This one will be of more interest to kernel developers.
Link to Spreadsheet: Results
RcrdBrt asked me to see which memory allocator performed best, and provided me with four identical kernels except for the allocators being SLQB, SLAB, SLUB, and SLOB.
SLUB turns out to have a slight performance advantage, as RcrdBrt had suspected.
The rank was:
SLUB - 980.56
SLQB - 972.17
SLAB - 967.72
SLOB - 967.72
Not much difference really, but every little helps, as they say at Tesco (Britain's leading super market/Religion).
Thanks to RcrdBrt for trusting me with this nice little study, and to Chainfire for the very useful benchmarking app. (See here)
Methodology
Technical detail (it's not necessary to read this!)
I used CF-Bench to obtain values for:
Native MALLOCS
Native Mem Read
Native Mem Write
Java Mem Read
Java Mem Write
...and used statistical methods to boil 10 passes for each allocator down to one final score: 10 passes is enough to get the mean and median within one percent of each other, i.e. establishes a confidence in the mean, assuming normally distributed data. Three standard deviations are subtracted from the mean, to show the minimum score we'd expect 99.7% of results to be above. This is to penalize variability (high peaks are less significant than a good consistency). The geometric mean is taken of the five end results to provide one final score.
Bedalus can't live without benchmarks. Closed UX benchmark yesterday and opened a new one today
One more instructive and helpful benchmark buddy
Most kernel developers prefer using SLQB. Any thoughts on why this is the case? Do they have a special reason for that or just a matter of not knowing better? Cheers!
RcrdBrt said it was the fashion! Tbh i don't know the differences between them, but some casual googling showed that slub has had success in other Linux arenas
apatal said:
Most kernel developers prefer using SLQB. Any thoughts on why this is the case? Do they have a special reason for that or just a matter of not knowing better? Cheers!
Click to expand...
Click to collapse
Buzzwordism. The other 3 are in kernel.
Sent from my Nexus Prime.
morfic said:
Buzzwordism. The other 3 are in kernel.
Click to expand...
Click to collapse
You mean all those allocators are present but they mention only SLQB because it's the trend?
Sent from my Nexus S
apatal said:
You mean all those allocators are present but they mention only SLQB because it's the trend?
Sent from my Nexus S
Click to expand...
Click to collapse
Correct.
Sent from my Nexus Prime.
Man, bedalus is really fast!
@morfic:
So, in the ICS stock 3.0.8 kernel, how many and which allocators are present?
Am I correct to say that all those four allocators are present, but only one is used?
Sorry for these dumb questions.
glennkaonang said:
Man, bedalus is really fast!
@morfic:
So, in the ICS stock 3.0.8 kernel, how many and which allocators are present?
Am I correct to say that all those four allocators are present, but only one is used?
Click to expand...
Click to collapse
SLAB, SLUB, SLOB are in most (mainline) Linux Kernel, SLQB is patched in.
bedalus, did you run this on GB too?
morfic said:
bedalus, did you run this on GB too?
Click to expand...
Click to collapse
I can do it if i have 4 identical gb kernels (except for the allocator) ... If you'd like it testing you'll have to build them for me?
Related
Results in Google Docs Spreadsheet
All ICS ROMs Benchmarked
Any regular visitors to this thread can't have failed to notice the big changes in the spreadsheet. In particular I've been studying the effect of the kernel the ROM developer includes. If you look at column H, you will see whether I recommend flashing another kernel to improve performance at stock speeds.
REMEMBER: Performance is not everything! Many custom kernels provide extra features!
If you do decide to switch kernels, and only desire stock speeds and a reduced feature set, I recommend two kernels in this post.
If you are new to this stuff read this.
SUMMARY OF RESULTS
#1st place: Pixel ROM 1.72
#2nd place: CyberGR-MOD|NS.NGN ICS v10 HYBRID
#3rd place: MIUI Catalyst 2.4.6
#4th place: AOSP+ N.5
#5th place: [MIUI][2.4.6] Oodie pRiMe rOcket fUeled
Where do I find all these wonderful ROMs?
Their XDA threads are all linked in the table, so click through to discover each ROM's feature set, included apps, theme, and evidence of good karma!
Looking for a Gingerbread ROM? Try this.
Where did all your other benchmarks go?
Kernel Features & Benchmarks: this thread
Battery drain: this thread
CPU Governors and I/O Schedulers: this thread
Power Saving Governors: this thread
Thanks to all the developers.
Testing Methodology
Each ROM is clean installed, allowed its own default settings, then all benchmark software is installed and run once the system is settled. I also perform one screen off screen on, just out of habbit (this used to resolve a graphics bug in gingerbread, which I'm not even sure exists in ICS). All results are recorded in the spreadsheet. Then I restart in recovery, wipe cache and dalvic (thanks morfic for dalvic wiper) and install morfic's T132-I kernel. Reboot, use nstools to select performance (default I/O scheduler for morfic's kernel is deadline). Reperform all benchmarks, recording them in the spreadsheet. The calculation for the overall CPU RAM I/O and GRAPHICS scores is identical to the one I use in my kernel benchmarking spreadsheet. Please refer to that thread for more detail. However, there is one difference. Quadrant behaves very differently in different ROMs depending on what graphics tweaks they incorporate. As such, no Quadrant scores are used in this ROM benchmarking study. For more details, see under 'Graphics' below.
The final step was to average out the results of both kernel scores, and ranking on that basis. Some would say that they are best ranked by looking at T132 scores only. I would say that is a fair point, and definitely worth consideration.
There are two reasons I incorporate the default kernel score:
1) Overall scores have a worst case scenario accuracy of within 2% of the true mean due to variability in the benchmarks (more detail on how I obtain this figure is available in my kernel benchmarking thread). For normally distributed data (which we can safely assume these benchmarks produce) averaging two sets of results will result in doubling the accuracy to within 1% of the true mean, making the ranking more reliable.
2) Combining the default kernel Scores for ROMs that used a poor performing kernel, or selected bad governor/scheduler combinations by default will suffer as a consequence. These problems can be solved if the user flashes their own preferred kernel and/or adjusts the kernel settings. However, many new users will not perform this step. For that reason, I want to give a score that at least somewhat reflects the performance as the developer intended.
-Graphics
One of the early findings is selecting force GPU rendering in the developer options improves 2D performance, in Quadrant only. Also, and only in Quadrant, 3D performance can be improved by deleting or moving /system/lib/egl/libGLES_android.so using root explorer or equivalent. These tweaks/hacks do not seem to affect other benchmarks in the slightest.
Because of the bizarre effects of these tweaks on Quadrant scores, I have removed them from the formula that calculates the overall GFX score, and this in turn impacts the overall average score. In short, it makes comparing each ROM's score fairer.
You can see evidence of how the graphics tweaks make Quadrant behave differently at the bottom of the spreadsheet, highlighted in light red. The host ROM for this particular comparison was DianXin or DX ROM for short. This is a reference to a post I make in DX ROMs thread, where I first decide that removing Quadrant is the best option: http://forum.xda-developers.com/showpost.php?p=22323688&postcount=194
-Android 4.0.1
These ROMs (at the bottom of the spreadsheet, highlighted in light red) were based on early maguro ports provided by koush. I have left them in, because although they are difficult to compare to 4.0.3 builds with the correct drivers and whatnot, they do compare to each other in one important respect: the kernel. One is stock, the other is built by eugene373 (galaxy nexus/nexus s dev.) Using eugene373's kernel, CPU, I/O, and Graphics were all raised to a high standard that is on a par, or above some of the current ROMs. This goes to show that a customised kernel can offer great benefits, even on a ported ROM.
-Android 4.0.4
As of the 5th of February, my table contains all the ICS ROMs that are listed in the XDA forums. There was only one 4.0.4 version at the time, and this was the best performing in the benchmarks. When the source code is released, I'm sure many ROMs will update. At this point they will benefit from increased RAM scores and javascript performance. If all the ROMs scores improve by the same proportion, the rankings will not be affected. I will attempt to included major releases, but I will not reperform these tests every time there is a new nightly.
Reserved 4
Reserved 3
Thanks
I was waiting for this
It might lead to troll wars though
By the way, i dont see Pete ICS rom in there, im using it.
Ibn Saeed said:
Thanks
I was waiting for this
It might lead to troll wars though
By the way, i dont see Pete ICS rom in there, im using it.
Click to expand...
Click to collapse
Thanks, and no troll wars please y'all.
Don't you think "popularity" is going a little too far? It will definitely lead to fights...all the devs are good and I don't think prominence is a good indicator of a good rom
But it's up to you
Sent from my Nexus S using XDA App
Personally, benchmarks are very misleading... Some of the most stable and well developed roms don't get the greatest scores because they weren't made with benchmarks in mind.
Sent from my Nexus S 4G using xda premium
Another great work! Thanks
Sent from my Nexus S using XDA App
mathkid95 said:
Don't you think "popularity" is going a little too far? It will definitely lead to fights...all the devs are good and I don't think prominence is a good indicator of a good rom
But it's up to you
Sent from my Nexus S using XDA App
Click to expand...
Click to collapse
just remember, popular doesnt mean better. but to get angry at the truth(whatever it might be) is just silly. its lies that you should be upset with.
Where's Peter Alfonso's ROM? Bugless Beast is a mainstream rom, second only to Cyanogenmod in scope and breadth.. The ROM thread is in the General forum, because he links to his page at Rootzwiki, but it's a more prominent ROM than many you've listed.
Please include it in the benchmarking as well.
Sent from my Nexus S using Tapatalk
http://forum.xda-developers.com/showthread.php?t=1430170
you didn't add and test this one
please add pete ics bugless beast.
Omissions fixed!
UPDATE: kwiboo's project removed, was an SDK port. So so tired now, got to sleep. Tomorrow begin again with Brainmaster's tweaked ICS
Which kernel is used to make the tests?
-----
Forget It...
"I'll be benchmarking them twice with whatever kernel they come with, and then with morfic's bare boned T132"
biotecsoul said:
Which kernel is used to make the tests?
Click to expand...
Click to collapse
Read the op.
Sent from my Nexus S 4G using xda premium
u could also add overnite battery drainage column with sync on or somethin like that, if not overnite then a period of 2 hrs idle .. how much does it drain ..
Holy **** man, what would it be if the NS comunity didnt have you??
Thumbs up!!!!
Sent from my Nexus S using xda premium
thegtfusion said:
Holy **** man, what would it be if the NS comunity didnt have you??
Thumbs up!!!!
Sent from my Nexus S using xda premium
Click to expand...
Click to collapse
Haha, that makes me feel good. Wait, Benchmarks > XDA thread > Positive XDA user feedback > Dopamine release = Addiction to benchmarking? Time to join crackflashers anonymous...
mathkid95 said:
Don't you think "popularity" is going a little too far? It will definitely lead to fights...all the devs are good and I don't think prominence is a good indicator of a good rom
But it's up to you
Sent from my Nexus S using XDA App
Click to expand...
Click to collapse
I doubt it would lead to fights between devs, well hopefully not. I can see users arguing stuff more than devs.
Benchmarks aren't that good anways.
BTW, your kernel sucks!!!!
Just kidding been running smoothly since I started using it the other day
-----------
Thanks for the ratings system. Looks nice!
What the title says.
This question gets asked over and over and over again. My answer....who cares. If you like what you are using, if its smooth and fast who cares right?
Sent from my HTC Vision using xda premium
I don't feel the need to explain how reliable (if at all) benchmarks are when it comes to measuring how fast a ROM performs.
So, to answer your question, my highest benchmarks were on ILWT and EliteMod, with both around 3,6-3,7K points on Antutu.
Spastic909 said:
This question gets asked over and over and over again. My answer....who cares. If you like what you are using, if its smooth and fast who cares right?
Sent from my HTC Vision using xda premium
Click to expand...
Click to collapse
Apparently you don't care, but I do.
crestofawave said:
I don't feel the need to explain how reliable (if at all) benchmarks are when it comes to measuring how fast a ROM performs.
So, to answer your question, my highest benchmarks were on ILWT and EliteMod, with both around 3,6-3,7K points on Antutu.
Click to expand...
Click to collapse
I already have ILWT. What's EliteMod? Kernel?
EliteMod is a CM7 Variant and uses it's own "EliteKernel", much likes Tsubus' ILWT one. Both ROM's performances are similar and you really can't go wrong with either of them, if you want as much speed as possible.
Noted differences between EliteKernel and ILWT kernel is that EliteKernel has smartassV2 / scary governors (+ others I can't remember) and can also overclock higher than ILWT. Other than that, they don't have much differences (that you are actually going to notice in terms of speed, etc).
convolution said:
What the title says.
Click to expand...
Click to collapse
I scored 4832 in Antutu on one of the first ICS ROMs by xTazx, if I remember correctly it had a kernel by Flinny. It had insanely high 3D scores, close to 130fps in the thing with all the lights.
Rizo Zizo benchmarked many of the major ROMs in this thread: http://forum.xda-developers.com/showthread.php?t=1376496
You'll see that a couple Sense 3 ROMs scored almost the same as ILWT, adding credence to the notion that benchmarks don't correlate to real world performance.
redpoint73 said:
Rizo Zizo benchmarked many of the major ROMs in this thread: http://forum.xda-developers.com/showthread.php?t=1376496
You'll see that a couple Sense 3 ROMs scored almost the same as ILWT, adding credence to the notion that benchmarks don't correlate to real world performance.
Click to expand...
Click to collapse
+1000 for having a clue
convolution said:
What the title says.
Click to expand...
Click to collapse
antutu-5059
ILWT CM7
overclock CPU at 1,574GHz
redpoint73 said:
Rizo Zizo benchmarked many of the major ROMs in this thread: http://forum.xda-developers.com/showthread.php?t=1376496
You'll see that a couple Sense 3 ROMs scored almost the same as ILWT, adding credence to the notion that benchmarks don't correlate to real world performance.
Click to expand...
Click to collapse
Guru Zeb said:
+1000 for having a clue
Click to expand...
Click to collapse
I agree entirely. Benchmark scores don't mean anything. People need to understand the idea of a bottleneck, and the fact that in the Desire Z the CPU is not the bottleneck, it's the RAM. I OCed CPU to 1800MHz and GPU to 350MHz (>100MHz OC) and my benchmark scores were slightly higher, well over 5000 in AnTuTu, but guess what, real life performance was no different. Proof: the Desire S and HD have a very similar chipset (msm7x30) and the same C/GPU but guess why they're so much faster and better at multitasking and better to use in general than the Desire Z? Because they have 768MB of RAM and we don't!
Stop worrying about benchmarks, and download a couple of different ROMs, GB, ICS, jb, sense, and make your own mind up about which one you like best.
Over my time here in XDA I've built up a few studies on the Nexus S.
Here are the links. There are summaries in each thread in the first and second (and sometimes third) posts of the main findings, but I've done some very quick and minimal overviews here also.
How to build a kernel: this thread
-This post takes you through installing Ubuntu right through to managing git, in order to compile your own kernel, and flash it in CWM recovery!
-Includes the tools you need to add MTP capabilities to your kernel.
User eXperience Benchmarks: this thread
-This shows which ROM and kernel threads have the best user feedback. The feedback is processed by an external sentiment analysis API. It is only meant as a rough guide, something of a quick reference particularly if you are looking for a new ROM to try. As of June 2012 this is no longer maintained.
All ICS ROM Benchmarks: this thread
-Freely available benchmark programs were used to determine which ROMs had the best performance. Generally the difference between the best ones is very small.
Battery Drain Benchmarks: this thread
-A collection of tips for getting the best battery life, and dispelling some myths.
Kernel Features and Benchmarks: this thread
-My original study, the one that started it all, the kernel benchmarks! Provides a quick table to compare which kernels perform best in CPU/RAM/IO/Database/Graphics.
CPU Governors and I/O Schedulers: this thread
-A study done to find which schedulers perform best in combination with the most popular governors. My two favourites (ondemand and smartassV2) combine best with noop.
Power Saving Governors: this thread
-A meta study done combining the findings from the battery study and governor study to show which governors will save the most power. Ondemand comes out significantly better than conservative, which surprised me. SmartassV2 is the best compromise for high performance and efficiency.
Does SuperCharging work?: this thread
-This was just a short study to find out if this script is suitable for the Nexus S, but it turned out it had no effect, positive or negative.
Kernel Memory Allocators: this thread
-This is another short study for kernel developers to show that SLUB performs best.
====
This thread is just an index. Our new mod veyka agreed would be a good idea to summarise all my studies for addition into the sticky roundup.
====
EDIT: Feel free to suggest any new things that I might be able to test!
im out of thanks for today.. but, awesome work regardless. thank you
simms22 said:
im out of thanks for today.. but, awesome work regardless. thank you
Click to expand...
Click to collapse
Me too, but thank you too!
Edit: ah just use the app, unlimited thanks!
bedalus said:
Me too, but thank you too!
Edit: ah just use the app, unlimited thanks!
Click to expand...
Click to collapse
really? unlimited? i dont use the app though, just the browser.
Uh, well, I haven't tested the limit... but I never get denied!
All your work was great I battery drain thread was very helpful
Sent from my Nexus S
Thanks for all your hard work.
IF , one of my assassins had done their job ..........
.. Generated via Tapatalk ..
Oodie said:
IF , one of my assassins had done their job ..........
.. Generated via Tapatalk ..
Click to expand...
Click to collapse
When they finally come for me, I will benchmark their performance.
bedalus said:
When they finally come for me, I will benchmark their performance.
Click to expand...
Click to collapse
nice answer
And nice guide for your very instructive benchmarks bedalus
Even when browsing through the app you'd still get that annoying thanks limit.
This is a really good overview for new NS users. (I doubt there are ever some since mostly would go directly for G-Nex)
What's next?
That "Oodie's Assassins' Performance Benchmark" sounds interesting, though.
PS: @Oodie, you should have kanged your assassins before sending them to bedalus' compound. That way, your assassins' would have at least injured bedalus
UPDATE: Memory Allocators studied. SLUB wins!
Great reference for a refresher or for a new comer. Really like the battery drain benchmarks and how you discuss some myths. Governors is great...love Smartass V2 with Matrix V20!!
Thanks @bedalus!! Totally deserves a sticky.
Sent from my Nexus S using xda premium
Thanks. It's in the sticky roundup!
bedalus said:
Thanks. It's in the sticky roundup!
Click to expand...
Click to collapse
Can we vote for the sticky or it is just for the moderators?
There was like ten stickies, I think veyka wants to keep everything in a list in one sticky.
I appreciate the sentiment though!
bedalus said:
There was like ten stickies, I think veyka wants to keep everything in a list in one sticky.
I appreciate the sentiment though!
Click to expand...
Click to collapse
thank you for all your work, love your kernels ... currently been using deadline/ondemand but since you say it is not best I will try smartassv2+noop.. (I've found smartass with other combinations give me a lot of lag)
BLaZeDRas said:
thank you for all your work, love your kernels ... currently been using deadline/ondemand but since you say it is not best I will try smartassv2+noop.. (I've found smartass with other combinations give me a lot of lag)
Click to expand...
Click to collapse
Ondemand for good balance between performance and battery saving. Smartassv2 for gaming. CFQ and deadline are the best governors, but deadline probably has the edge.
Thanks! Now this widens my perspective on choosing my own Governor and Scheduler Combo.
Here is the best kernel for stock based x10 roms.
n5-1 http://www.mediafire.com/?z8voe760ok24doy
ALL THANKS TO THJAP
You don't know what you've been missing! This is for locked bootloaders so flash it with xrecovery. For unlocked bootloaders flash baseband .71 with flashtool first which also gives you the stock kernel. Then you can flash this kernel through recovery.
Sent from my X10i using xda premium
Features?
Farik335 said:
Features?
Click to expand...
Click to collapse
Im not even sure whats different about this one over others. All I know is its way better than any other I've tryed.
Sent from my X10i using xda premium
There is no such thing as a "Best kernel"! it grealty depends on what rom you're using, which settings you apply, ram, number of apps, cpu frequencies, governors and i/o scheduler. It also greatly differs per phone!!!!!
This one doesnt work on jaf rom!
It makes my phone to stuck at bootscreen!
And after flashing back to dooms kernel,everything is lagging,so i had to restore a backup!really a great kernel!
After flashing this My phone speed is like flashing the new Adreno ICS Driver its really slow
OK OK OK
I know benchmarks don't mean anything to you guys.
Sent from my X10i using xda premium
Deleted!!!
JeffreyPiket said:
Read this and weep!!!!!
Click to expand...
Click to collapse
Man. I thought that would spark a reaction.
Sent from my X10i using xda premium
tonyreimer said:
OK OK OK
I know benchmarks don't mean anything to you guys.
Sent from my X10i using xda premium
Click to expand...
Click to collapse
Read it and weep!!!!!
Dani897 said:
benchmarks are basically useless to the end user. fun to compare but dont think about it too much. a benchmark like quadrant can show a low score based on one area of poor performance but there is no guarantee that that will cause a bottleneck for data. it can also generate a high score based on one area and there is no proof that it is benificial. for instance we use voodoo lagfix that increases filesystem io according to the method quadrant usses to measure it by a fairly significant amount. it has never been proven to reduce hesitation of the phone outside of perception of the user, it is a good mod and i dont knock it im just trying to be objective. the thing is that data goes throught many componants and is held in ram and cached on the cpu, as long as the useage is light the file system performance is barely an issue because there are several systems that act as buffers, only when processing long sets of data uninterupted will the certain speeds become a bottleneck.
for a good analogy it is like drinking through a straw vs drinking through a hose vs drinking through a 1ft diameter pipe. the straw may slow the speed you can drink while the hose will not, but the pipe can supply much more water yet it is of no benifit to the person drinking because he can only drink so fast.
add cacheing into the analogy and it is like filling a glass through a straw a pipe and a hose. the glass can supply the person drinking as fast as they need it, but only holds so much liquid, it is enough in for on person but if several people want to drink (multi tasking or heavy usage) the staw will be a bottle neck, the hose will not and the pipe though overkill may make things marginally faster. now if there is alot of caheing it may be more like filling buckets and then the pipe may help but only when there is an enormous amount of data being used and in that case it needs to be processed just as fast, this is when the cpu can become a bottle neck.
it is very difficult to write a benchmark that can take this into account so it is hard to tell the usefullness of the data the benchmark gives you without reading the results of the individual tests it performs and making judgment calls.
we can increase our benchmark scores for filesystem by changing the filesystem to ext4, and even better by going with nilfs or nilfs2. we can increase sequencial reads by increasing readahead but that reduces random performance and waists resources. it all becomes a big balancing act and the end results can vary.
there are also ways to change the ram timing, the dalvik cache heap size, alter the ram disks, we can make the rom smaller and remove stock apps that are not needed, turn off background services that use resources, change polling frequencies to save resources, change governors to ramp up faster, alter and compile drivers for more efficient use of hardware, and a long list of other things. some work extremely well, some don't. some will make the phone snappier and have no effect on benchmarks, some will greatly increase benchmarks and not the feel of the phone.
Click to expand...
Click to collapse
:highfive:
Check out the 2d and 3d on the bottom of the top screen shot. I challenge you to get that good on both of them with any other kernel on our old x10.
Sent from my X10i using xda premium
tonyreimer said:
Check out the 2d and 3d on the bottom of the top screen shot. I challenge you to get that good on both of them with any other kernel on our old x10.
Sent from my X10i using xda premium
Click to expand...
Click to collapse
No Comment... don't feel offended.. but
Did you read anything from the text Jeffrey posted?
You, well not you, but any good Dev can tweak a kernel and rom, to be good in Benchmarks.. while the general User Experience just plain sucks..
And to be honest, almost every ICS Rom with a somewhat decent optimized Kernel, has more 2D and 3D points than your screenshot.. i don't even need to post proof. it's a fact, even with the Xperia X10.
And there is no such thing as a best kernel...
Depending on the user, every Kernel has it dis-/advantages..
And thats why i think it's good for your ROM to support multiple Kernels, a thing many dev's are still missing, because when you have multiple Kernels, every user can use the Kernel best suited for his needs.
Tekkpriest said:
No Comment... don't feel offended.. but
Did you read anything from the text Jeffrey posted?
You, well not you, but any good Dev can tweak a kernel and rom, to be good in Benchmarks.. while the general User Experience just plain sucks..
And to be honest, almost every ICS Rom with a somewhat decent optimized Kernel, has more 2D and 3D points than your screenshot.. i don't even need to post proof. it's a fact, even with the Xperia X10.
And there is no such thing as a best kernel...
Depending on the user, every Kernel has it dis-/advantages..
And thats why i think it's good for your ROM to support multiple Kernels, a thing many dev's are still missing, because when you have multiple Kernels, every user can use the Kernel best suited for his needs.
Click to expand...
Click to collapse
Ya, I read it. I haven't tryed ics yet cause i use my phone as my internet connection. OK thats fine. I don't really care if not anyone else uses this kernel but as for me its the best one I've ever had. Peace
Sent from my X10i using xda premium
:silly: That's all!
etedeni fhall
JeffreyPiket said:
There is no such thing as a "Best kernel"! it grealty depends on what rom you're using, which settings you apply, ram, number of apps, cpu frequencies, governors and i/o scheduler. It also greatly differs per phone!!!!!
Click to expand...
Click to collapse
Totally agree with this post
Started this to keep the rom threads clean. Stay classy xda
Sent from my ASUS Transformer Pad TF300T using xda app-developers app
Ok, here goes. I flashed a 1600 OC capable kernel per the SeanZ JB Rom Hydro thread and it added a governor called "Pegasusq". It was suggested to me on a prior kernel that didn't include this mentioned one to just go with "On Demand".
Should I set Pegasusq to take over.
And also in that ROM thread above that I mentioned another user said they simply added the "SIO" Governor but I never got a response back as to it's usage benefits vs.s any of the above that I've mentioned.
Can anyone help here or explain a bit.
Also: To the OP - You mis-spelled Kernel <--- May want to fix so others take this thread more seriously. Just a suggestion.
I tested my friend some random class 6 or class 8 (he didnt know) card and here the results.
Antutu
http://db.tt/vwVFknnz
Androbench
http://db.tt/EUohUz5b
Quadrant
http://db.tt/xK9NGXXT
And offcourse hydro 4.1 rom and Data2SD_OC_1.8_Kernel, still waitingt that faster card. Damm the finnish postal service is so slow
Still trying to figure how to add vdd (voltage) change in kernel. I get voltages to show in setcpu, but im not able to "apply" them. Well mayby next days. http://db.tt/hslKzmAo
lazer9 said:
Ok, here goes. I flashed a 1600 OC capable kernel per the SeanZ JB Rom Hydro thread and it added a governor called "Pegasusq". It was suggested to me on a prior kernel that didn't include this mentioned one to just go with "On Demand".
Should I set Pegasusq to take over.
And also in that ROM thread above that I mentioned another user said they simply added the "SIO" Governor but I never got a response back as to it's usage benefits vs.s any of the above that I've mentioned.
Can anyone help here or explain a bit.
Also: To the OP - You mis-spelled Kernel <--- May want to fix so others take this thread more seriously. Just a suggestion.
Click to expand...
Click to collapse
Lol thanks for pointing that out
Sent from my SAMSUNG-SGH-I727
Niksunen said:
I tested my friend some random class 6 or class 8 (he didnt know) card and here the results.
Antutu
http://db.tt/vwVFknnz
Androbench
http://db.tt/EUohUz5b
Quadrant
http://db.tt/xK9NGXXT
And offcourse hydro 4.1 rom and Data2SD_OC_1.8_Kernel, still waitingt that faster card. Damm the finnish postal service is so slow
Still trying to figure how to add vdd (voltage) change in kernel. I get voltages to show in setcpu, but im not able to "apply" them. Well mayby next days. http://db.tt/hslKzmAo
Click to expand...
Click to collapse
This the data mod or normal?
Sent from my SAMSUNG-SGH-I727
seshaz said:
This the data mod or normal?
Sent from my SAMSUNG-SGH-I727
Click to expand...
Click to collapse
Like i say Data2sd kernel, so it data mod
Niksunen said:
Like i say Data2sd kernel, so it data mod
Click to expand...
Click to collapse
Cool i have to try that, maybe this weekend I'll do the formatting
Sent from my SAMSUNG-SGH-I727
lazer9 said:
Ok, here goes. I flashed a 1600 OC capable kernel per the SeanZ JB Rom Hydro thread and it added a governor called "Pegasusq". It was suggested to me on a prior kernel that didn't include this mentioned one to just go with "On Demand".
Should I set Pegasusq to take over.
And also in that ROM thread above that I mentioned another user said they simply added the "SIO" Governor but I never got a response back as to it's usage benefits vs.s any of the above that I've mentioned.
Can anyone help here or explain a bit.
Also: To the OP - You mis-spelled Kernel <--- May want to fix so others take this thread more seriously. Just a suggestion.
Click to expand...
Click to collapse
The sio doesn't seem to be performing as well with the i/o speeds for me
Sent from my SAMSUNG-SGH-I727
I tried Pegasusq and it didn't work out well for me. Caused many random reboots. Back to ondemand for now.
Remember, I'm on that 1600hz or whatever Kernel - don't remember the exact name, would have to dig through the last 5 or so pages of Seanz' Hydro JB thread to even figure out which one I'm on. I do have a 32GB Ext SDCard in this sucker but not sure what rating it is. - would it say on the card and if so, would a different (please be specific) Data2SD kernel be more to my suiting?
Thanks guys!
PS: And thanks for correcting the title of the thread
lazer9 said:
I tried Pegasusq and it didn't work out well for me. Caused many random reboots. Back to ondemand for now.
Remember, I'm on that 1600hz or whatever Kernel - don't remember the exact name, would have to dig through the last 5 or so pages of Seanz' Hydro JB thread to even figure out which one I'm on. I do have a 32GB Ext SDCard in this sucker but not sure what rating it is. - would it say on the card and if so, would a different (please be specific) Data2SD kernel be more to my suiting?
Thanks guys!
PS: And thanks for correcting the title of the thread
Click to expand...
Click to collapse
Yeah the rating should be on the card, a number with a letter C around it. Well at least most of them have it. Just check your card out and type whatever you see in it so we can help you identify it.
For example this sdcard is class 4 you can see the number inside the C: http://40.76.my/Malaysia/[email protected]
---------- Post added at 12:38 AM ---------- Previous post was at 12:10 AM ----------
By the way so many kernel development lately, makes me want to unlock my tablet this is what I was waiting for, since ASUS hasn't really tackled the I/O problem so far. First I need to get me a new sd card preferably class 10...and soon I'll take the jump.
jgaf said:
Yeah the rating should be on the card, a number with a letter C around it. Well at least most of them have it. Just check your card out and type whatever you see in it so we can help you identify it.
For example this sdcard is class 4 you can see the number inside the C: http://40.76.my/Malaysia/[email protected]
---------- Post added at 12:38 AM ---------- Previous post was at 12:10 AM ----------
By the way so many kernel development lately, makes me want to unlock my tablet this is what I was waiting for, since ASUS hasn't really tackled the I/O problem so far. First I need to get me a new sd card preferably class 10...and soon I'll take the jump.
Click to expand...
Click to collapse
Runs so much better with one of these kernels , this weekend I'm gunna go get a bigger sd and do the data2sd
Sent from my SAMSUNG-SGH-I727 using xda app-developers app
jgaf said:
Yeah the rating should be on the card, a number with a letter C around it. Well at least most of them have it. Just check your card out and type whatever you see in it so we can help you identify it.
For example this sdcard is class 4 you can see the number inside the C: http://40.76.my/Malaysia/[email protected]
---------- Post added at 12:38 AM ---------- Previous post was at 12:10 AM ----------
By the way so many kernel development lately, makes me want to unlock my tablet this is what I was waiting for, since ASUS hasn't really tackled the I/O problem so far. First I need to get me a new sd card preferably class 10...and soon I'll take the jump.
Click to expand...
Click to collapse
Lol, and crap! :laugh: I forgot, this is the ****ty 16GB I had swapped out of my Droid X: It's a (2) within the C
Wondering how this one would fair: SanDisk Ultra 64 GB microSDXC Class 10 UHS-1 Memory Card 30MB/s which I brought up in a recent discussion in the Accessories section? http://www.amazon.com/SanDisk-micro...64G-U46A/dp/B007WTAJTO/ref=reg_hu-rd_add_1_dp
lazer9 said:
Lol, and crap! :laugh: I forgot, this is the ****ty 16GB I had swapped out of my Droid X: It's a (2) within the C
Wondering how this one would fair: SanDisk Ultra 64 GB microSDXC Class 10 UHS-1 Memory Card 30MB/s which I brought up in a recent discussion in the Accessories section? http://www.amazon.com/SanDisk-micro...64G-U46A/dp/B007WTAJTO/ref=reg_hu-rd_add_1_dp
Click to expand...
Click to collapse
I ordered this card. Atleast it fast 95MB/s But delivery isnt very fast ordered almost week ago and not yet arrived. Well, it coming from china, so http://www.ebay.com/itm/SANDISK-EXTREME-PRO-MICRO-SDHC-16GB-95MB-16G-16-G-GB-MICRO-SD-UHS-I-C10-MEMORY-/221113862626?pt=PDA_Accessories&hash=item337b69c9e2
Compiling just new kernel, with more I/O SCHEDULERS, sio alreydy added, but now vr and deadline. I found here in xda forums, someone writings http://forum.xda-developers.com/showpost.php?p=22134559&postcount=4
and there i spotted " 2) Deadline
Goal is to minimize I/O latency or starvation of a request. The same is achieved by round robin policy to be fair among multiple I/O requests. Five queues are aggressively used to reorder incoming requests.
Advantages:
Nearly a real time scheduler.
Excels in reducing latency of any given single I/O.
Best scheduler for database access and queries.
Bandwidth requirement of a process - what percentage of CPU it needs, is easily calculated.
Like noop, a good scheduler for solid state/flash drives."
So i wanted to test it and compiled kernel with it. testing it soon, and then i have to go sleep
Niksunen said:
I ordered this card. Atleast it fast 95MB/s But delivery isnt very fast ordered almost week ago and not yet arrived. Well, it coming from china, so http://www.ebay.com/itm/SANDISK-EXTREME-PRO-MICRO-SDHC-16GB-95MB-16G-16-G-GB-MICRO-SD-UHS-I-C10-MEMORY-/221113862626?pt=PDA_Accessories&hash=item337b69c9e2
Click to expand...
Click to collapse
Hrmm. 95MB.. nice but I was hoping for both speed and storage space so how would you think (the one I posted above from Amazon) would fair with the 2SD Kernels?
lazer9 said:
Hrmm. 95MB.. nice but I was hoping for both speed and storage space so how would you think (the one I posted above from Amazon) would fair with the 2SD Kernels?
Click to expand...
Click to collapse
sorry but i don't know, have to test it before i can say anything sure.
But once i get my card i post bencmark results.
Niksunen said:
I ordered this card. Atleast it fast 95MB/s But delivery isnt very fast ordered almost week ago and not yet arrived. Well, it coming from china, so http://www.ebay.com/itm/SANDISK-EXT...862626?pt=PDA_Accessories&hash=item337b69c9e2
Compiling just new kernel, with more I/O SCHEDULERS, sio alreydy added, but now vr and deadline. I found here in xda forums, someone writings http://forum.xda-developers.com/showpost.php?p=22134559&postcount=4
and there i spotted " 2) Deadline
Goal is to minimize I/O latency or starvation of a request. The same is achieved by round robin policy to be fair among multiple I/O requests. Five queues are aggressively used to reorder incoming requests.
Advantages:
Nearly a real time scheduler.
Excels in reducing latency of any given single I/O.
Best scheduler for database access and queries.
Bandwidth requirement of a process - what percentage of CPU it needs, is easily calculated.
Like noop, a good scheduler for solid state/flash drives."
So i wanted to test it and compiled kernel with it. testing it soon, and then i have to go sleep
Click to expand...
Click to collapse
Looking forward to trying it
Sent from my SAMSUNG-SGH-I727 using xda app-developers app
Wow
Just updated my Kernel to Untermench's 1.6ghz thats listed in his OP.... all i can say is wow, this thing is fast. I would love to see what this tablet can do with a cranked GPU.... Im running Hydro 4.1 and have enabled all scripts with scriptmanager at boot, also this is no data2sd....just standard I/O, just goes to prove how much ASUS didnt optimize Jellybean for the TF300 beast
TEW999 said:
Just updated my Kernel to Untermench's 1.6ghz thats listed in his OP.... all i can say is wow, this thing is fast. I would love to see what this tablet can do with a cranked GPU.... Im running Hydro 4.1 and have enabled all scripts with scriptmanager at boot, also this is no data2sd....just standard I/O, just goes to prove how much ASUS didnt optimize Jellybean for the TF300 beast
Click to expand...
Click to collapse
Hmm very intresting. I look at it, and play a litle with my kernel, i have to look what speeds can i get.
Niksunen said:
Hmm very intresting. I look at it, and play a litle with my kernel, i have to look what speeds can i get.
Click to expand...
Click to collapse
also just to point out im using the SIO scheduler, My tablet really flies under my finger with this setup..... just looking for that bit of extra gaming umphh if you get my drift lol
TEW999 said:
Just updated my Kernel to Untermench's 1.6ghz thats listed in his OP.... all i can say is wow, this thing is fast. I would love to see what this tablet can do with a cranked GPU.... Im running Hydro 4.1 and have enabled all scripts with scriptmanager at boot, also this is no data2sd....just standard I/O, just goes to prove how much ASUS didnt optimize Jellybean for the TF300 beast
Click to expand...
Click to collapse
All scripts? I only enabled a few and only get around 5300 & like 5000 on I/O. I'm going to try all and see. That kernel is pretty legit though
Sent from my SAMSUNG-SGH-I727 using xda app-developers app