 Good evening. So I'm here tonight to talk a little bit about the, well, the talk is titled, The Open Source Silicon Ecosystem, with the Open Source Silicon Design Ecosystem. And I'm going to start talking about that, but then I guess I'll get into the foundation and why we were formed and what it is we aim to do. I'll give you a bit of a background first, or rather, a bit of an overview of the talk first. So I'm sure people have been around doing this longer than me, but I think we really started with having open source designs and RTL downloadable from the internet in the early 90s. So you had open cores start up in about 1999 with the original team, and then it went from there. So I'll talk a little bit about what it was like then, what it's like today, and as you heard from Al, it's pretty encouraging. There's some amazing stuff out there. But then I'll talk about why I think it's not good enough and why I think we could do more and where I think it could go. So a brief introduction about me. I've been involved in the Open Source project since 2008, although these days I don't really do much coding and that's partly because work keep me very busy and also work tell me not to. So I just organized the beer. But I'm a digital design engineer and through my involvement with the community I've been involved in the startup of the Free and Open Source Silicon Foundation, or FOSSI, as we abbreviate it to. Talking about the bad old days, that's really all we had. Unless you used vendor tools and a version of Model Sim, there wasn't a lot available other than Icarus and Wilson Snyder's Veripool tools. So Wilson Snyder is responsible for, if you didn't know, Emax mode in, sorry, Verilog mode in Emax. And Verilator later on, I'm not quite sure what year that was released but that would have been nearly 10 years ago I guess. 99, is that when it started? Oh, okay, I see. All right, so a little while, but you know, he's been there forever. And then OpenCores as well, which I'm sure most of you are aware of. And you know, I'm only mentioning microprocessors here, like the Open Risk and the Leon Spark. I'm sure there were others. But of course, you also had to think about things like microblazing Nios, which were getting used regularly both in academia and in industry back then. And there were no open source tools. Okay, I've just mentioned a couple of simulators, but there was nothing serious for verification or synthesis. So if you look at where we are today and where things have gone in the last, I think it's 15 years, it is remarkable. You've got, first of all, the one making a lot of news at the moment is the RISC-5 architecture, which has come out of Berkeley. And that really, you know, being involved in the Open Risk project, we knew that the architecture that we were using was quite old, Spark is quite old. There wasn't really anything out there which was doing things that people wanted, like improving code density with a mix 16, 32 bit instruction set. And addressing architectural issues that allow you to improve, you know, allow you to feasibly achieve super scalar implementations and or, you know, address the super low and deeply embedded stuff. So we think RISC-5 does that and does it quite well. And I'm quite happy that it's, I'm very pleased to see that it's backed by, they've got their own foundation now, they've got money, they've got corporate sponsors who are getting involved to, I guess, help guide the architecture. From that project, or rather, I'm not quite sure whether that's caused these other things, but the low risk project, I'm sure many of you are aware of, and that's very, very cool, very encouraging to see that there was somebody and then some groups of people who are keen to put their money behind making a truly open source SOC that is gonna be sold to hobbyists. That Goblin Core 64, I only found that today, that's actually a sort of big data, high data throughput processing extension to the RISC-5 architecture, which is being done at Texas Tech. There's a few other things on there, which are modern, you know, contemporary projects, which have been, had a lot of work put into them and been released as well. And I'll talk a bit later about, I think, why we also need to get academia involved a little bit more in this, because a lot of projects happen, a lot of fantastic work goes on in academia and then the guys graduate, they disappear, no one touches it again, it sits on a server. We're lucky if it's, you know, we can Google and track it down, but most of the time, I think this stuff never sees the light of day, despite the fact that a lot of university's IP contracts would allow it to be released under an open source license. So as Al mentioned before, this Clifford again and his work is remarkable, I think in finally open sourcing a tool chain, you know, for many years we were complaining about the Xilinx and our terror tools of the world, again, taking up a terabyte each and then things crash, you're not quite sure what's going on. So that's really a breath of fresh air. Of course, you know, things like Verilator and Icarus have continued to be developed, supporting newer versions of Verilog. And then there's been a load of Python-based EDA that's come along as well. So, Sebastian, I don't quite know how to say his last name, it's BordaQ or something. His work with the Milky Mist project, which then led into sort of caused him to go and make some Python-based HDL tools and then SOC generation tools as well, which are quite impressive. Chris Higgs and Stu's Cocoa TB, which we've heard about before at Oshog, that's a very, very cool tool, enabling basically Verilog test bench access from Python. So that's very useful, quite powerful. And then we've got build flow package management IP with Fuse SOC and of course, as I'm sure there's many other tools that I've missed here, which have come about, but you do see things like this being used in the main these days with sort of hobbyist, small industry level digital design. So that's good. So too, the licensing. So we've had the solder pad license come along, which is a permissive, Apache-based, seemingly lawyer-approved and vetted license, which is what we need. CERN as well have paid a lot of lawyers to go and look at a reciprocal license, which is definitely useful. And then my own little one, the OHDL there, which was MPL-based and it's a bit of a fine and replaced job, but in spirit, it's a file-level-based copyright license. So 10 years ago, we didn't have these. 10 years ago, everyone was using the LGPL or the GPL, which to me seems like madness because they're explicitly software licenses being applied to hardware designs. I didn't think that worked. So it's very pleasing to see that there are hardware-centric licenses out and about these days. So I think in summary, we've definitely come a long way in the last 10, 15 years and it's encouraging to see what is out there and what is being used. I'm sure many of you have come across some of these fantastic turnkey FPGA kits you can get now. I mean, I'm not sure if one is available from the RISC-5 community, but effectively you can download a blob, buy a board, press a few buttons, put it on, boot Linux, and to me that seems quite powerful. And of course, the community is growing and that's great and that's very encouraging. But I still think there's something more to do in this area and that was the conclusion a few of us came to a couple of years ago. Somewhat spurred on by Jeremy telling us that something had to be done and being involved in probably the last gasp of open causes, old owners trying to make something of it. Yeah, it's still not a completely rosy picture I think but with the way the community has gone and or rather with the lack of serious sort of groups and organizations, as I said there's some fantastic hobbyists and small industry guys who are doing incredible work. But what I think remains is something to pull it all together and something to coalesce all of the great work that goes on in the various communities and academia projects that occur to bring it together to make something that's greater than some of the parts in a way. So that's what a few of us behind the Fosse Foundation thought or probably all of us actually. So we formed this I think late last year and it's a community interest company here in the UK but we've got directors all around Europe who have been involved primarily from the open risk project for the last 10, 15 years. So the idea is that we're gonna do a few things. First of all is supplant open cause with Libra cause. Open cause made a few mistakes I think with the way that they enforced a registration model. They had a few concepts which didn't quite work. However we think that with sufficient community input and our own experience we can do something better. And for instance I'll get into Libra cause a little bit more but the idea is you don't want somebody to host all your projects. You don't wanna be a source for just something. You wanna be more like an IP digest or a listing. So you go along look for what you're after. A CPU, a crypto core will have a link to where it is on GitHub or the source hosting that's popular with those developers right. But as well have it as a place which will host best practices guidelines. Be a focal point for the community to come and the place where they share their IP, the place where they know they'll get eyeballs, the place where they know they'll be found. And so the idea is that you'll get academia, hobbyists and hopefully small industry. I'm not quite sure how we're gonna handle the case of perhaps someone who's selling commercial cores coming and putting like a shareware version on there. Maybe that's okay. I don't know, we'll cross that bridge when we come to it. But the idea is that it will be there for people to basically list their IP to make it easy to find. Because at the moment you've got fragmentation. Everybody is in their own corner of GitHub or whatever their repository management system is and you can't find it. I mean you have to be very lucky to stumble upon some of these projects sometimes. And the IP is great and it should be out there and it should be reused. So we're all about encouraging reuse. As I mentioned as well, the I think licensing in this field is quite important and I think we need to encourage people to understand the virtues of the different licenses and of course provide a list of those. And we have a little bit of a, if you go to the Fossey Foundation website, in fact, I think there should be a link at the end. It would be remiss of me to not put a link to the foundation on here, but if you go there, you'll find a little bit of a summary of the current state of licensing in the digital design world at the moment. Of course, we're all about drinking beer as well. So the idea is that we'll get people together each year. We've been doing that in Europe so far. We'd like to perhaps expand that around the world and foster all of the communities that exist wherever they are. We've applied for Google Summer of Code funding this year and I think we've got a few places. We have three places. I couldn't tell you what the projects, what the work that's getting funded is off the top of my head, but it's good stuff. It's fantastic to see Google supporting us. We're also running our own student design contest, which is worth mentioning here. So I believe the idea is you have to submit by the end of August and it has to be something to do with open source digital design. It doesn't have to be IP. It could be tools, it could be flows. There's more information about that on the foundation page as well. Actually, maybe that's on the Libra course page, but you'll find it. I'll get to it a bit later, but I do think as well getting industry involved in this area is important, be it small or big industry. I do think that's how the open source software movement got a lot of momentum and got their critical mass to play such a big role that they do now in industry and obvious lives. We also get a bit of a manifesto. The idea is that we want to support and champion open standards and design reuse. That's the idea. And of course, expose or bring more visibility to academia projects. So what's on Libra course at the moment is some documentation, some getting started best practices stuff. We've also got a bit of a blog planet on there, which I'm sure we could add more things to to make it more relevant. And then of course, the project on IP repository that I mentioned is still a bit of a work in progress. So at the end of the day, I think what we want to achieve is something that resembles a modern day open source software repository, something that has a lot of stuff that's relevant or is finished and mature and is there and you know it's there and you can go and get it and you can reference it. And as I mentioned before, I think industry is important to get involved here. I think it's a bit of a build it and they will come sort of idea, maybe not entirely, but I think that if you show a model that is or if you demonstrate a model that is usable by industry where they can come along and they can contribute bits of IP and they're worried about reciprocal licensing or being sued or something because somebody knows their micro architecture. If you can somehow get them involved, I think you're going to see a big swell in the sort of amount of participation we're going to get in this field. And obviously everybody benefits from that. And we'd like to see progression in the way that people develop and release and verify their cause. So that's all part of addressing the quality of the stuff that gets put out there. At the moment it's a little bit, it's very variable. People are learning, people are putting out cause which are very well verified but it's not obvious that you can go and run that verification yourself. So I think that's a big issue. And you know things like continuous integration, there's a project I think being run at the moment as well where we're trying to get up a continuous integration service, something like we have a website, you go and point it at your IP and you package your IP with a certain bit of metadata and the continuous integration tool can go and unpack it and run it and then obtain metrics about how many tests are failing or passing or what it is you're testing and coverage and things like that. So there's a couple of interesting ideas that we have at the moment that we'd like to implement. And so starting though from the very beginning we need to get this thing started and there are only a few of us at the moment there's probably only six I think, six or seven in the Foundation and I can't remember when we launched. It would have been mainly, well a main launch occurred last year at our conference which was in October and since then we've been very busy and inundated with people who are very interested to help us and like the idea and see that we are trying to do open cause right but we need more people to help us. If anyone's interested please give us an email or come and talk to us in IRC and we're running the conference again in Bologna and Italy this year in October but yeah this is definitely something that I think if it's done right we'll have a positive impact on the community and the amount and quality of open source IP that's available for people to use and if you're interested or motivated to come and give us a hand with Libre Cause or anything to do with the Foundation please give us an email or talk to me afterwards. So that's it, I'm happy to take questions.