Qualcomm will address the performance issues with first-generation Windows 10 on ARM PCs by engineering a new family of CPUs specifically for these products.
I actually saw this one coming, for a change. And if you also were paying attention to what Qualcomm announced so far this year, you might have noticed an interesting trend too. The mobile chip-making giant has started expanding its product lines by custom-designing chipsets specifically for different workloads instead of using the same chips for everything.
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
We saw this first with wearables, which I did report on. And we saw it with Internet of Things, which I did not. And I wondered, would we see this with PCs, too?
The answer, I was told, is yes. And I couldn’t be happier.
What this means is that second-generation, Qualcomm-based Always Connected PCs will not use the Snapdragon 845 processor, as we previously expected. Instead, they will utilize a new chipset, called the Snapdragon 850, which is being custom-tailored for PC use cases and optimized for performance.
Please read that again. Optimized for performance.
If you’ve followed along with my pre-release excitement about Windows 10 on ARM, and then my disappointment in actually using the first of these new PCs, you understand why this is such a big deal. Is, in fact, a potential gamer changer.
Windows 10 on ARM succeeds on a number of levels, but it suffers from two major usability issues: Application compatibility, which is tricky, and performance. Qualcomm is working with Microsoft and app makers to ensure that more applications work—and work well—when running on Snapdragon. But that latter issue, performance, is a bit trickier.
The performance problems we’re seeing today are caused in part because the first-generation ARM-based PCs utilize a Snapdragon 835 chipset which is optimized for smartphones. Its successor, the 845 is likewise optimized for smartphones.
So Qualcomm’s multi-year and multi-generational solution is to fork its Snapdragon chipset and create a new family of processors that specifically target PCs. The first of these is the Snapdragon 850.
It differs from the Snapdragon 835 in several important ways.
First, the 850 provides 30 percent faster performance on the PC, thanks to its faster Kryo 385 CPU cores, which run at speeds of up to 2.95 GHz, an improvement over the 2.6 GHz Kryo 280 in the Snapdragon 835. It also provides a 30 percent graphics performance boost.
Second, the 850 will provide 20 percent better battery life than its predecessor. Which is impressive, given that Snapdragon 835-based PCs already deliver about 20 hours of real-world battery life.
And third, Qualcomm is improving the speed of LTE-Advanced by about 20 percent, with the top download speed jumping to 1.2 Gbps thanks to a new Snapdragon X20 LTE modem.
At its Computex presentation today, Qualcomm will show off a reference design that it created for the Snapdragon 850. But the company expects the first 850-based PCs to hit the market in time for the holidays.
More soon.
skane2600
<blockquote><a href="#281206"><em>In reply to Winner:</em></a></blockquote><p>I agree, We also shouldn't ignore the possibility of "resume driven" development. Young developers who want to use whatever the "hot" technology is independent of business concerns to pad their qualifications. I suspect that's why we heard so much about HTML and JavaScript with regard to programming for Windows 8. Now the hot thing is "ARM" but it's not clear how Windows on ARM can help Microsoft's bottom line.</p><p><br></p><p>Update: I thought I made up the term "resume driven development" but it turns out it was already a thing: https://rdd.io/</p>
skane2600
<blockquote><a href="#281210"><em>In reply to lilmoe:</em></a></blockquote><p>While successfully implementing what you suggest could probably help Qualcomm, it doesn't really help Microsoft. For most people the combination of Windows and Intel is fine. The ARM solution would have perform significantly better than the Intel one to give people a reason to switch and even if they did, it would probably be a zero-sum game for MS (A Windows license on ARM, a Windows license on Intel, what's the difference?).</p><p><br></p><p>What you're really suggesting from a users's perspective is a brand new OS. Something that MS could have written years ago if it wanted to abandon it's cash cow. </p>
skane2600
<blockquote><a href="#281277"><em>In reply to Angusmatheson:</em></a></blockquote><p>Not exactly. One of the problems with RT was calling it "Windows" when from a user's point of view, it wasn't. Microsoft could have written a brand new legacy free mobile OS (unhampered by Windows, unhampered by *nix) designed from the ground up to be mobile. I can't say that it would have succeeded but it really couldn't have done any worse, IMO.</p>
skane2600
<blockquote><a href="#281358"><em>In reply to feedtheshark:</em></a></blockquote><p>Sounds more like wishful thinking to me than a technological solution. You essentially want more functionality (e.g. x86, Android) than current Windows in a "stripped down fast OS". There are always trade-offs and you can't just choose all you want independently when you design a product.</p>
skane2600
<blockquote><a href="#281891"><em>In reply to unkinected:</em></a></blockquote><p>I suspect that Microsoft took it on faith that the performance would be there. Some of the developers probably knew performance was going to suck, but sometimes in companies telling the truth to management can get you fired. </p>
skane2600
<blockquote><a href="#281768"><em>In reply to ghostrider:</em></a></blockquote><p>I think at best Windows on ARM would be revenue neutral unless they are planning to charge more for it than Windows on Intel (which would surely kill it). It's not as if they are going to get users who have no interest in traditional Windows to buy Windows on ARM. </p><p><br></p><p>As far as Win32 is concerned you may be right that they want to dump it, but to what end? Mobile is essentially dead to them so what benefit does a Win32-less Windows offer? Win32 compatibility is the only unique advantage that Windows offers. </p>
RR
<blockquote><a href="#281768"><em>In reply to ghostrider:</em></a></blockquote><p>First off nobody is trying to "save" anything, that is just typical blogger hyberbole. This is part of a road map that was conceived years in advance may be with an adjustment of 1 quarter here and there. </p><p>Second Apple just announced bringing iOS Apps to Mac, they have been going iPad pro for years, Google has been merging Android & Chrome of years now, etc. So, Microsoft should just stand still, because you know, I use Win32 and I don't want to be bothered by any changes. </p><p>These companies are responding to the opportunities and threats technological capabilities are presenting to them. And that is all WoA is about. It is not a luxury of choice. And no, WoA won't die, it will find its market niche(s), that may be large or small, but clear that are many permutations of this yet to be written.</p>
skane2600
<blockquote><a href="#282230"><em>In reply to RR:</em></a></blockquote><p>It's easy to see the parallels between what Apple and Google are doing (i.e. bringing their mobile apps to their desktop OS's) but they have nothing in common with what Microsoft is doing with WoA. </p><p><br></p><p>What new opportunity or threats does WoA address? I don't see any.</p><p><br></p><p>It's also worth keeping in mind that a small market niche is often the precursor to product death. We don't need to look any further than the Windows Phone to see an example.</p>