New Relic Studies Mobile Brower Performance

[nicehat]

 

Bad test, don't take it too seriously.

 

I assume modern browsers have such low overhead that hardware specs become irrelevant after a certain point. I think all modern mobile hardware is more than able to handle browsers without bottle necking it. The browser itself is the weakest link in the equation.

 

The same can be said on the other end of the spectrum. For example, the Q10 has a 1.5 Ghz Dual Core CPU. The Galaxy S4 also has a Snap Dragon but it's quad core and clocked @ 1.9 Ghz. The S4 already out specs the Q10 but has inferior browser benchmarks. Why? 

[ThePointblank]

I assume modern browsers have such low overhead that hardware specs become irrelevant after a certain point. I think all modern mobile hardware is more than able to handle browsers without bottle necking it. The browser itself is the weakest link in the equation.

 

The same can be said on the other end of the spectrum. For example, the Q10 has a 1.5 Ghz Dual Core CPU. The Galaxy S4 also has a Snap Dragon but it's quad core and clocked @ 1.9 Ghz. The S4 already out specs the Q10 but has inferior browser benchmarks. Why? 

Also consider that Blackberry has two browers in this comparison; Blackberry Browser 5.0, and Blackberry Opera Mini 7.1. The Blackberry 5.0 brower demolishes the Blackberry version of Opera Mini 7.1 by a long shot.

[LAwLz]

I assume modern browsers have such low overhead that hardware specs become irrelevant after a certain point. I think all modern mobile hardware is more than able to handle browsers without bottle necking it. The browser itself is the weakest link in the equation.

 

The same can be said on the other end of the spectrum. For example, the Q10 has a 1.5 Ghz Dual Core CPU. The Galaxy S4 also has a Snap Dragon but it's quad core and clocked @ 1.9 Ghz. The S4 already out specs the Q10 but has inferior browser benchmarks. Why?

The less overhead the browser has the more the hardware spec matters, not the other way around.

How can the browser be the weakest like anyway? Put enough hardware behind a slow browser and it will sooner or later outperform even a much more optimized browser (running on lower end hardware).

I would like to see some benchmarks showing the Q10 beating the Galaxy S 4 by the way.

[Murdoch]

how can you say there's not enough information in the source to come to a conclusion about the validity of blackberry10 but at the same time say that it is automatically biased towards less used browsers?  Either there is enough information to start drawing conclusions or there isn't.

 

The 'lack of information' comment was in regards to the conditions of the data collected. I even wrote out what it was lacking.

 

Although I can draw some conclusions about how the data collected could potentially skew results simply on the numbers alone.

 

My point was comparing 1,000 blackberry users to 750,000 android users (for example) is likely to give an advantage one way or the other (unless it's under standardised conditions). I based the 'biased towards less used browsers' on assumptions about the results and data set, this wasn't meant as a conclusion, I should have worded it very differently.

 

You're right, I can't categorically draw any conclusions because there's a lack of information.

[ThePointblank]

The less overhead the browser has the more the hardware spec matters, not the other way around.

How can the browser be the weakest like anyway? Put enough hardware behind a slow browser and it will sooner or later outperform even a much more optimized browser (running on lower end hardware).

I would like to see some benchmarks showing the Q10 beating the Galaxy S 4 by the way.

I've fired up Rightware Browsermark 2.0 on my Blackberry Z10 just now, and my score was 2573. Looking at the Rightware BrowserMark power board, the Samsung Galaxy S 4 comes in at around 2692 to 2942 depending on the exact brower (Mobile Google Chrome 28 or 31). The Blackberry 10 score's average comes in at 2644.

 

For something that has twice the cores, a 400MHz advantage, plus a more advanced GPU that has twice the ALU's compared to the Blackberry Z10, that's not much of an advantage. Depending on the brower, a supposedly much faster and more advanced phone either barely squeeked through with less than 2% performance at the low end, or just above 10% performance difference in terms of browser performance in Browsermark 2.0.

 

And that's with the known issues with Samsung phones cheating on a number of benchmarks.

 

So, in short, it appears that the Blackberry browser is remarkably well optimized browser that's performing very damn closely on lower end hardware.

[nicehat]

The less overhead the browser has the more the hardware spec matters, not the other way around.

How can the browser be the weakest like anyway? Put enough hardware behind a slow browser and it will sooner or later outperform even a much more optimized browser (running on lower end hardware).

I would like to see some benchmarks showing the Q10 beating the Galaxy S 4 by the way.

 

While in the PC world there will always be a place for high-end hardware and pushing the performance envelope, the vast majority of what we do is now, especially in mobile is far more dependent on an Internet connection than on 8 cores of processing power or the latest version of Windows running on a Wintel system. How users access the web - whether from a mobile device, a Macbook, a Chromebook, or a DIY desktop -- has become a matter of preference. The emphasis is less on hardware and more on the mobile browser itself and the internet connection that you use. Hardware matters not after a certain point. Your Xeon machine wont load a complex page any faster than my 8350, or an i3 for that matter. 

[LAwLz]

Hardware matters not after a certain point. Your Xeon machine wont load a complex page any faster than my 8350, or an i3 for that matter. 

Yep, and people said the exact same thing back in 2000...

Seriously, you are sounding very dumb and naive right now. We have already been through this phase before, with people saying that you don\t need more than 128MB of RAM and a 500MHz CPU. Then new software comes out that can take advantage of the hardware and all of a sudden hardware is very important again.

 

"Those who fail to learn from the mistakes of their predecessors are destined to repeat them"

 

Hardware is, and always will be, very important. The only reason why we don't do complex tasks on phones is because they are not powerful enough. The more powerful they become, the more complex things we can use them for. Oh and yes, it would load a web page faster, or at least render it faster. The difference would be pretty small though, but not everyone only use their computers for web browsing.

[nicehat]

Yep, and people said the exact same thing back in 2000...

Seriously, you are sounding very dumb and naive right now. We have already been through this phase before, with people saying that you don\t need more than 128MB of RAM and a 500MHz CPU. Then new software comes out that can take advantage of the hardware and all of a sudden hardware is very important again.

 

"Those who fail to learn from the mistakes of their predecessors are destined to repeat them"

 

Hardware is, and always will be, very important. The only reason why we don't do complex tasks on phones is because they are not powerful enough. The more powerful they become, the more complex things we can use them for. Oh and yes, it would load a web page faster, or at least render it faster. The difference would be pretty small though, but not everyone only use their computers for web browsing.

 

 

Ok, just take a deep breath and chill with the compliments for a second. 

 

Done?

 

We are talking about mobile. I used the PC example to illustrate a very confined point about browsers. Not anything else. I dont use my phone to render anything, I use it to casually flip through web sites. Which is what 100% of the population does with their phones. HTML5 doesnt load ANY faster after a certain leap in hardware. And as you said, if it does, its not noticeable. Again, dont get it twisted. Re-read what I said in the context of mobile browsers.

[LAwLz]

Ok, just take a deep breath and chill with the compliments for a second. 

 

Done?

 

We are talking about mobile. I used the PC example to illustrate a very confined point about browsers. Not anything else. I dont use my phone to render anything, I use it to casually flip through web sites. Which is what 100% of the population does with their phones. HTML5 doesnt load ANY faster after a certain leap in hardware. And as you said, if it does, its not noticeable. Again, dont get it twisted. Re-read what I said in the context of mobile browsers.

Oops, I thought I was in the thread about the new Galaxy S 5 specs, where everyone was saying "we don't need better hardware in phones!". Sorry about that.

Anyway, yes better hardware does improve the web browsing, especially if you take some of the new HTML5 tags into consideration.

Why do you think hardware doesn't matter? Things like the NAND flash performance matters A LOT for things like startup time. In Android the compiling speed matters as well (since it uses crappy JIT, which is heavily CPU bound.