Surface 3 powered by cherry trail

From the few reviews out I could think that INtel is not really trying as hard as it claims it does its Atom line.
Previous In Order hyper threaded cores remained in the wild mostly unchanged for a good while and it seems that it might be the same with their newer Out of Order cores.
The improvements upon the last renditions of their Atom @22nm ( the ones embarking a PowerVR GPU) appear to be pretty marginal.

Underwhelming...
 
liolio said:
From the few reviews out I could think that INtel is not really trying as hard as it claims it does its Atom line.
Previous In Order hyper threaded cores remained in the wild mostly unchanged for a good while and it seems that it might be the same with their newer Out of Order cores.
The improvements upon the last renditions of their Atom @22nm ( the ones embarking a PowerVR GPU) appear to be pretty marginal.

Underwhelming...
Click to expand...

Wasn't baytrail 6watts and now cherry trail is 2 watts. it seems to have increased in speed while they were at it
 
I think it's BayTrail 6W TDP vs. Cherry Trail 2W SDP.
Which means Cherry Trail's TDP could even be 8W, for all we know.
 
My 2 tablets with Atom Z3xxx have a 2W TDP/SDP. Though the Dell V8P actually goes almost to 3W when the IGP and CPU are loaded, according to HWMonitor. Ironically the bigger 10.1" HP Omni 10 locks harder to 2W.

Anyway it's clearly a major performance limiter. Atom Z3770 has worse GPU performance than some of the slower Z37xx models because of the faster CPU clocks reducing IGP clockability.
 
eastmen said:
Wasn't baytrail 6watts and now cherry trail is 2 watts. it seems to have increased in speed while they were at it
Click to expand...
Luckily there was some gains while moving from 22nm to 14nm.
Now it seems there are no significant architectural change and performances were more an issue than power. Actually Intel SOC were doing quite well under sustained load (and I guess even better now).

The issue is from their last 22nm SOC (embarking PowerVR GPU) I suspect that the user experience is mostly the same.
Things could get improved, multi-core scaling ain't that great from 2 to 4 cores, SIMD could see an update, Intel could now afford a slightly wider cores, etc.

At the other hand of the spectrum there are those low power Broadwell cores. They can ( and do) throttle a lot but Intel pushes the top clocks pretty high, yet their performances are unmatched by competing designs.

So I see 2 options Intel should have made better that those Cherry Trail (so improve on Silvermont micro architecture) or simply end that line of CPU:
I know it sounds crazy BUT they do have now others CPU for extremely low power operations AND their main line (unmatched) CPUs are getting a reasonable option (not to mention that if one wants Windows... sadly AMD offering is lacking... so no choice).

Core M are pretty impressive yet they are a tiny package 81mm2 iirc, there are no SOC though. In my opinion to really make a strong showing they have design a SOC based on their main line CPU (so Broadwell) something like that:
High-end: 1 Broadwell core, 2MB of L3, 16EU GPU, a 64 bit memory interface (single channel) supporting DDR4.

Intel "needs" to create further segmentation with its higher end Core M line (generating higher margins) I believe they should have done it through any combination of the options below (as usual I would add):
1) significantly lower max turbo clock speed (north of 2GHz instead of south of 3GHz).
2) Varying level of of SIMD support (/crippling).
3) On/Off X86_64 support.
4) On/Off SMT support.
5) On/Off DDR4 support
 
In 2015 forward, making a single-core SoC for an OS-carrying consumer product would be terrible in terms of performance/power.
 
Arnold Beckenbauer said:
Its SDP is 2 W, http://ark.intel.com/de/products/85475/Intel-Atom-x7-Z8700-Processor-2M-Cache-up-to-2_40-GHz

http://arstechnica.com/gadgets/2015/04/surface-3-review-smaller-slower-cheaper-better/2/
Click to expand...

Nice, better battery life than the iPad Air 2. Granted it's also bigger and heavier. But the ability to run all the productivity software I have in addition to it serving as a high end tablet makes this incredibly compelling as a content consumption tablet that can also get some work done. Pre-ordered.

Thank god, I'll never have to use a 16:9 tablet ever again.

If Surface Pro 4 comes with Skylake later this year, that might be interesting also from a light gaming perspective.

Regards,
SB
 
I wouldn't preorder it. Wait and see if it has any horrible problems upon release. I don't believe the hype and previews of anything anymore.

What are the popular touch-based Windows games these days? I haven't gamed much on my Windows tablets. I tried Shadowrun Returns for a bit, which is touch optimized, but it was still too annoying compared to playing on a real PC. Baytrail IGP can't handle much anyway. Actually I think 2D games are preferable because they don't wolf down the battery or heat up your hands.
 
Last edited:
liolio said:
At the other hand of the spectrum there are those low power Broadwell cores. They can ( and do) throttle a lot but Intel pushes the top clocks pretty high, yet their performances are unmatched by competing designs.
Click to expand...

They can throttle, but they don't have to. It's entirely up to the OEMs that design the devices as to whether it throttles or not. Anandtech recently did a nice piece about this.

It is entirely possible to design a Core-M device that is both fanless and never throttles (Asus UX305 in the Anandtech piece). It is also possible to design Core-M devices that throttle in workloads the device isn't intended to run (intended by the OEM).

For example, the Dell Venue 11 has a plastic shell which greatly limits the cooling ability of the device. It is going to throttle on sustained workloads. But this device isn't targetted at heavy and/or sustained workloads. In workloads the device is targetted at by Dell, it operates just fine. The max turbo is effective at making the device feel snappy, responsive and fast. Since those workloads (office applications, web browsing, watching a movie, etc.) don't feature sustained loads, it also means that it has ample time to cool off before heavy throttling kicks in.

If Dell wanted, they could have used a more expensive fully metal body to greatly increase heat dissipation. But they went with cheaper plastic.

Here's the thing. Core-M wasn't created for heavy sustained workloads. It isn't designed for heavy sustained workloads. It isn't marketed as a CPU for heavy sustained workloads. And thus, it isn't going to be particularly good for heavy sustained workloads. Broadwell-U is the CPU you want if heavy sustained workloads is what you want.

Core-M was designed for fanless devices with tablets being a major focus. As such it is designed for workloads that you most typically see from tablet users. Workloads that often feature quick bursts of activity and then significant idle time and/or low sustained demands on the CPU/GPU. And for the workloads it was designed for, it excells. For the ones it wasn't designed for, it's going to stumble. But even there an enterprising OEM can make it perform in workloads it wasn't designed for (but still slower than a CPU that was designed for it).

Regards,
SB
 
swaaye said:
Why preorder it? Wait and see if it has horrible problems first. ;)
Click to expand...

It's a 3rd generation Surface device. Everything I've read about it thus far has it following in the Surface 3 Pro's footsteps as a mature, well thought out, and excellently executed device.

The only cons I ever see are...

Accessories are an extra that greatly increases the price. They think the pen should be included (I agree) and the keyboard cover should be included (would be nice, but at this price point?). They'd like the kickstand to have the Surface Pro hinge, but understandable that it doesn't for a cheaper device. And that it isn't faster (but that's what the Pro line is for).

Build quality, feel, useability, screen, basically everything else has either been praised or deemed adequate.

Battery life at this point was the only thing that might have put me off getting it.

It's basically like a premium iPad, only the Surface 3 can run everything and supports an active digitizer pen. Yeah, yeah, it doesn't have all the mobile apps that the iPad does. But with access to desktop apps and the ability to use any add on device you could possibly want to use, who cares? And it's not as much of a fashion statement as an iPad, but again, I just don't care.

Regards,
SB
 
Silent_Buddha said:
It's a 3rd generation Surface device. Everything I've read about it thus far has it following in the Surface 3 Pro's footsteps as a mature, well thought out, and excellently executed device.

The only cons I ever see are...

Accessories are an extra that greatly increases the price. They think the pen should be included (I agree) and the keyboard cover should be included (would be nice, but at this price point?). They'd like the kickstand to have the Surface Pro hinge, but understandable that it doesn't for a cheaper device. And that it isn't faster (but that's what the Pro line is for).

Build quality, feel, useability, screen, basically everything else has either been praised or deemed adequate.

Battery life at this point was the only thing that might have put me off getting it.

It's basically like a premium iPad, only the Surface 3 can run everything and supports an active digitizer pen. Yeah, yeah, it doesn't have all the mobile apps that the iPad does. But with access to desktop apps and the ability to use any add on device you could possibly want to use, who cares? And it's not as much of a fashion statement as an iPad, but again, I just don't care.

Regards,
SB
Click to expand...

The only Windows "apps" I use are IE, Mail, Weather and Netflix. I spend almost all time on the desktop or in IE Modern. If Chrome was better at touch typing (auto complete/correct) and as smooth at the page swiping as IE, I'd use it at the desktop instead of IE and Mail.

The 3:2 aspect ratio is nice. My Windows tablets are 16:10. I had a 3:2 Nook HD+ for a long time though and liked that for reading even more. Also did 4:3 at one point. Anything other than 16:9 works for me.

It does sound like a nice device..... The problem is I have a ~$250 cut off. I got my HP Omni 10 for $180 refurbed. ;) I've had a lot of tablets and I don't see them as devices worthy of big money. Lots of compromise and lots of damage potential.
 
Last edited:
I am legitemately surprised at how badly Chrome behaves in WIndows 8 and the touch interface. I suppose this is Google's way of thumbing their nose at Microsoft's tablet OS, but it's just a sad state. I've become a Metro IE user specifically because of how badly Chrome behaved, and how well Metro IE works.

On my BigAss(TM) desktop box, and on my work laptop, I'm still a die-hard Chrome fan.
 
Silent_Buddha said:
It is entirely possible to design a Core-M device that is both fanless and never throttles (Asus UX305 in the Anandtech piece).
Click to expand...
I'm not sure if that's always true. If you look at notebookcheck.com's review of the UX305, the QHD+ model never throttled, whereas the FHD model did (not only that but the graphics clock was significantly higher on the QHD+ model too while running the stress test).
http://www.notebookcheck.com/Test-Asus-Zenbook-UX305FA-Subnotebook.140385.0.html (stress test 300Mhz GPU, 500Mhz CPU)
http://www.notebookcheck.com/Test-Asus-ZenBook-UX305-Subnotebook.135684.0.html (stress test 400Mhz GPU, 800Mhz CPU)
Both indicate package power of roughly 4.5W, though the throttling model reaches 5 degrees higher temperatures (but afaik cooling is identical too). Maybe it's luck of the draw if the cpu can actually sustain base cpu clock with such loads or it's something else but the difference is rather large (considering both cpu and gpu, that would be like 40% faster at the same 4.5W sustained power draw).
 
I think Silent Buddha made its point,
Silent_Buddha said:
They can throttle, but they don't have to. It's entirely up to the OEMs that design the devices as to whether it throttles or not. Anandtech recently did a nice piece about this.
It is entirely possible to design a Core-M device that is both fanless and never throttles (Asus UX305 in the Anandtech piece). It is also possible to design Core-M devices that throttle in workloads the device isn't intended to run (intended by the OEM).

For example, the Dell Venue 11 has a plastic shell which greatly limits the cooling ability of the device. It is going to throttle on sustained workloads. But this device isn't targetted at heavy and/or sustained workloads. In workloads the device is targetted at by Dell, it operates just fine. The max turbo is effective at making the device feel snappy, responsive and fast. Since those workloads (office applications, web browsing, watching a movie, etc.) don't feature sustained loads, it also means that it has ample time to cool off before heavy throttling kicks in.

If Dell wanted, they could have used a more expensive fully metal body to greatly increase heat dissipation. But they went with cheaper plastic.

Here's the thing. Core-M wasn't created for heavy sustained workloads. It isn't designed for heavy sustained workloads. It isn't marketed as a CPU for heavy sustained workloads. And thus, it isn't going to be particularly good for heavy sustained workloads. Broadwell-U is the CPU you want if heavy sustained workloads is what you want.

Core-M was designed for fanless devices with tablets being a major focus. As such it is designed for workloads that you most typically see from tablet users. Workloads that often feature quick bursts of activity and then significant idle time and/or low sustained demands on the CPU/GPU. And for the workloads it was designed for, it excells. For the ones it wasn't designed for, it's going to stumble. But even there an enterprising OEM can make it perform in workloads it wasn't designed for (but still slower than a CPU that was designed for it).

Regards,
SB
Click to expand...
I get your point it is all about trade offs, usages cases (and marketing concerns) though I did not wanted to make such an issue about it. You are right clock speed and turbo behavior should be let to OEM.
My point is indeed about case usages and how intel could further fine tune its chips (number of CPU cores, GPU slices and clocks speed) which lead me to question if Intel still needs Atom/Airmont architecture.

Edit: I read Anandtech review of that new dell tablet and those CoreM are indeed quite something. Airmont seems more and more pointless to me.
A 1core 8 Eu SOC, would be perfectly fine conf for lower end tablet
 
Last edited:
liolio said:
I think Silent Buddha made its point,

I get your point it is all about trade offs, usages cases (and marketing concerns) though I did not wanted to make such an issue about it. You are right clock speed and turbo behavior should be let to OEM.
My point is indeed about case usages and how intel could further fine tune its chips (number of CPU cores, GPU slices and clocks speed) which lead me to question if Intel still needs Atom/Airmont architecture.

Edit: I read Anandtech review of that new dell tablet and those CoreM are indeed quite something. Airmont seems more and more pointless to me.
A 1core 8 Eu SOC, would be perfectly fine conf for lower end tablet
Click to expand...

There's still a point for the Atom CPU line. It is less powerful but also far more frugal on power. This makes it easier to design the product with regards to size, heat dissipation, and cost. Sure you can make a Core-M product last as long on battery power, but you need more batteries, and thus more weight and a larger package. And better cooling. Etc.

It's a smaller chip and thus cheaper to manufacture. It's mean to compete with the mid and low end ARM offerings. It's a bonus that it also encroaches on higher end offerings. Although the GPU is still a weak point of the arch compared to the high end ARM offerings.

Core-M is there to straddle the line between laptop performance (for light workloads) and tablet portability.

Atom is there for tablet portability with the ability to run desktop apps.

Until the mid/low end tablet market doesn't exist, there's still a purpose for Atom. The Surface 3 represents the top end of the Atom product line spectrum. I'm sure there will eventually be other tablets based on the same CPU in the 150-250 USD price range. Something that is unattainable by Core-M, which I don't expect to hit price points much lower than 500 USD (maybe 400 USD when they start getting discounted, but certainly not much lower than that) in a tablet form factor.

Regards,
SB
 
swaaye said:
It does sound like a nice device..... The problem is I have a ~$250 cut off. I got my HP Omni 10 for $180 refurbed. ;) I've had a lot of tablets and I don't see them as devices worthy of big money. Lots of compromise and lots of damage potential.
Click to expand...

Yeah, it's a bit higher than I normally would pay for an Atom based tablet in the current market. But there's a myriad of features that have made me overcome that for this one. The aspect ratio, resolution, screen quality, build quality, battery life and active digitizer support are all key components for why I'm going to get one.

I don't expect to find another Windows tablet that combines all that into one package (and certainly not much cheaper if there was one). The only thing I'd wish for is a larger 12" screen. It would trade off tablet portability for a little more production oriented comfort, but I can make do with the 10" screen. And as a consumption device first and light productivity second, it's fine.

Regards,
SB
 
Albuquerque said:
I am legitemately surprised at how badly Chrome behaves in WIndows 8 and the touch interface. I suppose this is Google's way of thumbing their nose at Microsoft's tablet OS, but it's just a sad state. I've become a Metro IE user specifically because of how badly Chrome behaved, and how well Metro IE works.

On my BigAss(TM) desktop box, and on my work laptop, I'm still a die-hard Chrome fan.
Click to expand...

That's not the only thing. It's also badly coded for Windows (and apparently Mac). And likely on purpose.

I finally found out one of the possible reasons why it absolutely destroys battery life on Windows (and people are reporting it does the same on Macs).

https://code.google.com/p/chromium/issues/detail?id=153139

Apparently the bug/behavior was first reported to Google as early as 2010. Just lends credence to the theory that it's being deliberately done to make other devices look worse compared to Android/Chrome tablets/laptops as many reviewers love to use the Chrome browser. I really like some of the features in Chrome even if I really dislike some of Chrome's other behaviors. But I'm close to the point where I'm going to have to try something else as secondary browser. Really don't want to go back to Firefox but I might give that a try again. It has been over 5 years since I last used it, after all.

Regards,
SB
 
Last edited:
yups its been there for eons and there's an article on Ars that said a version of Chrome has made a workaround to fix it. But it seems the problem still persist.
 
You must log in or register to reply here.

Similar threads

P
Portable console for all PC games - PGS
2
Replies
30
Views
12K
StakanPortveina
S
E
Surface PRO pricing revealed !
8 9 10
Replies
199
Views
36K
dagamer
D
A
Halo 3 Forge: A Summary and Game Ideas
2
Replies
22
Views
13K
Tap In
T
A
Full transcript of John Carmack's QuakeCon 2005 Keynote
3 4 5
Replies
98
Views
32K
nemisis
N
Back
Top