why are smart phones and tablets so slow rendering webpages?

M

munkle

[H]F Junkie
Joined
Jan 16, 2005
Messages
11,799
why are smart phones and tablets so slow rendering webpages? Is it the cpu?
 
pretty much. If you look at browser benchmarks like sunspider, phones and tablets are about 10 times slower than PCs.

Code has a lot to do with it as well though, the latest ios 5 update made a dramatic improvement.
 
so how would an equally clocked arm cpu compare to an atom cpu? Is it just the clock speed or is it the architecture as well that makes it slower?
 
what device/OS/browser are you having issues with? I've never felt that my android phone is rendering slower then it should be. I switch between the default browser and dolphin HD on android (2.2) using a samsung intercept.
 
lilfiend said:
what device/OS/browser are you having issues with? I've never felt that my android phone is rendering slower then it should be. I switch between the default browser and dolphin HD on android (2.2) using a samsung intercept.
Click to expand...

any smart phone or tablet. Ipad 2, iphone 4, droid x, hp touchpad. All of have extremely slow browsers compared to a desktop/laptop.
 
munkle said:
so how would an equally clocked arm cpu compare to an atom cpu? Is it just the clock speed or is it the architecture as well that makes it slower?
Click to expand...

ARM was made to be as power efficient as possible, while x86, while Intel has made major strides in making it more efficient will likely never get it as low as ARM can make it. In Intel's favor however, because x86 is much more complex instruction wise it's often is a lot more powerful in what can be done per cycle.

But to just focus on clock speed or architecture isn't the correct way to gauge performance. Compiler choices, to operating systems, to familiarity of how each system works can affect performance. The gap will close with time.
 
A lot of it has to do with software as well. iOS 5 had large improvements, and ICS is supposed to bring large improvements as well. Mobile browsers still have some issues with Javascript as well.

I agree that the gap will close with time.
 
Take a look at the memory a typical desktop browser consumes. On a phone you have a fraction of that available. And there's no virtual memory either like on a pc, so a mobile browser is even more constrained.
 
Yeah the CPU is much much slower, little to no hardware acceleration, and maybe multi-threading isn't up to par either does not help.

Though I have to say using the generic browser with my Atrix the page renderings are quite fast, even with flash. Excessive flash use (ie: hardocp's main page, sorry but best example I have off-hand) does slow down scrolling considerably though, but still very usable.
 
munkle said:
so how would an equally clocked arm cpu compare to an atom cpu? Is it just the clock speed or is it the architecture as well that makes it slower?
Click to expand...

I've tried android on an atom. Things were actually really smooth UI wise. A single core atom scores around 170000+ on browsermark which is higher than all the dual core tablets. So x86 even in atom form which is pretty crappy still has an edge in performance.
 
Electrofreak said:
That's my point; the power of an Atom compared to an ARM SoC in a smartphone is moot if you can't run the Atom for more than a couple of hours on a 1500 mA battery.
Click to expand...

Using very crude math, I would give it 1.5 hours of battery life on a phone-size battery. :eek:

...2 if you go all-idle.
 
Last edited:
Zurginator said:
Using very crude math, I would give it 1.5 hours of battery life on a phone-size battery. :eek:

...2 if you go all-idle.
Click to expand...

I can't find the post, (I thought it was here on [H]ard|Forum) but a while back (maybe a year or so ago) I wrote a huge post comparing the wattage and TDP of the ARM Cortex-A8 and an Intel Atom, and tried to approximate processing potential but trying to compare MIPS across two different architectures using different instruction sets is like trying to compare apples to oranges.

I do vaguely remembering that my calculations placed the Atom at something like 1/6th as power efficient as the Cortex-A8.

Wish I could find it so I could compare the specs I cited to current-gen Atom and Cortex-A9 specs; I'd like to see if the gap is narrowing. :)

EDIT - And yeah, a little voice in the back of my head is saying "Didn't the Atom use about 1 watt? That would put it at 1.5 hours like Zurginator is saying."
 
Last edited:
MrCrispy said:
Take a look at the memory a typical desktop browser consumes. On a phone you have a fraction of that available. And there's no virtual memory either like on a pc, so a mobile browser is even more constrained.
Click to expand...

That's the issue right there. A PC browser is optimized for speed. A mobile browser is optimized for small(er) memory footprint. Don't forget, even though some high-end phones now have up to 1GB of RAM to play around with; plenty of low-end devices are limited to 256MB or less. Things need to be coded for the "least common denominator" so to speak.
 
Electrofreak said:
I can't find the post, (I thought it was here on [H]ard|Forum) but a while back (maybe a year or so ago) I wrote a huge post comparing the wattage and TDP of the ARM Cortex-A8 and an Intel Atom, and tried to approximate processing potential but trying to compare MIPS across two different architectures using different instruction sets is like trying to compare apples to oranges.

I do vaguely remembering that my calculations placed the Atom at something like 1/6th as power efficient as the Cortex-A8.

Wish I could find it so I could compare the specs I cited to current-gen Atom and Cortex-A9 specs; I'd like to see if the gap is narrowing. :)

EDIT - And yeah, a little voice in the back of my head is saying "Didn't the Atom use about 1 watt? That would put it at 1.5 hours like Zurginator is saying."
Click to expand...

Haha, I was much cruder with my math:
Typical netbook has a 6-cell 5200mAh battery. You typically see 6 hours out of that battery.

Typical smartphone has a 1500mAh battery. That's 3.5 times smaller. That means 1.7 hours of battery, and I rounded down due to the excessive amount of other factors I'm not going to bother looking into.
 
MrCrispy said:
Take a look at the memory a typical desktop browser consumes. On a phone you have a fraction of that available. And there's no virtual memory either like on a pc, so a mobile browser is even more constrained.
Click to expand...

On a PC the large memory usage is for bigger caches and back history. That doesn't really impact loading a new site any.

MrMike said:
A lot of it has to do with software as well.
Click to expand...

Nope, it's completely the hardware. iOS and Android both run the same software as their desktop counterparts. The difference is completely in hardware - particularly CPU and memory bandwidth.

I agree that the gap will close with time.
Click to expand...

Nope. The gap will always be there. However, smartphones may eventually be fast enough that it just doesn't matter, kind of like how even a low end PC is fast enough that hardware just isn't the bottleneck.
 
kllrnohj said:
Nope, it's completely the hardware. iOS and Android both run the same software as their desktop counterparts. The difference is completely in hardware - particularly CPU and memory bandwidth.
Click to expand...

iOS5 brought a sizeable improvement to web browsing performance for existing hardware. Android 4.0 is supposed to do the same. I think that's sufficient evidence there's still room to improve with software, no?
 
less resources to work with is why phones and tablets running a broweser are slower. ARM has a rather limited instruction set even compared to Atom but the up side is lower power. Available ram is also a HUGE factor, not to mention OS, speed of the internal ram, storage device (flash ram) diosplay size and the fact that all of the devices share the same pool of system ram that is S L O W.
 
I will say when I was playing with the Razr at the Verizon store before I ordered one I went to a few websites. I don't know if it was the 4G internet, or the newer processor, but I couldn't complain at all at how fast it loaded the webpages, including HardOCP. It was pretty blazing fast.
 
MrMike said:
iOS5 brought a sizeable improvement to web browsing performance for existing hardware. Android 4.0 is supposed to do the same. I think that's sufficient evidence there's still room to improve with software, no?
Click to expand...

It brought improvements in javascript, but those all applied to desktip as well . The software can improve, yes, but its the same as the desktop. There isnt a software delta between mobile and desktop is my point.
 
You must log in or register to reply here.
Back
Top