New Developments in Windows on Arm

Bill Sikkens: Hi, and welcome to Arm viewpoints. I’m Bill Sikkens filling in for Geof Wheelright. We’re going to be talking a little bit about Windows on Arm, the new tools and hardware available to enable more apps to switch to Arm. We have two exciting guests today, so let’s get started.

David Whaley is the director of strategic partnerships with Arm. David is currently working to ensure performance, support and optimization for the most important software and tools within the Windows on Arm and the Android ecosystems. Prior to joining Arm, David has worked on numerous mobile and desktop software products and managed key strategic partnerships at several prominent technology companies.

He is currently helping drive Arm’s developer recruitment efforts and streamline the developer ecosystem for Windows on Arm. David understands what is required to engage developers at early stage and build self-sustaining software ecosystems Dave’s insights on the innerworkings, into these ecosystems and experience managing customer software products and building the app developer partnerships has been instrumental in expanding the Windows on Arm ecosystem.

Rich Turner is a Microsoft veteran of previous roles in HBO and startups. He first joined Microsoft UK in 2000 after founding his own software consultancy and development company. He now heads up a team that modernizes the Windows native app platform and the related technologies and is responsible for some of the exciting innovations being announced at Microsoft’s Build 2200 event. Welcome David and Rich. Several years ago, Microsoft announced that it would run Windows 10 on Arm hardware. David, are you able to talk us through this and how this collaboration came to fruition?

David Whaley: Thanks, bill. Great question. Yeah, I’ve been at Arm about the time when just before these devices came to market I’m primarily a software guy, so I’ve been working.

With the different parts of the Windows team within Microsoft to make sure that Arm was working well, that it was optimized that really just the ecosystem readiness was there. I think there’s another part, there’s the sort of portfolio and a whole nother story around how the hardware came about.

I don’t know if rich, you could, if you have a better viewpoint of that, but I’m, I’ve been working on the software ecosystem since the beginning. And that’s been a very interesting journey.

So Rich, why don’t we ask you that? Do you want to speak a little bit broadly about the hardware side of this?

Rich Turner: It’s actually a fascinating story. If you think back, if we go back a little bit into history, for some who may not have followed how Windows is evolved over the years Windows NT upon which all modern versions of Windows have been built. Was always built from the outset to run on multiple architectures.

It was actually originally built on the Intel i860 RISC processor. Then the MIPS R3000, then the ex the 80386. Then we brought in a power PC and DEC alpha and Itanium. And then more recently x64 and then in 2010, ‘11, ‘12 timeframe. We brought out Surface RT, which was our first foray into surface devices, which ran a version of Windows on an Arm processor.

And then later in just three or four years ago, we then introduced Surface pro X, which ran on much more modern Arm, processors with more capabilities and more Ram and so on. This is the natural evolution really of Windows starting to adopt an exciting new processor or exciting new systems architecture that started to become quite powerful and quite flexible.

And we really just wanted to make sure that Windows and Windows, developers, and Windows, customers have a great experience when running Windows on any hardware platform they choose to use. And so we brought support for running Windows on, Arm devices. To the market just not that many years ago, what is it? Three or four years ago now, but yeah, it’s been super exciting as someone who first ran an Arm processor on an Acorn Archimedes back in 1987 is really excited to see becoming really interesting as a viable desktop platform now..

Bill Sikkens: So talking about this and David, I’m going to ask you this on the software side. What is it that separates Windows on Arm from other alternatives?

David Whaley: I think the what Arm brings to the game here really is increased battery life instant on. It can be always connected. It’s a lower power consumption, thinner designs, fanless designs.

So it’s it allows a form factor that people really want as well as the performance that. Expects that’s really what Arm is able to bring to Windows. And we’ve when we’re seeing that with the devices that are being released in the market. We think it’s very exciting.

It’s absolutely coming to fruition. It feels there’s people that are making investments, in the ecosystem now. And it just feels, it feels good.

Bill Sikkens: Yeah, that sounds absolutely amazing. And rich, I’m going to throw this one out at you .This

moore, kind of a hardware end of things. And that’s really who were the intended buyers for Windows on our laptops? Would it be consumers, business people, developers, and what are the models and markets? Will they expand over time?

Rich Turner: Basically yes. All of the above.

There you are. Hey that was easy.

As David said, the Arm has some very interesting characteristics in particular, around power management, long battery life, fanless cooling, or very low fan cooling, et cetera. Plus a lot of flexibility in what can be integrated onto an individual SOC. So we see Arm as a really interesting platform, especially for ultra mobile and mobile scenarios.

People have to work on large and large industrial engineering sites or have to travel around a lot with a device, having to carry a power brick with you and keep plugging. Is something that is a bit limiting in many scenarios. So we see Arm as a really interesting platform for in particular the very mobile worker which could be anyone from a doctor or a vet or an engineer working in a petrochemical refinery, that needs to be out and about all day long, all the way down to office workers who might just need to pick up a device once a week and take it to a given site or to a meeting and then come back and maybe leave it alone for awhile. There’s all kinds of scenarios in which, and a more mobile working pattern can free up people to work more flexibly. And we’re really excited about that. .

Bill Sikkens: And it does seem this would lead to smaller and thinner devices and all that kind of stuff, less power consumption. I think that’s absolutely amazing.

Rich Turner: Yeah, absolutely. I You look at when you go to a restaurant these days, oftentimes the waiters are carrying around a small tablet of some kind to take your order or to process payment, et cetera, et cetera.

So there’s all kinds of opportunities. I think we haven’t even begun to tap. Where smaller, more flexible devices or devices built from more specialists use will become really interesting and prevalent. So we’re keen for Windows to be able to have a place there as well and store.

Bill Sikkens: Now one other huge issue with any ecosystem of hardware and software is what you can run on it and what people are doing for it. And David, I think I’m going to throw this one at you because this is more software end of it. How is the Windows on our ecosystem developing?

David Whaley: Great question. I think that this is very this is the hard part. We’re overcoming years and years of legacy around x86, x64, you reap the most benefit of the Arm architecture by having an app that runs natively.. Just w without question Windows on Arm came out first with 32 bit app emulation worked great. But there was a power hit. There was a performance hit now with Windows 11 there’s 64 bit emulation. that’s very good. There’s also something called Arm 64, EC, which allows you to. To basically wall off certain libraries, certain portions of your app that that continued to run emulated, but the rest of it runs natively.

We’ve been doing some testing with that. Great performance great power efficiency, that sort of thing. But we are also looking very closely and working very hard to get the tools, the frameworks, the compilers, all the tools that everybody’s used to using for developing native apps,. Getting them ported over to Arm.

We’ve done yeah, we’ve done a ton of work with Microsoft. We’ve done some work Arm does its best work really in the open source community. That’s what we’re, that’s what we’re accustomed to. And so we’ve done work in Microsoft, thankfully has joined Linaro around the first of the year, which is which is the sort of the Arm source community.

Or one portion of that community. And there’s a Windows group there. That’s now working on a, app frameworks, compilers, LLVM is now available. Cute. We’ve done the electron and CF app frameworks and all those are being maintained and built by Linaro.

So we’re seeing a lot of. Which we think is great. We’re seeing more and more things happening that that, hopefully we see more Microsoft tools as well. And that’s what we’re, that’s what we’re hoping for. And we’re working to make happen..

Bill Sikkens: What do you think of as some of the mainstream stuff, say Netflix or Adobe illustrator or something like that? Are we seeing that coming to the ecosystem

David Whaley: eventually? Yeah. What we’re really shooting for is a self-sustained sort of organic ecosystem. Apps just appear. It’s not where, or it has to be a high touch by either Arm or Microsoft or anybody else.

So we are seeing apps that are coming. Like you mentioned Adobe illustrator, there’s a version of Adobe Photoshop. That are, that it’s native. Obviously all of the Microsoft office apps are native or at least run Arm 64 EC in support of their large plugin.

Zoom has come out as native. Amazon music, Kindle reader. There’s quite a few apps that, that that, that pop out in what we do the strategy that we’ve had is we try to engage those app developers with the apps that people expect to be running well on the platform.

And if they’re blocked with something we engage them, find out how they’re blocked and see if we, if there’s some way that we can prioritize, fixing what they’re blocked by. And maybe that’s going out to Linaro and saying, Hey, these guys are blocked. Python or something else that they use some math library that we could potentially port over or optimize for Windows on Arm.

If there’s that sort of situation that’s how we try to tackle it.

Bill Sikkens: Okay. And having these tools available, it makes a lot of sense, which I’m going to throw it off to you. Please go ahead.

Rich Turner: Yeah. I love what David was saying. This software really is the key to this whole thing.

If we don’t have native applications that run beautifully on Arm hardware, taking full advantage of the performance and the battery life. And so on the Arm devices offer, then we kinda missed a. Huge opportunity. And the emulation technology that we currently have is great in terms of being able to get things running initially, but ideally we’d want more and more applications to be ported to, Arm, native and to be able to take full advantage of the hardware underneath.

And that’s why we’ve been plowing enormous effort over the last year, 18 months or so into bringing. So many of our key developer tools, runtimes frameworks, and so on to themselves run natively on Arm and enable developers to build Arm applications, using tools that they’re familiar with. They’re comfortable with, visual studio, for example dotnet framework, our VC compiler, runtime, and tool chain bringing all of these tools to the developers so that developers. Are no longer blocked by not being able to have a tight and effective development loop, nor the developer tools that they would rely on to build the applications that end users will later run. So it’s super, super excited to finally bring real Arm native versions of visual studio, .net framework, visual studio, and working with David and Linaro in many different ways.

Both close source, as well as open source vendors, have libraries and tools to start bringing that ecosystem to run on Windows on am as well.

Bill Sikkens: No, that sounds amazing. Now, as an app developer myself, I’m going to throw this next question out at you and either one of you take it. Whoever wants to just jump in on it. This is all great. If I’m sitting down and writing new software, but what about my legacy code? You had talked about emulation, all that kind of stuff, but that almost always brings in a performance hit. Is there a path to being able to deal with that more directly?

Rich Turner: , so that’s a really good question. If I’m, if I, if you don’t mind. Cause you did say anyone take the question. But fundamentally from Microsoft’s perspective, we largely don’t want developers to even notice they’re running on Arm nor building for Arm. We want developers to be as productive as they are today. Building a traditional Windows, PC application, as they would be building the same application, but targeting Arm. We want them to have the same tools. We want them to see the same error messages if they have a bug in their code, as they might do, if they were to build on x64.

But we want the developers to largely to have a no op operate opportunity. How this is familiar. I’m familiar with this. I can just drop down this, the architecture that I’m targeting my application for, I hit rebuild and everything just works. It shouldn’t be a huge amount of effort to bring an awful lot of applications to run natively on Arm.

That’s the key goal with this. If we have to ask developers to go out of their way to do additional work that they have. Already prepared to do then that’s added an extra cost to them. That’s added an impedance that prevents them from porting their apps. So wherever we possibly can, we’d like the operation to be a no op. And for them to simply say, huh, it’s largely a recompile, maybe a couple of tweaks here and there, but it shouldn’t be a huge amount of effort.

Bill Sikkens: I know for one I’d find that. Absolutely amazing. David, go ahead.

No, I think the one thing I can add is, about, I’m going to say probably two years ago we were.

Most concerned about getting native apps on onto the platform. So developing on another architecture targeting Arm was perfectly okay. And that was two years ago. And since then, we’re really concentrating on Arm development for Arm on, on Arm. And that makes it even easier for the app developers to, compile, to test and really the next thing that we’re really focusing on is how does.

How does a developer test things how do they get these things into a CI/CD situation so that so that they act just like any other app that they’ve been building for the last few years and that’s the part that, that we’re focused on as well as the tools.

But also just the methodology and how. How developers actually do the work and we’ve, we’ve gone out and, there’s certain websites that we go out and talk to people, survey developers to see what they’re doing. And that helps us understand where we need to protect.

Rich Turner: Yeah that latter point you made actually about the CI/CD is super important as well. We’ve spoken with a lot of developers over the last few years, and I don’t know many developers these days that aren’t touched at least in some way by some form of CI/CD be it in house, hosted CI/CD and their own lab, build labs and infrastructure. Or using things like Azure dev ops and GitHub hub and so on to build that their applications in the sky. And so we really are very keen to help support those workflows as well, which is why Azure team at Microsoft have been built. VMs in the sky that Army in a VM infrastructure, which can then be used by Azure dev ops and get hub and others to then build out their CI/CD services as well.

So that you’ll be able to push your code to your repo that will automatically trigger a build process that builds your application. In an Arm VM using Arm native compilers and linkers and runtime, runtimes, and frameworks, and so on. And then crucially execute your built, binaries, and run all of your unit tests.

That’s a big gap that. That blocks a lot of people right now, where you can build cross compiling. You can build the application in the x64 VM, but you can’t run the tests. And so having the ability to actually run your full suite of unit tests as well and integration tasks and so on as part of your CI/CD is absolutely crucial.

So we’re super excited that Azure has now got Arm 64 VMs, and that we’ll be working over the next year. To figure out how to deliver all of the get hub and Azure dev ops services that developers will need to build and test their applications up in the sky as well. No,

Bill Sikkens: That sounds amazing.

Having a VM working like that has got to save a lot of time and money just, on the base of it going forward. And it seems like it could eliminate a lot of other problems. So going forward. What are we looking at for the future? What does the future hold for Windows on Arm? David, go ahead.

David Whaley: From our, from our standpoint, we think that this is I think we’ve proven out that the Arm architecture is more than capable of powering these class devices.

I think we’ve seen it there’s also, these types of processors running on, on infrastructure in the cloud it’s absolutely not the case that the. That every Arm, arc architecture, every Arm processor is running on a smartphone. It’s much more capable than that.

And we see, more and more and more devices coming out. And we expect the Windows on Arm platform to, to be successful.

Rich Turner: Yeah, the from the window side, sorry, from the Microsoft side of things, I should say this isn’t just about Windows. This is about all of our applications, developer tools and so on being available on, Arm, and by the end of 2022, our goal is to have the majority of our.

Mainstream key developer tool set available natively forearm in GA in January available. So it’s fully supported. So enterprises combine and so on. So by the end of this year, we’d like to, we’d like to have the vast majority of developers unblocked and being able to start porting their application or building applications, for Arm, if they’re starting from scratch as well.

And then moving forward. Into next year. We’d like to make sure that for example gaming is an exciting opportunity for developers to explore as well. And we want to make sure they have all the dev tools. They need to build the tools that game developers themselves need to use as well as the games themselves. Enterprise applications. For example, it’s a huge opportunity has again with its mobility and. They have relatively frugal power consumption. So on we think that portable devices are super interesting and important for hospitals and schools and large organizations. So you think about all those millions and millions of line of business applications, things that enterprise developers have been building for years. Wouldn’t it be great if they could literally just take. Quick recompile. And now it’s available for Arm, we think that’s a really powerful story as well. And so we’re working really hard to make those futures a reality as well.

Bill Sikkens: So it just seems like going forward the sky’s the limit in many ways on what you can do with productivity. And you’re talking about gaming on down the road and all that kind of thing. So I think it’s going to be exciting to see where this ends up. David Rich, thank you for joining us this week on viewpoints. Thanks to you both for sharing your insider perspective. And thank you for listening.