• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

ARM challenges Intel in PCs

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
ARM Chromebooks might become a bigger and bigger chunk of the market. That's definitely plausible, and meets the needs of lots of people. Windows on ARM, running emulated x86 and getting awful performance? I don't think anyone wants that.
 
Well they seem to be doing something right, as most of the tablets and phones in the world are using the ARM instruction set and more and more people are using such devices as their primary ones.
Yes they are doing marketing right,estimating performance and showing graphs that are not in scale...takes a lot of nerve to pull a stunt like that.
Yes everybody is using smartphones,I have one as well,doesn't mean it can do all I need which is why I also have a desktop.
 
Yeah they've come down to the level of what Intel has been doing for so long.
I don't understand the fuss here. The graphs are just cut-off at the bottom, they're otherwise linear in scale and match the predictions;

a4jM4tL.jpg


Also you need to understand that "estimating performance" means actually running the workloads through a cycle accurate simulator running on a server farm. Again I never saw issues with this.
 
I don't understand the fuss here. The graphs are just cut-off at the bottom, they're otherwise linear in scale and match the predictions;

a4jM4tL.jpg


Also you need to understand that "estimating performance" means actually running the workloads through a cycle accurate simulator running on a server farm. Again I never saw issues with this.
How much do you think fine-tuning of compilers adds to the variance of SPECint scores?
 
How much do you think fine-tuning of compilers adds to the variance of SPECint scores?
Do you mean compiler flags or actual difference between compilers? I'm not gonna go down that rabbit hole of an argument. The scores will be run with equivalent simple settings, which is the best apples-to-apples.
 
I don't understand the fuss here. The graphs are just cut-off at the bottom, they're otherwise linear in scale and match the predictions.
Sorry that wasn't clear, I was being ironic. Intel have been lying^W^Wdoing marketing BS for years, and they'll always win at that with their experience 😀
 
So is this the 5th or 6th time that ARM challenges Intel over the last 5 years?
Yea, another one of these threads. Somehow it is never pointed out by the OPs that the same criticism being directed to intel also applies to AMD, and that sales would be taken away from AMD as well. In fact, AMD would be hurt much more by x86 losing market share to ARM, since they are trying to get market share back, and have a lot less resources than Intel to weather the storm.
 
ARM Chromebooks might become a bigger and bigger chunk of the market. That's definitely plausible, and meets the needs of lots of people. Windows on ARM, running emulated x86 and getting awful performance? I don't think anyone wants that.
You can compile to ARM directly and I'm pretty sure that MS will start to encourage developers to do so soon.
 
Yea, another one of these threads. Somehow it is never pointed out by the OPs that the same criticism being directed to intel also applies to AMD, and that sales would be taken away from AMD as well. In fact, AMD would be hurt much more by x86 losing market share to ARM, since they are trying to get market share back, and have a lot less resources than Intel to weather the storm.
I think ARM still has a long way to go before we will see ARM based PCs suitable for use. Call me when I can build my own ARM PC just x86-64....
 
You can compile to ARM directly and I'm pretty sure that MS will start to encourage developers to do so soon.

You mean like how Microsoft has been encouraging developers to put their apps in the Windows Store... how's that working out for them?
 
You mean like how Microsoft has been encouraging developers to put their apps in the Windows Store... how's that working out for them?
Words cannot describe how much I hate the window's store. Even when I already have a program that does what I want, when I go to open a file, it wants to go to the Window's store first. Then it makes me browse to find the program I want. Seems like even when I check "use this program to open all files of this type" it seems to manage to send me back to the windows store after a while. But even with the annoyances of Windows 10, it is still far superior to Android for getting serious work done.
 
Microsoft could have done MUCH better on doing the hybrid Tablet/Desktop OS.

I don't know what they were smoking when making the Windows 8 interface. Of course, 8.1, and even 10 is a result of that.
 
As usual, you are going to have to wait until Apple does it (and does it right), so everyone else can copy them.

It doesn't require a genius to make it better.

I speculate one reason they went with the tile interface is because they want to somehow stand out. Well, if the usability is bad, then the uniqueness will work against you. Also, they put minimal effort on the interface. That's why you see idiotic crap like having two browsers and two control panels. It's like 8/8.1 was nothing but a 3rd party GUI add-on.

Use an Icon-based interface like Android and iOS. Heck, like Windows! Why is that such a bad thing to do Microsoft? Seriously. Their Pocket PC interface was better. Big companies fade into oblivion because they become more dumb the larger they get.
 
You mean like how Microsoft has been encouraging developers to put their apps in the Windows Store... how's that working out for them?

Thats not a store question - compiling for ARM is independent of the store. Currently for Windows for ARM you can compile the following native apps: ARM64 Win32 apps, ARM64 UWP apps - using Visual Studio or Clang as well as ARM64 Linux apps using gcc.
 
Thats not a store question - compiling for ARM is independent of the store. Currently for Windows for ARM you can compile the following native apps: ARM64 Win32 apps, ARM64 UWP apps - using Visual Studio or Clang as well as ARM64 Linux apps using gcc.

That's not what I was saying. My point was that Microsoft don't hold that much sway with developers any more.
 
That's not what I was saying. My point was that Microsoft don't hold that much sway with developers any more.

I am aware of this, but there is really not much to add here. Microsoft provided a comprehensive tool-set for native ARM compilation including native and remote debugger support - which makes it trivial for developers to compile for Windows on ARM. It has nothing to do with the question if Microsoft hold much sway with developers.
As example it took the VLC developers more than a year to port VLC-player to UWP, such that it can be put into the store. In contrast it took them just days to compile VLC player for Win32 ARM64.

In summary it comes down to market penetration of ARM Windows devices. In order to accelerate the trend - from my point of view - all passively cooled devices today should be ARM - as you get most performance out of the available power envelope.
As example, when running multi-core benchmarks on my Envy X2 (which has relatively old Cortex A73) it is killing everything in performance up to 15W actively cooled Intel/AMD devices.
 
Last edited:
I'm assuming that most of these if not all Windows ARM devices will be locked down where you can't install the software of your choice like Linux or one of the BSDs. I'm really hope I'm wrong about this...
 
Estimated SPECint scores for designs that are not yet finalized on nodes that are not yet shipping. By the time these appear in laptops they'll be against Intel's Whiskey Lake running at 31% higher clock speeds than the 7300U. The 8565U, a 14nm design, will be as fast as ARM's predicted 2020 5nm in single-thread performance. Sure, it'll use twice the energy but that's the price for being a few years earlier and two nodes behind.

For Windows, the emulation simply isn't fast enough and is awful for battery life. Apple and Google may end up with ARM-based laptops, since Google does web technology and Apple doesn't worry too much about backward compatibility. But Windows is stuck with x86 and its derivatives.
 
Last edited:
Back
Top