Related
web browser test on each tablet's native browser. Summary, the browser in Honeycomb is better than previous version. All three tablets are connected to the same wifi. check out the results yourself!
http://acid3.acidtests.org/
http://html5test.com/
http://www.webkit.org/perf/sunspider/sunspider.htm
Alternate Browsers
I ran those same on the Adam, on alternate browsers:
Netscape
Acid 3 97/100
HTML5 235+9
SunSpider 1820.3ms
Opera
Acid3 100/100
HTML5 234+8
SunSpider 1906.9ms
Both actually march or outperform the base Webkit browser in most things.
Anyone using JVQ after using JVP, what's the fuss about much improved browser? What has improved? Sunspider, scrolling speed, unzoom to tab previews?
Edit: So you don't have to read whole thread, JVQ browser is muuuuch smoother over already quite smooth JVP.
I dont using stock browser (because have not ability for gif animation, like in Froyo version Android), but somebody here in forum was say scrolling speed is signifited improved....
Regards
The browser in JVQ is vastly improved comparing to the one in JVP:
1) The scrolling is buttery smooth and it is not interrupted in the middle.
2) There is a capability to zoom out to tab view.
3) There are more animations, for example, when you open a tab there is a new animation.
4) The rendering is improved - phonearena website is "browse-able" now, it used to lag severely with JVP
In general it looks like samsung has managed to port back the browser from SGSII to SGS
NOTE: These are true only with odex stock rom. Deodexed rom has a broken browser - it lags too much.
Hope this helps
fungun1234 said:
The browser in JVQ is vastly improved comparing to the one in JVP:
1) The scrolling is buttery smooth and it is not interrupted in the middle.
2) There is a capability to zoom out to tab view.
3) There are more animations, for example, when you open a tab there is a new animation.
4) The rendering is improved - phonearena website is "browse-able" now, it used to lag severely with JVP
In general it looks like samsung has managed to port back the browser from SGSII to SGS
NOTE: These are true only with odex stoke rom. Deodexed rom has a broken browser - it lags too much.
Hope this helps
Click to expand...
Click to collapse
Yes i agree. i use stock JVQ rom. JVQ browser is better for me.
can someone upload browser apk?
I would like to try it on JVP before making a jump
Thanks
[email protected] said:
can someone upload browser apk?
I would like to try it on JVP before making a jump
Thanks
Click to expand...
Click to collapse
Don't do that, you will have to reflash your device.
I tried that myself and had a bootloop.
Why doesn't it work with deodexed?
Sent from my ever changing Galaxy S via XDA-Premium...
I can confirm this. On deodexed rom the browse is useless
NOTE: These are true only with odex stock rom. Deodexed rom has a broken browser - it lags too much.
Hope this helps
Click to expand...
Click to collapse
fungun1234 said:
The browser in JVQ is vastly improved comparing to the one in JVP:
1) The scrolling is buttery smooth and it is not interrupted in the middle.
2) There is a capability to zoom out to tab view.
3) There are more animations, for example, when you open a tab there is a new animation.
4) The rendering is improved - phonearena website is "browse-able" now, it used to lag severely with JVP
In general it looks like samsung has managed to port back the browser from SGSII to SGS
NOTE: These are true only with odex stock rom. Deodexed rom has a broken browser - it lags too much.
Hope this helps
Click to expand...
Click to collapse
JVP is much smoother than CM7 in scrolling so I guess this is imba fast. How about sunspider scores?
Here you go.
sunspider-0.9.1 scores:
============================================
RESULTS (means and 95% confidence intervals)
--------------------------------------------
Total: 6508.3ms +/- 2.1%
--------------------------------------------
3d: 979.1ms +/- 2.4%
cube: 379.0ms +/- 6.9%
morph: 322.6ms +/- 3.1%
raytrace: 277.5ms +/- 5.2%
access: 697.4ms +/- 6.7%
binary-trees: 30.0ms +/- 7.7%
fannkuch: 231.0ms +/- 7.9%
nbody: 343.2ms +/- 10.8%
nsieve: 93.2ms +/- 13.3%
bitops: 494.5ms +/- 13.5%
3bit-bits-in-byte: 33.4ms +/- 12.2%
bits-in-byte: 72.3ms +/- 9.9%
bitwise-and: 170.5ms +/- 25.5%
nsieve-bits: 218.3ms +/- 16.0%
controlflow: 25.3ms +/- 13.3%
recursive: 25.3ms +/- 13.3%
crypto: 428.9ms +/- 8.0%
aes: 170.9ms +/- 6.1%
md5: 131.6ms +/- 7.8%
sha1: 126.4ms +/- 18.6%
date: 927.6ms +/- 8.4%
format-tofte: 404.2ms +/- 17.4%
format-xparb: 523.4ms +/- 4.1%
math: 638.7ms +/- 8.1%
cordic: 276.4ms +/- 20.5%
partial-sums: 266.6ms +/- 2.7%
spectral-norm: 95.7ms +/- 6.3%
regexp: 144.5ms +/- 17.1%
dna: 144.5ms +/- 17.1%
string: 2172.3ms +/- 1.5%
base64: 210.6ms +/- 3.4%
fasta: 358.5ms +/- 2.0%
tagcloud: 356.4ms +/- 2.7%
unpack-code: 939.3ms +/- 2.9%
validate-input: 307.5ms +/- 1.5%
Can't understand why such a low scores compared to cm7. I guess something with JIT... As long as it performs smooth, I don't care.
So I'm currently on JVP from RomKitchen, I have no idea about this odexed/deodexed stuff. I just want to know if I can use the Browser from JVQ on my current setup, like just copy the APK over or something, because my current browser is quite laggy and glitchy to use. Thanks!
dylandylan1 said:
So I'm currently on JVP from RomKitchen, I have no idea about this odexed/deodexed stuff. I just want to know if I can use the Browser from JVQ on my current setup, like just copy the APK over or something, because my current browser is quite laggy and glitchy to use. Thanks!
Click to expand...
Click to collapse
No, you can't use apk from JVQ on JVP. At the moment, the only way to improved browser is to use stock odexed JVQ (deodexed JVQ breaks browser).
Well the biggest difference between JVP's and JVQ's browser is:
JVP: CRAP!
JVQ: AMAZING and smooth!
lol
I normally only use deodexed roms mainly they are a tad faster and alot of themes are for deodex only. But after using stock odex JVQ for now im sticking to it till the deodex lag is fixed. Browser is very smooth good improvement from JVP.
Even 720p video works in this new browser, wow
Even Vimeo in HD (flash) works flawless This GPU is finally fairly pushed to work
The new JVQ browser is as fluent as the iPhone 4 browser, if not better.
onetik said:
The new JVQ browser is as fluent as the iPhone 4 browser, if not better.
Click to expand...
Click to collapse
Yep. Thanks Samsung, great job on previous flagship model. HTC see and learn
Wow, browser on JVQ really is amazing!
So much improved... My SGS soon will be ideal
No matter what, Youtube videos HD stutter and lag intolerably. Running CM9 @ 1.5GHz, and on HC/ARHD HC it was perfectly smooth. It was also laggy on stock ICS. I believe this is an ICS issue, and was wondering if anyone has a fix.
Yes instal geuvors kernel test 13 and not 13b
Dont run from mobil site, click "desktop" in the lowest part of the site.
Or start videos in stick it player when browsing
Sent from my Transformer TF101 using Tapatalk
Got same stuttering here in browser and youtube app. Stock, latest revolver and Revolution HD roms with test kernel 14 and 15 on both. No difference. The cpu is through the roof when watching HD content. The "force gpu" setting on/off makes little difference to the overall framerate. Will try test13 and report back.
The youtube app played everything smooth in HD on Honeycomb, so I went back to HC to test if it ran smooth again. It did not for some reason. Don't know what to make of this..
It is perfectly smooth for me. Check what version of the YouTube app you are running in the Settings. The latest is 3.5.5. I found that for me cm9 did not identify itself correctly on the Market and the YouTube app would never update leaving me with an old version.
cholywell said:
It is perfectly smooth for me. Check what version of the YouTube app you are running in the Settings. The latest is 3.5.5. I found that for me cm9 did not identify itself correctly on the Market and the YouTube app would never update leaving me with an old version.
Click to expand...
Click to collapse
Please watch This in youtube app and report back again.
Maybe with high overclock it is semi smooth, but I don't believe it is as smooth as on HC. Check on your PC how smooth it CAN play for reference....
I have tried your video in my Youtube app, fullscreen and HD enabled.
I do have some stuttering, but really not that much. It might be happening every minute or so : the screen seems to freeze before it catches up. The rest of the time, it plays perfectly fluid.
And TBH, it is really bearable.
Mercvtio said:
I have tried your video in my Youtube app, fullscreen and HD enabled.
I do have some stuttering, but really not that much. It might be happening every minute or so : the screen seems to freeze before it catches up. The rest of the time, it plays perfectly fluid.
And TBH, it is really bearable.
Click to expand...
Click to collapse
Thanks for checking
What is your rom / kernel? What CPU speed?
PS yes between stuttering it is fluent, but the stuttering is quite annoying, I have it every 5-10 seconds or so in the beginning of the movie.
FF666 said:
Please watch This in youtube app and report back again.
Maybe with high overclock it is semi smooth, but I don't believe it is as smooth as on HC. Check on your PC how smooth it CAN play for reference....
Click to expand...
Click to collapse
So i tried with youtube app, stutter not really watchible
Opened with stick it videoplayer and it was 100% lagless. Also it was floating over the screen so that was awsome
Edit: The quality was not the same, even if HD was enabled
Sent from my Transformer TF101 using Tapatalk
FF666 said:
Please watch This in youtube app and report back again.
Maybe with high overclock it is semi smooth, but I don't believe it is as smooth as on HC. Check on your PC how smooth it CAN play for reference....
Click to expand...
Click to collapse
Edit: ignore me i accidentally clicked off HQ :X
Sent from my Transformer TF101 using Tapatalk
FF666 said:
Thanks for checking
What is your rom / kernel? What CPU speed?
PS yes between stuttering it is fluent, but the stuttering is quite annoying, I have it every 5-10 seconds or so in the beginning of the movie.
Click to expand...
Click to collapse
Running the last Megatron rom, with Guevor15 OC'ed at 1,4. The stuttering definitely doesn't happen as often as what you're reporting, but it must have to do with my overclock.
I don't really remember how well it might have worked on HC, cause I usually don't watch much Youtube videos of that quality...
Overclocking isn't going to help mutch since YouTube uses the gpu to render the video's
same problem here. Never stuttered on HC.
Running latest Android Revolution HD 3.2.1 [email protected]
Still stuttering and lag when there's alot of movement. If I overclock to 1500mhz it goes a bit smoother, but HC had no stutter (without overclocking).
Youtube app version is 3.5.5. Guess it's the youtube app that doesn't take full advantage of the gpu in ICS, or some nvidia driver issue.
Just have to live with it I guess, until Asus comes out with a fix. If they still continue to support this old tablet
EDIT: Using Setcpu [email protected] and governor Interactive i managed to get a bit smoother performance (test16 kernel). Lags a bit at the start of a video, and some stutter when there's ALOT of action/movement.
Try this, download sempervidlink from google play, it enables you to send videolinks like youtube to any player (mxplayer) and choose which quality you want it in. Open a actionfilled video to see if it helps
Sent from my Transformer TF101 using Tapatalk
anarox said:
Try this, download sempervidlink from google play, it enables you to send videolinks like youtube to any player (mxplayer) and choose which quality you want it in. Open a actionfilled video to see if it helps
Sent from my Transformer TF101 using Tapatalk
Click to expand...
Click to collapse
Thx! will try this later today.
I Actually went back to Revolver 3.11 Honeycomb with no OC. All videos run smooth without any stuttering (youtube app). I guess Honeycomb manages to use the GPU, while ICS uses the CPU. If you tick the "show cpu thingy" under dev options and play a HD video, the bars are through the roof. damn annoying! This is the only issue i've got with ICS. no deep sleep of death or random reboots hehe
-bana
I'll also have to try the above method, since I also get stuttering and clipping of the search bar (not sure if thats related)
Just use guevor's v20 kernel and he posted some NVIDIA codec updates. I think it will be included into v21.
I can (almost) confirm that the new codecs helps, it's not as smooth as on hc, I don't think, but it's certainly better.
The test video posted earlier is still choppy for me but other stuff that was choppy before is better.
Link to the codecs.
http://forum.xda-developers.com/showpost.php?p=26430343&postcount=1466
Guys. this is an ics issue. i can confirm on 4 different devices.
nexus s 4g gingerbread was all good, ics no
epic 4g touch gingerbread great, ics leaks, cm9, no good
stock galaxy nexus, no good.
now the transformer, hc good. ics no good.
i have almost 40mbps down at my house there is no reason it should stutter.
it seems to me the first 2 or 3 videos are fine, after that it lags. i have to shut it down and kill it for it to play right again.
Hello All
In post #19 I have an update with tests of Beta versions of most popular browsers.
I've been reading yesterday a browser comparison between Chrome, Firefox, IE and Opera on both Windows and Linux platforms and I've decided to do something similar for Galaxy Note 2.
Now for some starting info:
Rom - Phoenix v13
Kernel - Perseus 36.3
OC - yes, 1.8Ghz, GPU untouched, a bit of undervolt also.
The system is fairly pure, default speedups as found in Phoenix installed. Rom is ODEXED
Not much apps installed.
Contenders: all currently downloaded from Playstore
- Samsung Browser
- Chrome 28.0.1500
- Firefox 22.0
- Opera 15.0.1162
- Maxthon 4.0.6.2000 B2844
They were all put under the following tests:
- HTML5 Test - Link
- HTML5 Benchmark - Link - Basically a 2D game benchmark.
- Octane - Link
- Browsermark 2.0 - Link
- Mazesolver (Microsoft) - Link
- Lawnmark (Microsoft) - Link
There is a story about that benchmark involving cheating as stated here but it does not affect my comparison.
- Peacekeeper - Link
- Sunspider 1.0 - Link
- V8 Benchmark (succeeded by Octane but still valid and does offer different results) - Link
- Dromaeo - Link - be carefull, it takes about 20 minutes or more to complete.
So off to results:
...........XXXXX................Stock browser.....Chrome (+%).........Firefox (+%).........Opera (+%)..........Maxthon (+%)......Dolphin (+%)........UC (+%)
HTML 5 Test.................434.....................410 (94%).............422 (97%).............428 (98.6%).........419 (96.5%).......419 (96%).............404 (93%)..........more is better
HTML Benchmark..........43......................1911 (4400%)........2731 (6351%).......2043 (4751%)......77 (179%).........107 (248%)............113 (263%)........more is better
Octane..........................2402...................2900 (120%)..........2622 (109%).........2783 (116%).......2375 (99%)........2451 (102%)..........2441 (101.6%).....more is better
Browsemark 2.0............2525...................2427 (96%)............2306 (91%)...........2897 (115%).......2123 (84%)........1956 (77%)............1963 (78%)..........more is better
Mazesolver (20)............1.9......................4.9 (257%)..............6 (315%)..............6.3 (331%)..........1.6 (84%)...........2 (105%).................1.9 (100%)........less is better
Mazesolver (30)............9.6......................20 (208%)..............91 (948%)............25 (260%)...........4.9 (51%)...........5,6 (58%).................5,9 (61%)..........less is better
Mazesolver (40)............30.......................101 (336%)............112 (373%)..........109 (363%).........28 (93%)............29 (97%)..................31 (103%)........less is better
Lawnmark.....................538......................873 (162%)............1071 (199%)........679 (126%).........515 (96%)...........not_finished *2............not_finished *3..........less is better
Peacekeeper ...............873......................891 (102%)............605 (69%) *1........891 (102%).........927 (106%).........929 (106%)............857 (98%)..........more is better
Sunspider.....................1041....................812 (78%)..............863 (83%)............806 (77%)...........1160 (111%)........1080 (103.7%).......1050 (101%)........less is better
V8 Benchmark..............2690....................3101 (115%)..........2677 (99.5%).......3191 (119%)........2546 (94.6%)......2597 (96%)............2553 (95%)..........more is better
Dromaeo.......................167.15.................179.92 (107.6%)....146.42 (87.6%)....186.03 (111%).....168.6 (100.8%)...161.94 (97%).........164,44 (98%)..........more is better
*1 - this result is a strange one since Firefox is the only browser that allowed almost all videos to play, there were some problems with buffering so the videos skipped some but they worked, not one other browser did. So I don't understand why the score is so low, maybe other parts were bad.
*2 and *3 - in both cases browsers simply closed without any error sometime after starting benchmark (less than 3 minutes), consecutive tries resulted in closing few seconds after starting benchmark. Both browsers had to be fully restarted (as in killed in apps manager or by phone restart)
HTML5 test is only a measure of how well browsers conform to HTML standard, somewhat like old ACID3 test.
Now, HTML5 Benchmark on stock and Maxthon were a slideshow, <2 fps and it took more than 30 minutes to complete. I have no idea why as it is a simple 2D platform game and other browsers run quite well even up to 60 fps in case of Firefox.
Mazesolver had to be run more than 10 times for every settings and median calculated because tests were prone to big variations. As such I would mark that test as unreliable. Lawnmark, despite being unfairly skewed for IE 10/11 is an equal oportunity test for other browsers, so the results "stay"
To compare browsers I will count Octane, Browsermark, Peacekeeper, Sunspider, V8 and Dromaeo benchmarks.
Results:
Chrome is on average 10,43% faster than Stock Browser
Firefox is on average 4.48% slower than Stock Browser
Opera is on average 14,33% faster than Stock Browser
Maxthon is on average 4.43% slower than Stock Browser
Dolphin is on average 4.23% slower than Stock Browser
UC Browser is on average 5.06% slower than Stock Browser
And the Winner is ....... Opera
Now all those results should also be counted against power consumption and here is where things change a lot. I started those test with full battery and all the browsers did the same amount of work actually doing those tests. The problem is that while Chrome, Stock, Maxthon and Firefox required roughly the same amount of power, Opera used more than twice that. I'm attaching screenshots of those results.
View attachment 2133846 View attachment 2133847 View attachment 2133848 View attachment 2133849View attachment 2133855 View attachment 2133856 View attachment 2133857 View attachment 2133858
On other topic, I think almost 6 hours of screen on time with almost all that time CPU was working full tilt and screen was @100% is a great result. There were only short times when I left the phone to cool off, mostly with screen on but not working hard
Any comments or questions, just ask
Update for Naked Browser - the scores were in order: 2470 (103%), 2019 (80%), 824 (94%), 1087 (104.5%), 2729 (101.4%) and 168.69 (101%)
Total score for it: Naked Browser is on average 4.18% slower than Stock Browser
Next Boat Browser - the scores were in order: 2483 (103%), 2109 (83%), 850 (97%), 909 (87%), 2479 (92%), 165.98 (99%)
Total score for it: Boat Browser is on average 2.16% slower than Stock Browser
ONE Browser - the scores were in order: 2342 (97.5%), 2097 (83%), 680 (78%),1801 (173%), 1280 (47.5%), 166.62 (99.7%)
Total score for it: ONE Browser is on average 27.88% slower than Stock Browser
From all those scores of different browsers it's apparent that some of the use a common rendering engine (Firefox, Maxthon, Dolphin, UC and Naked and probably Boat too) because their results are within 1% of each other while Samsung, Opera and Chrome are much different. So between those above you can choose based on features set that they have because performance is about equal.
If you are interested in modern games based on HTML5 code, Firefox and One browsers are the way to go with Chrome and Opera about 30% slower. HTML5 Benchmark which I did not include in final score (but tested all browsers with) shows great disparity between browsers.
The scores:
Stock : 43 points
Chrome : 1911 points
Firefox : 2622 points
Opera : 2043 points
Maxthon : 77 points
Dolphin : 107 points
Naked : 96 points
UC : 113 points
Boat : 78 points
ONE : 3041 points
I do not understand how it's possible to get a 30 times difference in modern times in a fairly important category but there you have it. I hope it's not similar to Microsoft's Lawnmark benchmark that before every iteration followed (as is standard for some reason) 4ms pause while IE 10 and 11 allowed for "no pause" policy. The results - IE 10/11 finished in 15 seconds, other browsers took upwards of 10 minutes. The funny thing is, some Chrome addicts took the benchmark code and set a level playing field by removing the wait time for every browser, and lo and behold, IE was 2 times slower than Firefox and 3 times slower than Opera with Chrome somewhere in the middle.
It gives some thoughts as in both cases (HTML5 benchmark and Lawnmark) after benchmark run, the phone / cpu is completely cool. I do not at the moment have any monitoring app installed, but CoolTool would be easy to use to verify that.
In light of recent "strange" benchmark results of newer Intel chips that happend to be Intel compiler using a flaw in Antutu benchmark code all benchmark results can be suspected.
You can argue that compiler optimisations are the lifeblood of fast computing, but when you optimize code by hand to get best score in a benchmark, it's not clean fight. Because no other app is going to have the same optimizations so the benchmark result is misleading to the real life performance of the hardware. All that on top of usefullness/uselessness of benchmarks in the whole
And yet non of these browsers run as smoothly as the stock samsung browser when scrolling and zooming on any web page...
Where is dolphin?
True, Stock is smooth as butter, but there are no benchmarks that test smoothness so that conclusion is up to the user. I may test dolphin but there will be no power use tests for it because I would not be able to get comparable data.
Any other requests for browsers?
Hell Guardian said:
Where is dolphin?
Click to expand...
Click to collapse
2nd Dolphin and Naked Browser.
Sent from my GT-N7100 using xda app-developers app
I prefer the stock browser regardless of the statistics.
Nice comparison! I personally switch between Firefox and 'UC Browser'.
I find UC Browser really good for full screen/desktop content, but some websites don't display correctly.
I think I'm leaning towards boat - gestures are really helpful on Note 2 than it is on other phones. The only thing I miss is inverted browsing. There's a night mode which is in many ways better than inverted browsing, but in many sites it's not useable compared to the proper inverted browsing.
Also obviously much better control over flash plugins and everything. Seriously, give it a go.
dazza7111 said:
2nd Dolphin and Naked Browser.
Sent from my GT-N7100 using xda app-developers app
Click to expand...
Click to collapse
Can't test naked browser as it's not available for download in Poland (I mean free version). Paid is available but I won't buy it just to test it. (Or you can share .apk)
Dolphin is allready tested, I will update results later.
Sent from my GT-N7100 using xda premium
Naked browser is free on xda I think or The Dev website
Nice benchmarks!
I'd like to see Boat too.
Have been using it for some time and I really like it. It feels both fast and smooth!
Browser benchmarks give points for data preloading, using proxies, and other items like ability to load an HTML5 video placeholder.
Benchmarks do not evaluate browser integrity.
Naked Browser does not preload data, send your data through a proxy, or use resources on things like HTML5 video placeholders.
--Naked Browser developer
aminaked said:
Newer versions of Stock Browser (ICS+) preload web site data by default.
Naked Browser does not preload data. [As developer of Naked Browser I have scoured multiple versions of Stock Browser source code.]
The greater issue is Android web browser integrity. It turns out that even Firefox/Opera/Chrome are suspect.
Spying not only violates privacy but wastes battery, bandwidth, and CPU.
Click to expand...
Click to collapse
I agree, I also do not like to waste bandwidth and battery. On the other hand I do like to quickly open subpages..... in the end I will go with saving money For now If I may ask, on what code is Your browser based (mostly?) and is there any hope of porting this great speed in HTML5 execution (like in Firefox or One browsers) ?
mat9v said:
I agree, I also do not like to waste bandwidth and battery. On the other hand I do like to quickly open subpages..... in the end I will go with saving money For now If I may ask, on what code is Your browser based (mostly?) and is there any hope of porting this great speed in HTML5 execution (like in Firefox or Safari - One browsers) ?
Click to expand...
Click to collapse
I wrote it from scratch but looked at stock, Zirco, Firefox, and all available code. It uses WebView on your device so HTML5 depends on how much capability Google WebView has on your device.
Almost every Android web browser including mainstream ones are spyware. This is why I made Naked Browser. Here is the XDA thread for it if you want to talk more about it: http://forum.xda-developers.com/showthread.php?t=1929663
Next to stock the only browser comes close to it for smoothness is Mozilla. But wish it had an exit option and full screen
Tried them all. Naked is alright, but the UI is awkward!
Sent from my GT-N7100 using Tapatalk 2
jujuburi said:
Next to stock the only browser comes close to it for smoothness is Mozilla. But wish it had an exit option and full screen
Tried them all. Naked is alright, but the UI is awkward!
Sent from my GT-N7100 using Tapatalk 2
Click to expand...
Click to collapse
I'm testing now Beta versions of Mozilla, Opera, Chrome and Dolphin. Results will follow maybe tomorrow but it's apparent the differences are sometimes "game changing"
mat9v said:
I'm testing now Beta versions of Mozilla, Opera, Chrome and Dolphin. Results will follow maybe tomorrow but it's apparent the differences are sometimes "game changing"
Click to expand...
Click to collapse
Don't forget to enable chrome://flags features in chrome beta browser.
Sent from my GT-N7100 using Tapatalk 2
Manjunath324 said:
Don't forget to enable chrome://flags features in chrome beta browser.
Sent from my GT-N7100 using Tapatalk 2
Click to expand...
Click to collapse
Which futures should I enable? I never been in that happy chrommy place
OK, here be the Beta versions of Main Android Browsers benchmarks:
...........XXXXX..............Chrome standard(+%).....Chrome beta (+%)....Chrome beta exp. (+%)........Opera beta (+%).......Dolphin beta.(+%) .Firefox beta (+%)
HTML 5 Test.................410 (94%).....................429 (99%).................444 (102%).........................428 (99%)..................482 (111%) ...........426 (98%) ..........more is better
HTML5 Benchmark.......1911 .............................2462.........................2723....................................2248..........................3170.......................2674...................more is better
Octane..........................2900 (120%).................3040 (127%).............3064 (128%).......................2923 (122%)..............2304 (96%) ...........2893 (120%).......more is better
Browsemark 2.0............2427 (96%)...................2347 (93%)...............2694 (107%).......................2934 (116%)..............2659 (105%)..........2623 (104%).......more is better
Peacekeeper ...............891 (102%)...................881 (101%)...............872 (100%)..........................918 (105%)................Failed....................636 (73%)...........more is better
Sunspider.....................812 (78%).....................805 (77%).................780 (75%)............................790 (76%)..................935 (90%).............820 (79%)............less is better
V8 Benchmark..............310 1 (115%)................3119 (116%).............3255 (121%)........................3219 (120%)..............2734 (102%).........3114 (116%)........more is better
Dromaeo.......................179.92 (107.6%)...........179.81 (107.6%)......181.02 (108.3%)...................184.4 (110.3%)..........151.31 (90%)........142.92 (85%).......more is better
Some thoughts are neccessary here:
1. I am no fan of Firefox, but in both Browsermark and Peacekeeper benchmarks it was the only benchmark to actually display most of the tests correctly as in playing videos, displaying all 3D content or correctly resizing webpages during tests.
2. There is sometimes large difference in speed when requesting desktop versions of webpage ns mobile (even if on the first glance they look the same) and it can affect benchmark results. For example HTML5 benchmark on some browsers takes 1/2 screen, on others is resized to fit and then on others don't even fit the screen. I tried to keep things equal by manually resizing the "game window" to fit the screen but I do not know the impact of manual resize on the score. The other thing is benchmarks in portrait and landscape mode again differ, sometimes by 5-10%, as such all results are from benchmarks run in portrait mode.
3. Dromaeo benchmark is "made by Mozilla" and curiously it shows the worst scores on Firefox (both mobile and desktop versions as tested on my laptop PC). Results are displayed in "runs/s" so it should read that "the higher the better", my PC scores over 450 in Firefox and over 600 in Chrome so it should be ok. Still, why does Firefox score so low, when in other Java Script benchmarks it is one of the fastest, I have no idea.
To compare browsers I will count Octane, Browsermark, Peacekeeper, Sunspider, V8 and Dromaeo benchmarks.
Chrome is on average 10,43% faster than Stock Browser
Chrome beta is on average 11.27% faster than Stock Browser
Chrome beta with experimental options turned on is on average 14,88% faster than Stock Browser
Opera beta is on average 16.21% faster than Stock Browser
Dolphin beta is on average 0.5% faster than Stock Browser
Firefox beta is on average 3.17% faster than Stock Browser
Personally on my Note 2 I would leave 3 browsers - Stock (smooooth), Opera (faaaast) and Firefox (compatible!!), although I would miss great sync options offered by Chrome. Naked browser, while a great effort, has to be worked on some more to win, either of compatibility, interface or speed, because for now the only thing that speaks for it is that it does not spy on us (which I can not test anyway )
And the winner is........... still Opera but by much smaller margin
Excellent, thanks for testing them out for us.
Just got my T-mobile Note 4 today and installed Chrome right away.
I noticed instantly how chrome felt sluggish compared to my Nexus 5. I ran a few benchmarks (https://www.webkit.org/perf/sunspider/sunspider.html and http://html5-benchmark.com/ ) And chrome came in much slower than the stock 'internet' app. For example on the JS benchmark I was getting ~1.6 seconds for Chrome, 300 ms on 'internet' app. This doesn't sound right at all!
Any idea of what's going on? any settings I could tweak?
Same with the note 3, that's why I use Firefox
Common problem. This should fix it.
Alik4041 said:
Common problem. This should fix it.
Click to expand...
Click to collapse
YOU'D think with all the bloatware sam sung puts on our devices that this would be one known to the massess, or why doesn't Google bake this into chrome so we don't have to even go through this
Sent from my SM-N910T using Tapatalk
What? Never heard of this before? What does it do? It doesn't really ha acne a description. But thanks!
Can't believe I've never heard of this or seen it before! Really makes Chrome much faster!
Actual benchmarks
I tried to install the 'Chrome Samsung Support Library' but it did nothing to improve Chrome lower benchmarks vs the stock browser on my Note 4. I actually benchmarked firefox, chrome and stock on 4 different benchmarks, here are the results:
Code:
[B]peacekeeper (higher is better)[/B]
stock: 1310
chrome: 900
chrome Samsung support: 957
firefox: 648
[B]octane JS ( higher is better )[/B]
stock: 5712
chrome: 5470
chrome Samsung support: 5306
firefox: 3429
[B]sunspider JS ( lower is better )[/B]
firefox: 1094ms
chrome: 884ms
chrome Samsung support: 834ms
stock: 432ms
[B]kraken JS[/B]
stock: 5348 ms
chrome: 5841 ms
chrome Samsung support: 5466ms
firefox: 11658 ms
Stock is always faster, even up to 44% faster on the peacekeeper benchmark. This is disappointing. I'm not sure if people have noticed, but the stock browser is actually Chromium based, just type 'chrome://version' on the stock and you'll see all the info. Therefore there is nothing particularly different to the Chrome available on the app store, so stock and chrome must share a huge amount of codebase. The difference may lie on startup command line settings and some tweak here or there that probably Samsung knows very well and hasn't bothered to make sure chrome runs well.
This is great. Thanks for those benches.
Personally I hate Chrome.....the stock browser is awesome and much lighter and snappier.
:good:
bartolo5 said:
I tried to install the 'Chrome Samsung Support Library' but it did nothing to improve Chrome lower benchmarks vs the stock browser on my Note 4. I actually benchmarked firefox, chrome and stock on 4 different benchmarks, here are the results:
Code:
[B]peacekeeper (higher is better)[/B]
stock: 1310
chrome: 900
chrome Samsung support: 957
firefox: 648
[B]octane JS ( higher is better )[/B]
stock: 5712
chrome: 5470
chrome Samsung support: 5306
firefox: 3429
[B]sunspider JS ( lower is better )[/B]
firefox: 1094ms
chrome: 884ms
chrome Samsung support: 834ms
stock: 432ms
[B]kraken JS[/B]
stock: 5348 ms
chrome: 5841 ms
chrome Samsung support: 5466ms
firefox: 11658 ms
Stock is always faster, even up to 44% faster on the peacekeeper benchmark. This is disappointing. I'm not sure if people have noticed, but the stock browser is actually Chromium based, just type 'chrome://version' on the stock and you'll see all the info. Therefore there is nothing particularly different to the Chrome available on the app store, so stock and chrome must share a huge amount of codebase. The difference may lie on startup command line settings and some tweak here or there that probably Samsung knows very well and hasn't bothered to make sure chrome runs well.
Click to expand...
Click to collapse
Even with the support library installed, I think it is slow. Coming from a Note 2 where I used Chrome as my default browser. Never tried any benchmarks, but it just seems way way slower than it ever did on the Note 2.