a relatively clean architecture that could support multiple address space sizes, as well as multiple usage cases (servers, portable devices, embedded systems, Internet of Things, etc.). This design allows trade-offs in the chips for speed and power without moving away from the basic architecture.

One of the other features of the open architecture is that it can be used without the oppressive licensing fees of some other architecture vendors. It allows for hardware vendors (either commercial or non-commercial) to create chips that follow the basic architecture but also add extensions as desired.

The article was published on December 1, 2018. Since I received the link on December 1, 2019, enough time had passed for the normal peanut gallery of commenters to make their thoughts known.

Many of these comments seem to have come from people who have never produced a piece of sophisticated hardware in their life (and perhaps never even produced a piece of sophisticated software). While some of the people had intelligent and insightful knowledge to provide, other commenters were often ignorantly rude.

As an example, when it was pointed out that the project was still in its early stages, people wanted to know when they could have "their chip" or "their board" for whatever proposed purpose. As a response, other commenters would start to detail the issues around taking a proposed architecture and turning it into real silicon, not the least of which is fitting it to a manufacturing process, scheduling the production, and setting up the supply chain for getting the chips onto a board and into the hands of the "customers." While these steps are in progress, it takes time (and money) to formulate. This process is rolling out, but apparently not fast enough for the anxious (and rude) commenters.

Then came a discussion about how much the boards might cost. Many of the commenters were shocked by the prices of "developer boards," boards that would go into the hands of people working on the compilers, operating system, and libraries to help make the entire system available. In the free software space, a lot of these people, even today, are volunteers, and not able to afford a chip and board that are made in relatively low production numbers.

Therefore, companies like SiFive [1], which is one of the first companies to actually produce silicon and boards supporting the architecture, are not charging "Raspberry Pi" prices for their high-end development boards. Instead, they have innovative programs that allow developers to get the boards needed to do real work, as well as generate the funds necessary to create new boards. In addition, SiFive has boards that allow people to experiment with the architecture at a lower price than SiFive has to charge for their higher end boards.

Many of the commenters were complaining about the business plans of companies that used closed-source tools to make creating the silicon easier and faster. While these companies are using closed-source tools, other companies and groups were creating open source techniques, which may take more person months but that were still available for those who wished to use completely open source techniques.

Finally, there were the naysayers, who (when they finally started to comprehend the roadmap to bringing out this architecture) questioned why it was necessary and lamented how long it would take.

For the naysayers, I have these thoughts: Linus started the Linux Kernel project in late 1991, and it was more than two years (early 1994) until it reached version 1.0 level. Distributions appeared in late 1993 and gained speed in mid-1994. It was not until 1998 that commercial databases started to support their products on top of GNU/Linux (a sign of acceptance in the commercial Unix space) and 2000 when Linux was used for embedded systems. It still took a decade for GNU/Linux to be accepted by a lot of businesses.

The year 2019 will mark the 25th anniversary of the release of version 1.0 of the Linux kernel and the creation of the Beowulf supercomputers (using Linux at their core). The latter project now drives all 500 of the fastest computers in the world.

Some things take time and money. In the scope of things, RISC-V is moving at the speed of light. Instead of complaining, join the project.

Infos

  1. SiFive: https://www.sifive.com

The Author

Jon "maddog" Hall is an author, educator, computer scientist, and free software pioneer who has been a passionate advocate for Linux since 1994 when he first met Linus Torvalds and facilitated the port of Linux to a 64-bit system. He serves as president of Linux International®.

Buy this article as PDF

Express-Checkout as PDF
Price $2.95
(incl. VAT)

Buy Linux Magazine

SINGLE ISSUES
 
SUBSCRIPTIONS
 
TABLET & SMARTPHONE APPS
Get it on Google Play

US / Canada

Get it on Google Play

UK / Australia

Related content

comments powered by Disqus

Direct Download

Read full article as PDF:

Price $2.95

News