What are the more reliable, and accurate apps for benchmark, stressing, and monitoring?
Sent from my SGH-T989 using Tapatalk 2
There are no benchmarks that accurately determine what the user experience will be like.
le rustle face
Keion said:
There are no benchmarks that accurately determine what the user experience will be like.
le rustle face
Click to expand...
Click to collapse
Interesting, I base user experience on function of software in conjunction with hardware, or compared to original factory software. The only accurate way to determine this is by stressing hardware to it failure point, each time a software component is changed/new code implements are added
So the question still remains, what do the people at xda use to benchmark, stress and monitor Android devices
Sent from my SGH-T989 using Tapatalk 2
Androbench for I/O (or RL Benchmark for SQLite performance and SDCardTester or similar app for R/W speeds), Linpack for CPU, AnTuTu and CF-Bench for CPU/GPU/general performance (some people use Quadrant for GPU/CPU too, but it's been said not to be relevant anymore).
Related
This is the final part in a 5 part series of Android performance benchmarking apps. Quadrant is probably the most popular Android performance benchmark app available today. First made popular in videos showing the (then) upcoming Galaxy S, it's been making the rounds as people jump on board to test their Android phones against the big guns. Quadrant is an all-around, full-featured performance test that measures all aspects of your phone's performance for a full rundown of how powerful it truly is. Some of the things analyzed are the CPU, 3D performance, and Memory, among others. Check out the video for a full demo of the app. Be sure to download today for a great app to compare Android phones, compare ROMs, or just have fun using one of the hottest apps on Android. Then go brag to all your iPhone friends. Do it.
Let me know in the comments which app is your favorite, or if you have any other great benchmarking apps!
Part 5: Quadrant for All-Around Performance Testing
For Part 4, check out Fps2D for 2D FPS Benchmarking
For Part 3, check out Linpack, the Most Popular CPU Performance Test
For Part 2, check out Necore for 3D Graphics Performance
For Part 1, check out Benchmark Pi for CPU Performance Benchmarking
I always liked Linpack.
Wouldn't it be nice to have all of this in one application, with similar framework and standardized test methods?
And would it be nice if you were able to extract ALL of the information out of your phone regarding Build, CPU features, Video subsystem, network subsystem, etc?
I wonder what such an app should be called? (seriously, I can't decide on a name. PhoneAnalyzer just seems... blah)
Dont forget GLBenchmark.
I dont think there is one for the Evo but I used the Incredible one.
http://glbenchmark.com/
Doward said:
Wouldn't it be nice to have all of this in one application, with similar framework and standardized test methods?
And would it be nice if you were able to extract ALL of the information out of your phone regarding Build, CPU features, Video subsystem, network subsystem, etc?
I wonder what such an app should be called? (seriously, I can't decide on a name. PhoneAnalyzer just seems... blah)
Click to expand...
Click to collapse
Definitely would be. Quadrant gets close. I think a combo of Quadrant, Phonalyzr, and SystemApp would be an app not to be reckoned with
Has anyone here used the Antutu Benchmark app. I was comparing the scores between my Desire S and a Desire HD. My DS beat DHD in all aspects but Databse IO performance.
Can anyone explain what the Database IO is and how I can improve the performance?
Ps. My database IO score was very low between 20-40 whereas the DHD had over 200. I also seen other people with the same handset as me with scores over 150, is it just my handset?
Sent from my HTC Desire S using XDA App
tkr1992 said:
Has anyone here used the Antutu Benchmark app.
Click to expand...
Click to collapse
See results of Antutu on several Desire S and my Desire:
http://forum.xda-developers.com/showthread.php?t=1054393
tkr1992 said:
Can anyone explain what the Database IO is and how I can improve the performance?
Click to expand...
Click to collapse
I guess it is benchmark of queries and updates in sqlite databases (that are used quite often for persistent data storage of small data items).
i am running infused 1.3.1 andon quadrant i would score 1500-1750 but on smartbench i would get 915-1015 on cpu and pretty close to 3000 on gpu , i have two questions, are those score from smart bench good and secondly which should i take more seriously quadrant or smart bench
silver258 said:
i am running infused 1.3.1 andon quadrant i would score 1500-1750 but on smartbench i would get 915-1015 on cpu and pretty close to 3000 on gpu , i have two questions, are those score from smart bench good and secondly which should i take more seriously quadrant or smart bench
Click to expand...
Click to collapse
None of it matters... The phone runs good doesn't it? that is all that should matter
Sent from my SAMSUNG-SGH-I997 using XDA App
Wtf good is an insult its beyond good i came from iphone 3gs lol
Sent from my SAMSUNG-SGH-I997 using XDA App
if you buy the full version of quadrant you get more information, i am totally against the idea of just stacking scores like quadrant does. that being said smart benchmark attempts to make a simple compsite score that is more relevant than the otheres do using some kind of algorithm to balance the weight of each componant, i dont agree with that either.
if you want to benchmark to figure out what you have improved specifically then get antutu or pay for quadrant but never look at the cumulative scores, it is totally irrelevant.
if you want to see how the performance compare acording to an algorithm get smart benchmark
if you want to be impressed with the phone, run a 3d/opengl benchmark. or otherwise play some H.264 codec .mkv videos at 720p (1080p also works but at a lower bitrate and class, the 720 video would likely look better) then try the same video on an atrix and an inspire.
Well...
I'm currently running
Faux 0.12 OC'd to max on 'ondemand'
Android Revolution HD 3.04
and I only get 2500-2600 on Quadernt benchmarks
I've heard of people that get 3000+ D:
How do I increase performance???
Quadrant is ****. That's all
¤Violets are red,
Roses are blue,
Wait, what?
I have no idea what I'm saying,
But you lost the game¤
Quadrant is also inconsistent. It will piss you off to no end, my friend. Try CF-Bench. I use that to see whether or not any "improvements" are such.
Matt
Get a crayon and draw a line.
Get a pencil and write any number you like under it.
As good as quadrant, any day.
Do you want a fast phone, or do you want some imaginary numbers?
Use cf bench and linpack to judge if changes made your phone better or worse.
As far as comparing numbers from one phone to another, forget it.
P.
I've heard linpack is the worst app ever from a lot of people so I don't trust that, at all.
Sent from my HTC Sensation Z710e using xda premium
I've heard a lot about CF-Bench but never got around to using it. I definitely know not to use Quadrant but have always heard good things about Antutu (which I normally use). How does CF-Bench compare with Antutu for performance measuring accuracy?
I'm not usually a benchmark kinda dude, bt antutu can suck some pistachios, I always get inconsistent results with it, bt if anything ll run cf when I use a diff kernel just to c if any improvements
Sent by my feau gee
Doesnt clock both processors. Use cf bench please. Thats why you are getting low scores
Sent from my HTC Sensation Z710e
yep, agree on CF bench, but if you really like higher quadrant points, you should try 2 things:
1. run quadrant a minimum of 5 times in a row, will increase your points on the last runs
2. put in class 10 sd card, will increase points a lot in comparison to class 4 stock card ..
Agreed on running CF Bench, Antu tu isn't bad either.
hlenforcer said:
yep, agree on CF bench, but if you really like higher quadrant points, you should try 2 things:
1. run quadrant a minimum of 5 times in a row, will increase your points on the last runs
2. put in class 10 sd card, will increase points a lot in comparison to class 4 stock card ..
Click to expand...
Click to collapse
Whats the difference in sd class?
Some benches test r/w speeds of SD card and they affect bench score.
Alfhea said:
Whats the difference in sd class?
Click to expand...
Click to collapse
Sent from my HTC Sensation Z710e using XDA App
I found that while browsing news so I though I'd share it here.. I'm disappointed because Samsung actually cheating us. :/
--
Smartphone benchmarks never really indicate how a device will ultimately perform, yet tech enthusiasts exalt results like they’re the end all, be all. Check any review; the nebulous collection of numbers are always held in high regard despite them never really doing much to affect the overall experience. Still, companies take them quite seriously—in Samsung’s case, a little too seriously.
According to a new report from AnandTech, Samsung might be fibbing its way to more favorable Galaxy S4 benchmarks. Has your device suddenly come to a crawl? Of course it hasn’t; benchmarks shouldn’t change your perception of a flagship as powerful as the S4. Still, it’s embarrassing that Samsung would resort to such technical tactics, like allegedly using code dubbed “BenchmarkBooster.” Yes, your device takes steroids.
AnandTech found that Samsung set the GPU of the Exynos 5 Galaxy S4 to run higher when benchmarked—higher than normal everyday use. When engineers tested the device, the S4’s Exynos 5 ran at 533MHz during benchmarking, and only ran at 480MHz during regular use. Not an enormous difference, but large enough to call shenanigans.
In addition, AnandTech found that when running CPU benchmarks with apps such as AnTuTu and Quadrant, the device’s Cortex A15 clocked at 1.2GHz; an unofficial benchmarking app, GFXBench 2, revealed that the device actually runs at 500MHz when it’s not juicing. Seems fishy, no?
AnandTech’s findings should in no way effect your final opinion on the Galaxy S4, though it does highlight some shady Samsung tactics. It’s likely the Korean company isn’t the only one to fib benchmarking tests, though; the company is just the one that got caught
Source : http://www.technobuffalo.com/2013/07/30/samsung-caught-manipulating-galaxy-s4-benchmark-results/
Gsmarena clearly stated that Samsung is Cheater?
They played the game with the emotion of common consumer...and ...now i am thoroughly dissapointed by Sammy's behaviour
The manipulation of Benchmarks for me confirms the rumors about the Octacore:
SamMobile News from 30 May 2013
Regards
Tanis64 said:
The manipulation of Benchmarks for me confirms the rumors about the Octacore:
SamMobile News from 30 May 2013
Regards
Click to expand...
Click to collapse
So they released an unfinished chip into the market. Only reason could be marketing. The look at us we have eight core phones angle. People go wow eight is more than four and buy it. So lame.
-- Sent from the mighty Note 2 --
Already mentioned by @AndreiLux few decades ago.
Sent from my iPotato
LegendJo said:
I found that while browsing news so I though I'd share it here.. I'm disappointed because Samsung actually cheating us. :/
Click to expand...
Click to collapse
So is Intel cheating you to when the Intel Core i5-2500k CPU have anormal speed of 3.3 GHz, but have a turbo speed of 3.7 GHz?
Tom-Helge said:
So is Intel cheating you to when the Intel Core i5-2500k CPU have anormal speed of 3.3 GHz, but have a turbo speed of 3.7 GHz?
Click to expand...
Click to collapse
if that speed is only available if your running prime95 or intel burn then....yep.
You can't compare the two, they are totally different.
Tom-Helge said:
So is Intel cheating you to when the Intel Core i5-2500k CPU have anormal speed of 3.3 GHz, but have a turbo speed of 3.7 GHz?
Click to expand...
Click to collapse
No, because Intel didn't mislead anyone. Samsung clearly did.
yeahmann said:
No, because Intel didn't mislead anyone. Samsung clearly did.
Click to expand...
Click to collapse
Really??
Have you not read this research paper?
http://www.abiresearch.com/press/intel-apps-processor-outperforms-nvidia-qualcomm-s
A modified version of AnTuTu was used to cheat in benchmark.
http://www.eetimes.com/author.asp?section_id=36&doc_id=1318857
http://forums.anandtech.com/showthread.php?t=2330027
CLARiiON said:
Really??
Have you not read this research paper?
http://www.abiresearch.com/press/intel-apps-processor-outperforms-nvidia-qualcomm-s
A modified version of AnTuTu was used to cheat in benchmark.
http://www.eetimes.com/author.asp?section_id=36&doc_id=1318857
http://forums.anandtech.com/showthread.php?t=2330027
Click to expand...
Click to collapse
That paper had nothing to do with what he's saying. Turbo is still valid and not misleading.
Secondly, you can blame AnTuTu, not Intel, for that benchmark discrepancy. And then again it was only due to compiler change and no actual cheating per se in the conventional matter.
CLARiiON said:
Really??
Have you not read this research paper?
http://www.abiresearch.com/press/intel-apps-processor-outperforms-nvidia-qualcomm-s
A modified version of AnTuTu was used to cheat in benchmark.
http://www.eetimes.com/author.asp?section_id=36&doc_id=1318857
http://forums.anandtech.com/showthread.php?t=2330027
Click to expand...
Click to collapse
From Exophase:
But frankly, I blame AnTuTu in all of this. They allowed themselves to be manipulated (probably for a price), despite constantly warning against other people cheating their numbers. I don't know if they're displaying a complete lack of integrity or a complete lack of understanding of how their own software works, or something in between the two, but whatever the case I hope they lose all credibility and whatever revenue the program brings them.
Click to expand...
Click to collapse