Why is Windows 10 Mobile not as smooth and fast as WP 8.1?

astondg

New member
Jan 7, 2013
409
0
0
Visit site
I would be interested to see the real difference between with & without animations. I think in a lot of cases the animations are adding minimal overhead and are there to hide system & app load times.

I am also keen to see some real metrics on the performance differences between Windows Phone 7, 8 & Windows 10 Mobile. My Lumia 920 certainly seemed to run better on WP8.1 than W10M but I didn't record any real metrics an the Lumia 920 isn't a W10M supported phone. To have discussion like this it would be good to have metrics, performing various tasks like booting, SMS, getting maps directions, etc., from the same device running an officially supported version of WP8.1 and W10M as well as a WP7 device.
 

milkyway

New member
Apr 16, 2014
764
0
0
Visit site
I would be interested to see the real difference between with & without animations. I think in a lot of cases the animations are adding minimal overhead and are there to hide system & app load times.

I am also keen to see some real metrics on the performance differences between Windows Phone 7, 8 & Windows 10 Mobile. My Lumia 920 certainly seemed to run better on WP8.1 than W10M but I didn't record any real metrics an the Lumia 920 isn't a W10M supported phone. To have discussion like this it would be good to have metrics, performing various tasks like booting, SMS, getting maps directions, etc., from the same device running an officially supported version of WP8.1 and W10M as well as a WP7 device.

you mean something like this? ;)
Benchmarking Redstone vs Threshold on the Lumia 1020/920 etc.
 

a5cent

New member
Nov 3, 2011
6,622
0
0
Visit site
I would be interested to see the real difference between with & without animations. I think in a lot of cases the animations are adding minimal overhead and are there to hide system & app load times.

When people thought of OS 'inefficiencies', the UI animations have ALWAYS been the first thing people instinctively pointed their fingers at. This has been the case ever since WP7. At least up until W10M, those people have ALWAYS been wrong.

You are right that some animations exist only to hide app launching times. The GPU plays no role in launching apps though. During that time the GPU would just sit there and "twiddle its thumbs" while the CPU loads and initializes a newly launched app. Instead of doing nothing, the CPU tells the GPU to use that time to animate some element of the UI. If the CPU happens to finish loading before the GPU finishes the animation, the animation is terminated prematurely. As a result, these UI animations neither add anything to the time/overhead required to load an app, nor do they make the OS any less efficient.

Animations are just an obvious thing to point at. So obvious in fact, that the developers were way ahead of the people pointing at them as possible inefficiency culprits...

...at least it's been that way so far. I don't know how W10M handles this. IMHO there's a good chance that W10M handles that NO worse than WP8 and WP7 did before it.
 
Last edited:

Tien-Lin Chang

New member
Feb 9, 2015
207
0
0
Visit site
To me it's one of the two(or both) reason:

1)short-sighted at product plan during WP8 days when they already knew (according to their story) they need buying time moving to this one-code one-platform goal. So they making all those old/low-spec devices just couldn't provide the raw power to running a more complex code (not meaning better functionality) than before.

2)limit capability of tuning the performance in old devices when the SoC are just too old and QCoM just not willing (MS not willing to pay the among QCoM wants?) to provide driver support for the new platform. So there's not much fine tuning can be done in terms of performance...

either way, it's a big shame to the PM responsible for all these. They knew they couldn't afford to reboot again yet they still plan their product line like this is just hard to justify for them....
 

TechFreak1

Moderator
May 15, 2013
4,611
5
38
Visit site
In essence WaaS doesn't allow much time for code "polishing" and not to mention older devices that are eligible for AU/Threshold will hardly get any firmware updates.
 

vezycash

New member
Jun 20, 2013
111
0
0
Visit site
Just why?
WP8.1 was Windows in name only. WP10 is very close to desktop Windows 10 and it inherits virtually all its problems: sluggishness, crash prone, ram hungry, hot device, 100% processor usage, OS rot (needing a reset often), apps which crash hardware (WP10 skype used to crash the camera),... eventually, WP would need antivirus as well.

Will it ever be?
Doesn't hurt to dream. Eventually, we all have to wake up and face reality

Everything was so smooth and great in 8.1, what happened?
Politics. 99% of reason why courier was canned - because it used a modified version of Windows XP. This move set Microsoft back in the tablet market for at least 5 years before Windows team could deliver a touch screen friendly OS.

The same thing happened with WP8.1 someone wanted to consolidate power at the expense of consumers. We got promises of a beautiful future but in the mean time...
 

a5cent

New member
Nov 3, 2011
6,622
0
0
Visit site
^ You'd have gotten a "like" from me if you hadn't gone quite so overboard. 😀 OS rot is actually one of the things the UWP definitively solves. Critisizing 100% CPU usage is also a bit strange, as a CPU core typically runs at either 100% or 0%. There is no in-between when it comes to CPU technology. Any CPU usage graphs telling you otherwise is dumbing the measurement results down and ultimately lying to you. You're also downright wrong on the bit about requiring antivirus software.

Basically, you're conflating the traditional Win32 and the more modern WinRT (UWP) runtime environments. Those are two very different things and W10M only has the later. Many of the issues you mentioned are specific to the Win32 environment which doesn't exist on W10M.
 

Padraig McGloin

New member
Jul 11, 2014
5
0
0
Visit site
I have the Windows 10, Redstone 1 release on my Lumia 930 and it seems to work perfectly, smooth, no restarts and the battery life has become weirdly awesome ... I have like 40% at the end of the day now.
 

astondg

New member
Jan 7, 2013
409
0
0
Visit site

No, because the 1020 & 920 aren't officially supported by W10M so the OS and firmware hasn't been optimised for them AND they were locked out of the latest Insider builds (i.e. the builds at the end of the cycle when Microsoft was making optimisations for release). To be fair to W10M it would have to be with a supported device (e.g. Lumia Icon, 1520, 930, 640, 640XL, 730, 735, 830, 532, 535, 540, 635 1GB, 636 1GB, 638 1GB, 430, 435, BLU Win HD w510u, BLU Win HD LTE x150q, MCJ Madosma Q501).

The results may still be the same but that's the kind of test we need to know for sure. I would also be interested in seeing a WP7 device in the test as well, performing the same tasks.
 

EspHack

New member
Jun 11, 2013
1,279
0
0
Visit site
No, because the 1020 & 920 aren't officially supported by W10M so the OS and firmware hasn't been optimised for them AND they were locked out of the latest Insider builds (i.e. the builds at the end of the cycle when Microsoft was making optimisations for release). To be fair to W10M it would have to be with a supported device (e.g. Lumia Icon, 1520, 930, 640, 640XL, 730, 735, 830, 532, 535, 540, 635 1GB, 636 1GB, 638 1GB, 430, 435, BLU Win HD w510u, BLU Win HD LTE x150q, MCJ Madosma Q501).

The results may still be the same but that's the kind of test we need to know for sure. I would also be interested in seeing a WP7 device in the test as well, performing the same tasks.

I bet they get SMOKED by wp7 :p
 

mildor

New member
Jul 24, 2013
132
0
0
Visit site
I know, I used to hate android because it was always laggy and ugly, and Loved💘 Microsoft Windows Phone because it was fast and clean...
That all change, now android is not that laggy anymore and is much cleaner (still ugly) and Windows10Mobile is not that clean anymore and is laggy 😞
 

milkyway

New member
Apr 16, 2014
764
0
0
Visit site
No, because the 1020 & 920 aren't officially supported by W10M so the OS and firmware hasn't been optimised for them AND they were locked out of the latest Insider builds (i.e. the builds at the end of the cycle when Microsoft was making optimisations for release). To be fair to W10M it would have to be with a supported device (e.g. Lumia Icon, 1520, 930, 640, 640XL, 730, 735, 830, 532, 535, 540, 635 1GB, 636 1GB, 638 1GB, 430, 435, BLU Win HD w510u, BLU Win HD LTE x150q, MCJ Madosma Q501).

The results may still be the same but that's the kind of test we need to know for sure. I would also be interested in seeing a WP7 device in the test as well, performing the same tasks.

I don't think it matters if your phone is officially supported or not. That newer phones with Snapdragon 400 and above run W10M way better than the ones with Snapdragon S4 is nothing you can correct with firmware updates
 

astondg

New member
Jan 7, 2013
409
0
0
Visit site
I don't think it matters if your phone is officially supported or not. That newer phones with Snapdragon 400 and above run W10M way better than the ones with Snapdragon S4 is nothing you can correct with firmware updates

To be a fair test it has to be a supported phone. If I had one I'd run some tests myself but I've only got a 920 :(
 

a5cent

New member
Nov 3, 2011
6,622
0
0
Visit site
I would also be interested in seeing a WP7 device in the test as well, performing the same tasks.
There are of course huge hardware differences between WP7's and today's modern devices. I'm sure you are aware, but still want to point out that a stop watch comparison between a WP7 device and a modern one running W10M would tell us very little.

Efficiency is about how much computing power must be brought to bare to complete a certain amount of work in a given amount of time. Consider two identical hardware platforms with different OSes, where both are set to complete the same task (work). If one completes the job quicker, then that one is doing the work more efficiently (and will require less battery power to do so).

WP7 devices were at most dual-core 1.2GHz devices with 512MB of RAM. Yet their UI was just as responsive and at least as fluid (if not more so) than what we have today, despite being comparatively under powered (massively). I'm guessing it took high-end WP7 devices about 50% longer than my L830 to launch an app, but they had less than a fourth of the computational power, so the WP7 device is still more efficient despite being slower. If W10M was similarly efficient, our current devices would tear up their modern counterparts. My point is that just timing tasks, without taking the hardware into consideration, won't tell you anything about efficiency.

Otherwise I'm completely with you. Firmware can make huge differences in performance so any test should use hardware that is officially supported.
 
Last edited:

EspHack

New member
Jun 11, 2013
1,279
0
0
Visit site
I don't think it matters if your phone is officially supported or not. That newer phones with Snapdragon 400 and above run W10M way better than the ones with Snapdragon S4 is nothing you can correct with firmware updates

steve on AAWP has actual stopwatch numbers for app startups and such things, sd400 just barely tops s4 phones, sd200 is way behind both, so we can only wonder why 435 got approved while 92x didnt

some people seem to think a generational improvement on cpu means a 100% speed increase, far from that, its just that things started to be done for quad core cpus not dual cores, cores on s4 cpus are clock for clock faster than sd400 and they tend to be clocked higher too not to mention its more than that, fundamentally we have a product intended to rule(s4) and another intended to be good value(sd400) if you look at memory, nand, gpu and other things, the s4 absolutely crushes the sd400, and in simple terms: you cant expect todays i5 to beat yesterdays i7 just because its new, in fact, just go check that out, you might be surprised
 

astondg

New member
Jan 7, 2013
409
0
0
Visit site
There are of course huge hardware differences between WP7's and today's modern devices. I'm sure you are aware, but still want to point out that a stop watch comparison between a WP7 device and a modern one running W10M would tell us very little.

Yeah, definitely. It's an important point and I don't want to draw any conclusions about WP7 OS based on that comparison, I just think it would be interesting to see how WP7, on the devices at the time, performed in the same types of tasks. Some of the difference could even be the UI structure of WP7 vs 8.1 & W10M (e.g. the People Hub integrations) so not a fair test of OS level efficiency from that point of view either, just interesting to see how long typical user interactions take across the generations.

But the real test I want to see is WP8.1 vs W10M on the same, supported, device performing the same tasks.
 

Tien-Lin Chang

New member
Feb 9, 2015
207
0
0
Visit site
I don't think it matters if your phone is officially supported or not. That newer phones with Snapdragon 400 and above run W10M way better than the ones with Snapdragon S4 is nothing you can correct with firmware updates

https://en.wikipedia.org/wiki/List_of_Qualcomm_Snapdragon_devices#Snapdragon_400_series

The S4 plus used by 920/925/1020/ATIV S/8X has two Krait 200 core pumping 3.3 DMIPS/MHz * 2 * 1.5Ghz = 9.9
The SD400 used by most of the lumia devices has four Cortex-A7 core pumping 1.9 DMIPS / MHz * 4 * 1.2Ghz = 9.12

It means S4 plus egdes the max raw power over SD400 (you have to be sure WP8/W10M can make fully use of the four core or it lose more)

The S4 plus has dual channel LPDDR2@533Mhz
The SD400 has single channel LPDDR2/3@533Mhz

It means in the best case SD400 has the same memory bandwidth of S4 plus, or it has only half of the later one.

Adreno 305 got about same power as Adreno 225

So yes, if SD400 run W10M way better than S4 plus, it's firmware/driver issue rather than SD400 being much powerful than S4 plus and it's the issue that MSFT who failed(unwilling?) to solve, not couldn't be solved.

When discuss something, we shouldn't confuse the result with reason.
 

paul87russell

New member
Aug 23, 2014
26
0
0
Visit site
Tbh as the platform moves forward, I think we'll become somewhat android-ish. In the sense that your experience will depend on what kind of SD you have and how much RAM..

Not allowing 512mb phones to officially upgrade and keeping a certain threshold for phones that are allowed to use continuum, are (at least according to me) indications of this

That is quite true IMHO. I read an article I can't quite remember where, i think in android central, where samsung submitted a patent for dual/simultaneous boot of android and windows in their phones
 

milkyway

New member
Apr 16, 2014
764
0
0
Visit site
https://en.wikipedia.org/wiki/List_of_Qualcomm_Snapdragon_devices#Snapdragon_400_series

The S4 plus used by 920/925/1020/ATIV S/8X has two Krait 200 core pumping 3.3 DMIPS/MHz * 2 * 1.5Ghz = 9.9
The SD400 used by most of the lumia devices has four Cortex-A7 core pumping 1.9 DMIPS / MHz * 4 * 1.2Ghz = 9.12

It means S4 plus egdes the max raw power over SD400 (you have to be sure WP8/W10M can make fully use of the four core or it lose more)

The S4 plus has dual channel LPDDR2@533Mhz
The SD400 has single channel LPDDR2/3@533Mhz

It means in the best case SD400 has the same memory bandwidth of S4 plus, or it has only half of the later one.

Adreno 305 got about same power as Adreno 225

So yes, if SD400 run W10M way better than S4 plus, it's firmware/driver issue rather than SD400 being much powerful than S4 plus and it's the issue that MSFT who failed(unwilling?) to solve, not couldn't be solved.

When discuss something, we shouldn't confuse the result with reason.

but the SD400 was already faster than the S4 Plus on WP8.1: https://www.youtube.com/watch?v=1fy4BCHB0eY
 

Tien-Lin Chang

New member
Feb 9, 2015
207
0
0
Visit site
That score means nothing if we don't know how they test it and how the score had been put together.

Like adreno 305 support a little bit newer graphic standard than adreno 225 so if you heavily score that part then you can generates the illusion of 305 smashing 225 which is not the case in real life. Its an old trick in those 3dmark days in 1998-2006(maybe still in today). The only thing I think SD400 devices stands a chance is the storage, it's newer so it has the chance to fit in faster NAND(I have no idea what Nokia put in during their low-spec days).

Still I think it's a shame if the raw power is there yet failed to deliver
 

Members online

Forum statistics

Threads
322,919
Messages
2,242,896
Members
428,005
Latest member
rogertewarte