-
- diegoneiRetired Mod & AmbassadorAll I can see is that the javascript issue is still there (the hanging on the end of the loading. the page is all there, but javascript is still being processed so the bar doesn't finish loading).
For the rest, the test does not pitch both devices in the same condictions so it can't really be a reference. Have both devices on the same networkl both with sim cards in and then we can do testing. Even better if both are sim-less and on wifi, so signal strenght doesn't get in the way.12-16-12 07:06 PMLike 0 - Anybody notice how the bb10's keyboard takes up a little more screen space than the Iphone 5 and SG3?12-16-12 07:55 PMLike 0
-
-
-
So of course the Galaxy S3 is going to be faster because it has a 1.5 GHz dual krait core (which is faster than cortex A9) CPU vs. dev alpha B with a downclocked 1 GHz dual core cortex A9 CPU.
If you interpolate the sunspider scores, a fully clocked 1.5 GHz Blackberry 10 will have about the same sunspider score as the iphone 5 --- which is going to be a lot faster than the Android counterparts.12-17-12 12:32 PMLike 0 - Multiple crackberry members including xsacha (the guy who wrote the Playbook rooting software) had stated that the dev alpha A is downclocked from 1.5 GHz to 1 GHz. And dev alpha B users have posted about the same sunspider scores as dev alpha A users --- thus the dev alpha B is also downclocked from 1.5 GHz to 1 GHz.
So of course the Galaxy S3 is going to be faster because it has a 1.5 GHz dual krait core (which is faster than cortex A9) CPU vs. dev alpha B with a downclocked 1 GHz dual core cortex A9 CPU.
If you interpolate the sunspider scores, a fully clocked 1.5 GHz Blackberry 10 will have about the same sunspider score as the iphone 5 --- which is going to be a lot faster than the Android counterparts.
Case in point, the iPad and iPhone (pre iOS6 versions), when upgraded to iOS6 (which had a newer nitro javascript engine) showed about a 20% improvement in JS performance without any change in CPU at all.
Another example, the stock Android JB browser has considerably better Sunspider scores than Chrome running on the same device.
This shows that:
1. RIM can improve the JS performance without any changes to the actual hardware
2. You can't directly predict JS performance based on clock speed alone
3. Sunspider is not an accurate measure of overall CPU performance12-17-12 12:44 PMLike 3 - Sith_ApprenticeMod Team EmeritusYou may be right, but it's dangerous to interpret sunspider scores like that. The basic formula for calculating sunspider scores is not solely dependent on the CPU, and we really have no BB10 units running at a different clock speed to see the actual performance ratio (I don't think we do anyway, correct me if I'm wrong).
Case in point, the iPad and iPhone (pre iOS6 versions), when upgraded to iOS6 (which had a newer nitro javascript engine) showed about a 20% improvement in JS performance without any change in CPU at all.
Another example, the stock Android JB browser has considerably better Sunspider scores than Chrome running on the same device.
This shows that:
1. RIM can improve the JS performance without any changes to the actual hardware
2. You can't directly predict JS performance based on clock speed alone
3. Sunspider is not an accurate measure of overall CPU performancemithrazor likes this.12-17-12 01:01 PMLike 1 - You may be right, but it's dangerous to interpret sunspider scores like that. The basic formula for calculating sunspider scores is not solely dependent on the CPU, and we really have no BB10 units running at a different clock speed to see the actual performance ratio (I don't think we do anyway, correct me if I'm wrong).
Case in point, the iPad and iPhone (pre iOS6 versions), when upgraded to iOS6 (which had a newer nitro javascript engine) showed about a 20% improvement in JS performance without any change in CPU at all.
Another example, the stock Android JB browser has considerably better Sunspider scores than Chrome running on the same device.
This shows that:
1. RIM can improve the JS performance without any changes to the actual hardware
2. You can't directly predict JS performance based on clock speed alone
3. Sunspider is not an accurate measure of overall CPU performance
The last time Apple dumping a brand new javacript engine onto the webkit community was back in March --- the only thing that matter is whether RIM has incorporated this javascript engine into the BB10 browser.
New low level JavaScript interpreter to boost WebKit performance more than 200%
We are sufficiently close to the BB10 launch that I don't see any chance of RIM suddenly migrates to a different webkit JavaScriptCore branch right now --- and there is no new Apple dumping brand new javacript engine in the last couple of months. The source code that RIM posted was based on webkit source code back in August 2012. Nothing extraordinary happened from August to December.
https://github.com/blackberry/WebKit...0_09/ChangeLog
Right now, RIM is fixing bugs and maybe even adding new "chrome" features (like bookmarking abilities). What you see basically on the dev alpha's is going to be on the final BB10 browser --- as far as webkit core and JavaScriptCore are concerned.
PS: the dot's benchmark in the video is really about benchmarking the GPU --- which the other smartphones have dual core GPU's or triple core GPU's vs single core GPU on the dev alpha B.Last edited by samab; 12-17-12 at 01:54 PM.
mithrazor likes this.12-17-12 01:42 PMLike 1 - Webkit javascript performance is directly related to how close you are to the main webkit's JavaScriptCore development tree. Apple also has the advantage by rewriting the javascript engine from scratch in-house and then dump the entirely new (and formerly proprietary) javascript engine to the open source webkit community afterwards. Then other open source webkit participants have to spend months trying to understand those newly dumped formerly proprietary source code --- while Apple has months of head start and claims the javascript speed record. This is the main reason for these big Apple speed jumps.
The last time Apple dumping a brand new javacript engine onto the webkit community was back in March --- the only thing that matter is whether RIM has incorporated this javascript engine into the BB10 browser.
New low level JavaScript interpreter to boost WebKit performance more than 200%
We are sufficiently close to the BB10 launch that I don't see any chance of RIM suddenly migrates to a different webkit JavaScriptCore branch right now --- and there is no new Apple dumping brand new javacript engine in the last couple of months. The source code that RIM posted was based on webkit source code back in August 2012. Nothing extraordinary happened from August to December.
https://github.com/blackberry/WebKit...0_09/ChangeLog
Right now, RIM is fixing bugs and maybe even adding new "chrome" features (like bookmarking abilities). What you see basically on the dev alpha's is going to be on the final BB10 browser --- as far as webkit core and JavaScriptCore are concerned.
Again, you may be right that RIM will not do any further optimizations to the JS engine, but again there's nothing stopping them from doing so.12-17-12 01:55 PMLike 0 -
-
There is NOTHING to do further optimization on. The big speed difference is going to be whether RIM incorporated the new javascript engine (Low Level Interpreter). Any further optimization would be the speed increase between ios 6 to ios 6.01 to ios 6.1 --- not very much.
Webkit Core and JavascriptCore are two pieces of critical browser components that have to be fully betatested --- it means that RIM certainly won't spring a brand new webkit core or a brand new javascriptcore onto the final BB10 handset without testing them on the dev alpha.12-17-12 02:14 PMLike 0 - again brings me back to the score and performance. yes a higher score but thats numbers. looks good on paper. in real life looks to load a little more info than the i5 because of the full html so there for it ran a tad slower. id like to see them clear the cache of both devices and go to about 10 websites to see which is faster on average.12-17-12 02:36 PMLike 0
- again brings me back to the score and performance. yes a higher score but thats numbers. looks good on paper. in real life looks to load a little more info than the i5 because of the full html so there for it ran a tad slower. id like to see them clear the cache of both devices and go to about 10 websites to see which is faster on average.12-17-12 02:45 PMLike 0
-
The LTE Playbook which sports a full speed 1.5 GHz OMAP4460 but runs on ancient Playbook OS 2.01 --- gets a better sunspider score (1400 sunspider) than people with dev alpha's (which is around 1500 sunspider).
BlackBerry PlayBook 4G LTE hands-onzyben likes this.12-17-12 03:45 PMLike 1 - I think the real test will come when we have our hands on the actual hardware, NOT A Dev Alpha, and when we have some gold code, NOT some crippled alpha developer only software! That being said what is shown in the videos does look promising, and what little we have been shown of production like hardware has looked good too. I'm pumped!diegonei likes this.12-17-12 03:54 PMLike 1
- RIM has the technology to guarantee the fastest browsing experience. Just how it is folks
RIM patent app will have you barely browsing the web at incredible speeds12-17-12 04:00 PMLike 0
- Forum
- BlackBerry 10 Phones & OS
- BlackBerry 10 OS
Browser on BB10 "A game changer"
Similar Threads
-
Let us support the following dev to port his ios game on BB10
By kojita in forum General BlackBerry News, Discussion & RumorsReplies: 3Last Post: 11-13-12, 04:07 AM -
TK victory bb10 concept game changer opinions?
By nickthebold in forum General BlackBerry News, Discussion & RumorsReplies: 22Last Post: 09-04-12, 10:29 PM -
Google Chrome browser on BB10 devices?
By Marty_LK in forum BlackBerry PlayBookReplies: 4Last Post: 06-05-12, 07:04 PM -
Web browsers on 8830
By fasteddie#CB in forum BlackBerry 88xx SeriesReplies: 0Last Post: 09-16-07, 10:28 PM -
Surf two browsers on your bb
By tomdick in forum General BlackBerry News, Discussion & RumorsReplies: 7Last Post: 08-17-07, 07:59 AM
LINK TO POST COPIED TO CLIPBOARD