Microsoft is claiming this week that its hardware partners are “on track” to deliver the first ARM-based Always Connected PCs by the end of the year. Like many of you, I’m looking forward to this. And I’m curious to see how—or even if—Qualcomm can shake up the PC industry.
So what’s happening here? There seems to be a lot of misinformation—or at least misunderstand—around these two separate but related pushes.
Sign up for our new free newsletter to get three time-saving tips each Friday — and get free copies of Paul Thurrott's Windows 11 and Windows 10 Field Guides (normally $9.99) as a special welcome gift!
"*" indicates required fields
As you may recall, Microsoft announced last December that it was partnering with mobile chip-making giant Qualcomm to port Windows 10 to the ARM platform that powers most smartphones and tablets. This isn’t another halfhearted port like Windows RT or Windows Mobile, however. It’s “full” Windows 10, including compatibility with (Intel-based) the Windows desktop applications that customers actually need.
The Qualcomm partnership is two-fold.
First, Qualcomm provides the leading ARM designs today, and Windows 10 is specifically being ported to the high-end Snapdragon 835 chipset, which has been shipping for months on various flagship Android handsets. That chipset is powerful enough to run the x86 emulation software that makes Windows desktop application compatibility possible.
Second, Qualcomm is the only ARM chipmaker that is big enough and powerful enough to offer a counter to Intel. This is important to Microsoft because AMD, despite a nice Ryzen push this year, has effectively disappeared from the market and no longer offers effective competition to Intel, a slow-moving behemoth that won’t innovate unless pushed. This is key to Microsoft, as it needs Windows 10 running on modern, reliable, highly-connected, and very portable hardware; that is not Intel’s strong suit.
We’ve only seen a handful of Windows 10 on ARM demos since that initial announcement. And we’ll need to collectively test a variety of hardware devices—ASUS, HP, and Lenovo are among the PC makers on board here—to see whether Microsoft’s promises are true. That is, will ARM-based Windows 10 PCs deliver the performance—especially with x86 desktop apps—that users expect from PCs and the battery life they expect from modern mobile devices? That’s an open question.
So let’s get to the misunderstandings. There are a few.
First, Microsoft is not making one or more new Windows product versions, or SKUs, that specifically target ARM. Instead, it has specifically said that it will port Windows 10 Home, Pro, and Enterprise to the platform. (This was last December, and the firm has since announced Windows 10 S, which will also be ported to ARM.) In other words, Windows 10 is Windows 10. And when you run, say, Windows 10 Pro on ARM, you get exactly the same capabilities as you when you run Windows 10 Pro on an x86 PC. Because it’s the same product. Boring, right?
Second, Microsoft in May announced something called the Always Connected PC initiative, which establishes a new class of portable PCs that include Embedded SIM (eSIM) capabilities for universal compatibility with cellular data systems worldwide. This initiative has been incorrectly tied to ARM by many because, after all, ARM chipsets—like the Snapdragon 835—do include integrated cellular data chips. It’s a natural fit.
Well, don’t be fooled by that: You will, in fact, be able to buy Always Connected PCs that run on both Intel and Qualcomm platforms. And there is a longer list of PC makers—ASUS, Dell, HP, Huawei, Lenovo, VAIO, and Xiaomi—that are known to be building Intel-based Always Connected PCs. You’ll have more choice on Intel than on ARM. Especially at first.
But there is a related issue to consider. It is Qualcomm, and not Intel, that has the expertise in this area. It’s mobile chipsets have offered integrated cellular data connectivity for years, after all. And more to the point, many PC makers—including Dell, HP, Fujitsu, Lenovo, Toshiba, Panasonic, and, wait for it, Microsoft—already offer Qualcomm cellular modems in their PCs. Because they’re the best in the business.
But the Snapdragon 835 still has the edge from a connectivity standpoint. Not only is this platform more efficient than anything Intel offers, but its cellular modem technology is better, too: The Snapdragon 835 features the Snapdragon X16 LTE modem, which supports Gigabit LTE for download speeds of up to 1 Gbps (1000 Mbps).
“That’s three to seven times faster than the average U.S. broadband speed,” Qualcomm says. “The Snapdragon X16 modem is designed to support an ‘Always Connected’ Internet experience that’s faster than most Wi-Fi networks in homes and offices.”
Can that type of connectivity speeds overcome any of the presumed slowness of emulated x86 code? Will these ARM-based Always Connected PCs deliver battery life that exceeds, hopefully dramatically, that of their Intel-based competition?
I cannot wait to find out.
Stooks
<blockquote><a href="#173214"><em>In reply to glenn8878:</em></a></blockquote><p><br></p><p>"AMD's Ryzen has forced Intel to respond."</p><p><br></p><p>Big time. Paul and the technical details don't always jive. Ryzen is very popular especially among those the that stream or do a lot of video work. Dell, HP and Lenovo just last week announced Ryzen workstations.</p><p><br></p><p>If I was building a new multipurpose desktop today it would be using a Ryzen 1800x.</p>
Stooks
<blockquote><a href="#174028"><em>In reply to ChristopherCollins:</em></a></blockquote><p>I was more focused on Paul's comment about how AMD has not done anything even with Ryzen. </p><p><br></p><p>You are correct both price cuts of the current mainstream intel products and the rushed i9/"8th" gen stuff is all a reaction to Ryzen.</p><p><br></p><p>Yes ARM is better for power consumption but it depends on the work load. For light to medium demanding stuff it should be fine, but if you need powerful CPU's ARM falls by the wayside quickly. I personally do not care about the power savings, I want the power.</p>
skane2600
<blockquote><a href="#174691"><em>In reply to MikeGalos:</em></a></blockquote><p>While you could say that a lack of 64 bit support is not that significant, I can't see how such a limitation could be "ideal". Obviously 100% compatibility would be better.</p>
skane2600
<blockquote><a href="#174719"><em>In reply to jimchamplin:</em></a></blockquote><p>Immaterial? What does Rosetta have to do with this discussion?</p>
skane2600
<blockquote><a href="#174731"><em>In reply to jimchamplin:</em></a></blockquote><p>Well, it's similar in the sense that it allowed you to run some programs designed for one platform on different platform. That doesn't mean that the limitations of Rosetta have to be adopted by Windows on ARM.</p>
Stooks
<p>"<span style="color: rgb(0, 0, 0);">you get exactly the same capabilities" "I cannot wait to find out"</span></p><p><br></p><p>Neither<span style="color: rgb(0, 0, 0);"> can I. However as they say "the devil is in the details".</span></p><p><br></p><p><span style="color: rgb(0, 0, 0);">Lots and lots of questions will be answered. Like you say it has the same capabilities. Does that mean all of Windows, including all of the apps that come pre-installed have been ported? Or does that mean just the core OS and those pre-installed apps are running through this new emulation….and are less "optimal" in terms of performance?</span></p><p><br></p><p><span style="color: rgb(0, 0, 0);">I think the use case for these devices will be mostly used in vertical markets. A smaller, lighter device with better battery life that runs Win32 apps with "OK" performance but for specific use cases that is acceptable. </span></p>
Stooks
<blockquote><a href="#173240"><em>In reply to jimchamplin:</em></a></blockquote><p>It is not as easy as you suggest. ARM does not have full API support with x86/64. If it was that easy Windows RT (Windows on ARM) would have had actual apps and been a hit. RT died because no one ported their apps.</p><p><br></p><p>The whole deal with this OS is that you DONT have to port your apps and you get the advantage of small devices, better battery life and cell support.</p><p><br></p><p><br></p>
Stooks
<blockquote><a href="#174520"><em>In reply to wright_is:</em></a></blockquote><p>I fully understand it. My initial question was…is it just the core OS that has been re-written for ARM or the core OS and all the included software (base install), like Edge, Mail, People, etc been re-written/ported to native ARM?</p><p><br></p><p>My preference if using one of these devices would be for all apps to be written in native ARM code for performance reasons. That will never happen or RT would have been a real thing. I am at least hoping that Microsoft has ported all of its apps to native ARM…..but since not all of their software is even UMP on x86 I am not holding my breath.</p>
skane2600
<blockquote><a href="#173240"><em>In reply to jimchamplin:</em></a></blockquote><p>It's never as simple as just a compile-time option when moving from one architecture to another. Except perhaps for C programs that use nothing more complicated than the standard C libraries.</p>
skane2600
<blockquote><a href="#174717"><em>In reply to jimchamplin:</em></a></blockquote><p>Theory and practice are often quite different. If you have personally ported a moderate-sized application from Intel to MIPS or Alpha and all you did was change a compiler option, I'll defer to your superior experience. </p>
skane2600
<blockquote><a href="#175071"><em>In reply to stephenf:</em></a></blockquote><p>It's not that anyone forgot, it's just that these portability efforts are never 100% effective. For example, read the end of this link on RISC bit alignment: msdn.microsoft.com/en-us/library/aa294366(v=vs.60).aspx It's an old reference, but it illustrates the general principle.</p><p><br></p><p><br></p>
skane2600
<p>Given the fact that Windows 10S apparently requires different drivers even though it's running essentially the same OS on the same processors, it seems unlikely that Win32 apps that rely on special drivers are going to run "out of the box" on Windows on ARM.</p><p><br></p><p>I expect less than 100% compatibility and mediocre performance, but we'll see. People have been excited by this idea for years but it's not clear there's much real value over what we already have.</p>
skane2600
<blockquote><a href="#174684"><em>In reply to Tallin:</em></a></blockquote><p>It's depends on your definition of "driver". Any required installation is part of the driver deliverable. Two pieces of software that differ even in one byte are different. It doesn't matter what part of the software is different.</p>
Stooks
<blockquote><a href="#174696"><em>In reply to skane2600:</em></a></blockquote><p>"It's depends on your definition of "driver"</p><p><br></p><p>EXACT same driver. As in ZERO difference or effort from a engineering/coding perspective. They did NOT have to go back and re-write them just for Windows 10s. They are in different packages as in wrappers.</p><p><br></p><p>Not to mention….no one but schools are using Windows 10s. Joe consumer can't even buy Windows 10s.</p>
skane2600
<blockquote><a href="#174699"><em>In reply to Stooks:</em></a></blockquote><p>It sounds like you've never written an installer if you think that supporting installation doesn't require any extra effort from an engineering/coding perspective.</p><p><br></p><p>I understand that you are saying that the post-installation part of the drivers are identical but that doesn't make a driver as delivered exactly the same. If it did, there would obviously be no issue to resolve at all.</p>
skane2600
<blockquote><a href="#174792"><em>In reply to MikeCerm:</em></a></blockquote><p>So the XPS13 has more than 10 hours of battery life?</p>