Hi All Mighty Developers!
You all work hard to satisfy our needs on ROM and App developent but from our point of view (end user) there is no objective and comparable results of developed ROMs for HTC devices.
My idea is to develop a software that is capable of measuring device:
- MTBF (Mean time between failures) or between soft/hard reset
- Average and low high levels of RAM usage
- Average CPU levels, Net usage, etc.
- Device Manufacturer, Type,
- Device OS, ROM, ROM version
This is a small footprint of data which can be uploaded to a site. At server side data could be analysed and integrated into a structural view to hellp users decide to install ROM or not.
ROM developers should be allowed to register their ROM and decide weather to include the client software or not. They can maintain their ROM versions and follow its data.
Regards,
cina.
you are right. these is no objective way to measure performance of roms. the systematic benchmarks are totally usless
i think its much more useful to develop a simple application that does a messure of how long it takes to complete a set of well known tasks.
the application would for example run a script to visually start mail appication then run file explorer then create a tcalender item then change the time then copy a file to the SD card and read from it. the output would be the time it takes to do this. this should be extremely useful for rom cookers to really help them optimise their roms
anybody willing to give this a shot. it shouldnt be that complicated
That is a one shot execution of measuring performance of a software which is influenced by many factors - currently running applications, memory usage, etc.
If execution of a software needs to be measured (objectively and comparable) I think it needs to run several times - so a monitoring application should be developed which records timing of every execution and then averages results.
MTBF and the other parameters I've mentioned in post #1 also needs a monitoring method to develop. Small footprint as a service separated from the GUI which is only running during configuration, uploading result, etc.
I have a lot of ideas with this. But I'm not a PPC developer. I have some experience in desktop application development so I might help some desperate programmers with a basic specification and design if needed.
Idea is absolutely useless, imho.
Too many variables.
I am using soft i've added to rom to FEEL difference between roms, checking:
- overall speed in real life conditions,
- doc file reading test(ebook reader and word),
- emulators test(cpu and gfx speed test),
- overall GUI speed,
- cam performance differencies,
- vid playback differencies,
- battery behaviour in similar conditions,
- mem drain after launching/using exactly same set of soft(interesting - mem drain is always related to phone subsystem),
- some synthetic tests related to gapi, gdi etc(BUT NEVER SPB bench, that program is ridiculous!),
- filetypes read check(to test associations/codecs).
And you know what..i haven't had single wm crash, batt drain or weak performance(cpu,batt) for 4 years, always used best base available(funny thing is TO FIND such base...many builds are needed..), tweaked to the max, these tests always worked for me, all weak/problematic builds ALWAYS showed potential problems/weaknesses after such test set.
First thing is to know your device, feel it(let's say).
I think that ONE program to measure these:
MTBF (Mean time between failures) or between soft/hard reset
- Average and low high levels of RAM usage
- Average CPU levels, Net usage, etc.
- Device Manufacturer, Type,
- Device OS, ROM, ROM version
Click to expand...
Click to collapse
is just not good idea - vide spb benchmark - results ARE NOT reliable at all(to be honest - that soft is BULL****).
I assume it is just because it is ONE program(and it serves salary, not REAL benching), such thing has nothing to reallife usage schema.
sry for my 2 cents.
upd:
MTBF (Mean time between failures) or between soft/hard reset
Click to expand...
Click to collapse
this is seriously ridiculous - HOW can one test something like that, LOL?
pupakota said:
this is seriously ridiculous - HOW can one test something like that, LOL?
Click to expand...
Click to collapse
Well not that much ridiculous I think! In server world one important measure is availability which is primary based on MTBF. Supposing that an ideal PPC never needs to be soft/hard reseted an average measure of uptime of your ROM is an approximate of stability. Statistics I mean.
Averaging thousands of results of all previously cooked ROMS stability of device is measured which can be a valuable feedback to the manufacturer.
(I think I think in wide usually and open as well. Probably my nickname is producing a lot of negative feedback to my ideas. Maybe I should test this more. )
cina said:
(I think I think in wide usually and open as well. Probably my nickname is producing a lot of negative feedback to my ideas. Maybe I should test this more. )
Click to expand...
Click to collapse
Ohh, do not take it personally, please...
Averaging thousands of results of all previously cooked ROMS stability of device is measured which can be a valuable feedback to the manufacturer.
Click to expand...
Click to collapse
Do You think they WOULD CARE? It's all about salary, nothing more.
---
Notice, that NOBODY is using stock roms, too(and it is quite clear, there's reason for that ).
It may be interesting - reporting cooked roms performance to...stock rom manufacturer
pupakota said:
Ohh, do not take it personally, please...
Click to expand...
Click to collapse
Paranoid a bit I am. Anyway thx for the encouragement.
pupakota said:
Do You think they WOULD CARE? It's all about salary, nothing more.
---
Notice, that NOBODY is using stock roms, too(and it is quite clear, there's reason for that ).
It may be interesting - reporting cooked roms performance to...stock rom manufacturer
Click to expand...
Click to collapse
I know everything is driven by money. Even in healthcare where I currently work.
I don't want to repeat myself. But the basic idea was to help end users choose the ROM that best suits their needs. And to help chefs with feedback on their jobs.
Parallel with medical work all doctors need feedback about their patients to achieve better level of practice. A chef should learn from data collected from end-users and enhance their ROM development.
I understand that a lot of you cook for themselves and publish the result (one time cookers or beginners). But there are big names of cookers with huge experience working hard reading the posts about their ROMS and trying to debug based on that. Wouldn't it be easier to get collected data on ROM performance/stability, bug reports, SW incompatiblity, etc. in an organized way?
Anyway you said that
pupakota said:
And you know what..i haven't had single wm crash, batt drain or weak performance(cpu,batt) for 4 years, always used best base available
Click to expand...
Click to collapse
How do you know? Probably you are a lucky one that got an exceptional stable hardware, a battery with no weak cells and a wireless chip with no misaligned molecule in its layers. Others probably have weak hardware, SW from other source that you have never tested. Are you quite sure that your ROM is prepared for "all" situation? Probably not and no one can be sure about that. (And do not take it personal). "Perfection cannot be reached but we should try it hard." - source unknown.
So literally I propose a software that measure some performance/stability/compatiblity issues of cooked ROMs and I've just listed my ideas. Anyone who will develop a software like this will add its own ideas and probably suppress/emphasize some of mine. I intended just to inspire developers for such a solution. I didn't want to give a clear solution anyway but I offered some help with a work like that.
Best regards,
Cina.
Allright, good luck, really...
Paranoid a bit I am.
Click to expand...
Click to collapse
Me too, mate, me too
I understand that a lot of you cook for themselves and publish the result (one time cookers or beginners).
Click to expand...
Click to collapse
btw i am not beginner. i stopped making public roms 2 yrs ago.
How do you know?
Click to expand...
Click to collapse
..because i can see it.
my softpack was tuned for almost 2 yrs, practically every single apps' oem is made by me, every single app is reconfed by me.
my methods for getting optimal os base were tuned for almost 5 yrs(and regardless of base, builds are always made with same, OLD methods).
zero HTC soft inside. zero useless soft inside. zero useless dlls, zero useless reg entries. bases with ANY kind of problem are thrown out and not used by me.
i do not need any external info from anyone to get optimal results. strict methods have very important feature: elimination of sudden misbehaviours, unknown problems, so only problems with any of: performance, batt,whatever, may come from base then.
That's all.
i do not know anything about others' point of view on external performance/etc. data ideas, but it may be hard to believe that good cooks are not having similar ways. You know, i see that someone tells: "That base is good", allright, 20 minutes later i have ready rom, so it's quite fast way to check, if it is good base, then practical subsystems tests in reallife verificate REAL usability. btw hard to imagine such build list, if we cannot have EVEN clear com 2,3,5 builds list, lol.
sorry for LQ english(or engrish..).
"Perfection cannot be reached but we should try it hard."
Click to expand...
Click to collapse
So true...
Thanx. And I appreciate your posts. Really.
pupakota said:
btw i am not beginner.
Click to expand...
Click to collapse
I've already got that.
Related
I've recently come across a kernel patch that according to reports it greatly improves performance in resource management. People are reporting amazing results. You can read about it here, and you can find the current patch here.
As I haven't really messed with kernels even though I'm a linux user (I don't even know if it's possible to apply a generic kernel patch in Android kernels) I thought I'd bring this to your attention, in case it can be applied to our 2.6.32.x.
I have taken a look at the patch.
It indeed works wonders in real world multithreaded workloads, but we will fully be able to take advantage of it (the patch) once we have dual core smartphones.
Let me ask you 1 thing: how much do you multitask on your handset?
;]
Unless you have multiple services running in the background and hogging your res, this isn't such a big deal (yet).
adwinp said:
I have taken a look at the patch.
It indeed works wonders in real world multithreaded workloads, but we will fully be able to take advantage of it (the patch) once we have dual core smartphones.
Let me ask you 1 thing: how much do you multitask on your handset?
;]
Unless you have multiple services running in the background and hogging your res, this isn't such a big deal (yet).
Click to expand...
Click to collapse
does it take a lot of work to apply this patch to a DHD kernel?
maybe it can make things a lil bit smoother : )
adwinp said:
Let me ask you 1 thing: how much do you multitask on your handset?
;]
Unless you have multiple services running in the background and hogging your res, this isn't such a big deal (yet).
Click to expand...
Click to collapse
That's true, we don't really multitask much (if not at all)
But since this manages the resources in a better way when running multiple apps, wouldn't it make a difference on the long run (for example if you don't reboot your device some of the apps idle as background processes) making the device running smoother? Or does it only benefit us when running multiple processes at the same time?
Just some food for thought, I don't really know about this myself but from what I understand you could shed some light!
PS: If someone decides to compile a kernel with that patch, in case they believe it will make a difference, I'm willing to flash and report back.
I say, if it's not too much work, than let's do it, and we'll see what happens
easier to observe what happens, than find out what would if...
i multitask alot on my DHD, go from checking an email straight to the browser then back to email and then to sending an SMS
I'll try patch it into my kernel. To see if it has anything to bring to android
Second look at it it may take a bit of work
Apache14 said:
I'll try patch it into my kernel. To see if it has anything to bring to android
Second look at it it may take a bit of work
Click to expand...
Click to collapse
super cool C:
Apache14 said:
I'll try patch it into my kernel. To see if it has anything to bring to android
Second look at it it may take a bit of work
Click to expand...
Click to collapse
You will be in for some hacking. The patch doesn't apply cleanly.
You can make a quick diff against the 37-rc2 kernel though; it's mainly scheduler stuff.
So I'm just wanting to increase my own understanding of how things work, plus I see people ask this question about specific rom's often. Given 2 different roms with identical kernels and identical setups, how do you explain different power usage and battery life?
In other words, why would you see a noticeable difference in battery life with one rom vs another, given all apps equal and kernels equal?
kingston73 said:
So I'm just wanting to increase my own understanding of how things work, plus I see people ask this question about specific rom's often. Given 2 different roms with identical kernels and identical setups, how do you explain different power usage and battery life?
In other words, why would you see a noticeable difference in battery life with one rom vs another, given all apps equal and kernels equal?
Click to expand...
Click to collapse
Hi,
I think it's a well known fact that not all phones are created equal.
Plus,no two people use their phone in the same way.
So,you will get variations all the time.
I think the OP is talking about 1 user and 1 phone having different results (different power usage and battery life) with roms that use the same base, kernel, etc.
I personally think that certain devs tweak some settings a little and the slight difference we notice comes from this. But since I'm not a Pro, I'm going to let someone else answer it, as I been looking for this answer.
Thanks OP for asking it.
rasco702 said:
I think the OP is talking about 1 user and 1 phone having different results (different power usage and battery life) with roms that use the same base, kernel, etc.
I personally think that certain devs tweak some settings a little and the slight difference we notice comes from this. But since I'm not a Pro, I'm going to let someone else answer it, as I been looking for this answer.
Thanks OP for asking it.
Click to expand...
Click to collapse
Hi,
Sorry, I mis-read.
You are right about the Devs doing their own thing,and tweaking every ROM differently.
They are always trying to squeeze every drop of improvement out of the ROM for the user.
Mike 1986 is a good example of this.
On every changelog of his are the words:
"Other minor changes and fixes"
rasco702 said:
I think the OP is talking about 1 user and 1 phone having different results (different power usage and battery life) with roms that use the same base, kernel, etc.
I personally think that certain devs tweak some settings a little and the slight difference we notice comes from this. But since I'm not a Pro, I'm going to let someone else answer it, as I been looking for this answer.
Thanks OP for asking it.
Click to expand...
Click to collapse
That's exactly what I'm asking. I'm not wondering why I get different results from you, I'm wondering why I get different results from one rom to the next with MY phone and MY same apps etc.
From here: http://forum.xda-developers.com/showthread.php?t=1987032
I came to that thread via LifeHacker, in which they featured it. I'm going to try this now on my freshly-updated-to-Jelly-Bean Note 10.1.
If you want to share a feedback, please post it here. If you want to thank the developer, suggest something, or buy him a beer, please post it on his thread.
APK: Click here (original direct link from his thread)
lambgx02 said:
Hey everyone,
So, I was experiencing significant lag as we all do from time to time, and decided I was going to get to the bottom of it.
After tracing and debugging for hours, I discovered the source of 90% of Android's lag. In a word, entropy (or lack thereof).
Google's JVM, like Sun's, reads from /dev/random. For all random data. Yes, the /dev/random that uses a very limited entropy pool.
Random data is used for all kinds of stuff.. UUID generation, session keys, SSL.. when we run out of entropy, the process blocks. That manifests itself as lag. The process cannot continue until the kernel generates more high quality random data.
So, I cross-compiled rngd, and used it to feed /dev/urandom into /dev/random at 1 second intervals.
Result? I have never used an Android device this fast. :good:
It is literally five times faster in many cases. Chrome, maps, and other heavy applications load in about 1/2 a second, and map tiles populate as fast as I can scroll. Task switching is instantaneous. You know how sometimes when you hit the home button, it takes 5-10 seconds for the home screen to repopulate? Yeah. Blocking on read of /dev/random. Problem solved. But don't take my word for it .. give it a shot!
Update!
I've built a very simple Android app that bundles the binary, and starts/stops the service (on boot if selected). I'll be adding more instrumentation, but for now, give it a shot! This APK does not modify /system in any way, so should be perfectly safe.
This is my first userspace Android app, so bear with me!
Note that this APK is actually compatible with all Android versions, and all (armel) devices. It's not at all specific to the Captivate Glide.
Caveats
There is a (theoretical) security risk, in that seeding /dev/random with /dev/urandom decreases the quality of the random data. In practice, the odds of this being cryptographically exploited are far lower than the odds of someone attacking the OS itself (a much simpler challenge).
This may adversely affect battery life, since it wakes every second. It does not hold a wakelock, so it shouldn't have a big impact, but let me know if you think it's causing problems. I can add a blocking read to the code so that it only executes while the screen is on. On the other hand, many of us attribute lag to lacking CPU power. Since this hack eliminates almost all lag, there is less of a need to overclock, potentially reducing battery consumption.
If you try it, let me know how it goes.
ROM builders - feel free to integrate this into your ROMs (either the .apk / application, or just the rngd binary called from init.d)!
If anyone's interested, I've launched a paid app on the Play store for non-xda users. As I add features I'll post the new versions here as a thanks to you guys (and xda community at large for being such a great resource). But if anyone's interested in the market's auto-update feature, just thought I'd mention it.
Cheers! :highfive:
Click to expand...
Click to collapse
I don't know if it's a placebo effect, but it seems that enabling this entropy seeder makes the device faster for real ^^
Does this need root?
Sent from my Nexus 4 using xda app-developers app
tuffballa said:
Does this need root?
Sent from my Nexus 4 using xda app-developers app
Click to expand...
Click to collapse
Yeah, I forgot to add.
Redemption rom 4.1.2 is pretty quick and only lag was keyboard opening.
This mod keyboard instant opening now .
jje
Sent from my GT-N8010 using xda premium
Post one on the portal has more info .
http://www.xda-developers.com/android/entropy-seed-generator-not-all-its-hacked-up-to-be/
jje
Without reading the portal piece I guess the reason you may perceive speed up as that the device will be spending longer on higher cpu cycles by this polling.
In other words it is not a good thing for battery.
You can try playing with the governor settings to increase responsiveness, but "lag" can be caused by quite a lot of things. Random seeding shouldn't be one of them.
Sent from my GT-N7000 using Tapatalk 2
A lot of developers are stating that this is a fluke on anything running ICS+. Any boost in performance is 99% a placebo effect.
Placebo effect or not, I can notice improvement in the note general operation.
Sent from my GT-N8000 using Tapatalk HD
Disprove it or walk away.
Any dev saying it's a placebo effect needs to hold up data to prove the assertion, otherwise, the dev is just speculating. Being an Android dev can mean a variety of things, so the title doesn't grant authority.
Any dev claiming random data seeding shouldn't have an effect, should contact the APK writer and ask to see the trace logs, i.e. duplicate the experiment and verify the results. Otherwise, see above.
Everyone else, download it. Enjoy a faster UI experience. I confirm it works, which is at least as good as any dev who hasn't tested it.
--#
7April said:
Any dev saying it's a placebo effect needs to hold up data to prove the assertion, otherwise, the dev is just speculating. Being an Android dev can mean a variety of things, so the title doesn't grant authority.
Any dev claiming random data seeding shouldn't have an effect, should contact the APK writer and ask to see the trace logs, i.e. duplicate the experiment and verify the results. Otherwise, see above.
Everyone else, download it. Enjoy a faster UI experience. I confirm it works, which is at least as good as any dev who hasn't tested it.
--#
Click to expand...
Click to collapse
Have *you* checked logs to see what's happening as you claim you can see a performance hike.
I tried it on 3 devices earlier, all did exactly what I suspected. Pushed the cpu time up quite a bit, but that's what this does. It's forcing an io which causes the cpu to ramp, so of course it feels more responsive, it's starting from a higher cpu clock and less time in the lower click rates.
One of my services it caused serious slow down, die to pegging the cpu as it didn't play nicely with the particular aml quadcore governor.
Sent from my GT-N7000 using Tapatalk 2
fards said:
Have *you* checked logs to see what's happening as you claim you can see a performance hike.
I tried it on 3 devices earlier, all did exactly what I suspected. Pushed the cpu time up quite a bit, but that's what this does. It's forcing an io which causes the cpu to ramp, so of course it feels more responsive, it's starting from a higher cpu clock and less time in the lower click rates.
One of my services it caused serious slow down, die to pegging the cpu as it didn't play nicely with the particular aml quadcore governor.
Sent from my GT-N7000 using Tapatalk 2
Click to expand...
Click to collapse
Couldn't agree more with your comments, I also tested this on a variety of devices and got exactly the same results, on one of my devices I was loosing 5% battery per hour!
seeder is working on all devices?
xda just had an apology article for featuring this lol
Hey guys,
So, I was wondering.. I am new to the all android experience, I am learning and I am enjoying tweaking and messing around with my phone.
Problem is that I am really un knowledgeable when it comes to kernels and roms..
I tested at least 5 kernels today and tbh apart from system performance on antutu i'm not seeing much, i'm wondering how people can reach 19 hours on here when no matter which rom or kernel I pick my phone just keeps draining pretty fast when i'm playing games..
I want to learn, to understand more.. but for that, I need help.
I want to know how to select a good rom, a good kernel, what features to look at, how to mess with voltages and etc..
I would love to learn, so someday I can be the one assisting other android users.
MrFishiz said:
Hey guys,
So, I was wondering.. I am new to the all android experience, I am learning and I am enjoying tweaking and messing around with my phone.
Problem is that I am really un knowledgeable when it comes to kernels and roms..
I tested at least 5 kernels today and tbh apart from system performance on antutu i'm not seeing much, i'm wondering how people can reach 19 hours on here when no matter which rom or kernel I pick my phone just keeps draining pretty fast when i'm playing games..
I want to learn, to understand more.. but for that, I need help.
I want to know how to select a good rom, a good kernel, what features to look at, how to mess with voltages and etc..
I would love to learn, so someday I can be the one assisting other android users.
Click to expand...
Click to collapse
Only one way to do that : READ READ READ READ and read some more. And don`t be afraid to try things like flashing roms, kernels, recovery`s, mods. Always read the OP`s of the devs well so you`ll know what to do,
gee2012 said:
Only one way to do that : READ READ READ READ and read some more. And don`t be afraid to try things like flashing roms, kernels, recovery`s, mods. Always read the OP`s of the devs well so you`ll know what to do,
Click to expand...
Click to collapse
Thing is that most people here are already.. " in the zone " when it comes to knowledge, so they speak the language that I don't understand.. it's like chinese when I first look at it.
How does a beginner evolve like that?
Btw: I was wondering maybe I should test kernels and roms for a period of time instead of periods of hours? because I don't see immediate results I just switch a rom or a kernel not really giving it a chance to evolve
MrFishiz said:
Thing is that most people here are already.. " in the zone " when it comes to knowledge, so they speak the language that I don't understand.. it's like chinese when I first look at it.
How does a beginner evolve like that?
Btw: I was wondering maybe I should test kernels and roms for a period of time instead of periods of hours? because I don't see immediate results I just switch a rom or a kernel not really giving it a chance to evolve
Click to expand...
Click to collapse
All you need to know about adroid http://forum.xda-developers.com/showthread.php?t=2510966
Learn how to setup and use adb, unlock the bootloader and return to stock if need be? See http://forum.xda-developers.com/showthread.php?t=2513701.
Also read this usefull thread http://forum.xda-developers.com/showthread.php?t=2507905.
Read this also (for beginners very helpfull) http://forum.xda-developers.com/showthread.php?t=2510966
Installing usb drivers http://forum.xda-developers.com/showthread.php?t=2510966.
This is a very good start to stop beeing a` NOOB` and eliminate the risk of bricking your phone.
All users start from scratch bro, we all live and learn till the day we die
First don't care about bemchmarks. They mean nothing in term of user experience.
Second a kernel must be tried at least during one or two days.
Third there is no best Rom our best kernel. Golden rule. Each person has different needs and tastes.
And be careful when you see amazing battery times. What matters is phone usage. Good phone network ? Wi-Fi? Localisation always on? And so on.
What surprises you when you play your battery drains? Normal.
What makes difference is 'idle' time. Use Greenify, fight backend running apps.
Some people spend days tweaking their phones, and though don't have time to use it...
MrFishiz said:
my phone just keeps draining pretty fast when i'm playing games..
Click to expand...
Click to collapse
this is normal... what do you expect when you stress the CPU/GPU?
Thing is, you don't just change kernels and run. Many of the efficiencies you gain with custom kernels comes from the tweaks you make to the system made possible by the kernel. These are where you can get increased performance in many respects. Keep in mind, though, no matter what you do, the processors will still be hungry under load and there is not much to do about screen use except brightness control. (i.e. games will kill your battery faster than just about anything else)
Sent from my Nexus 10 using XDA Premium HD app
Have you checked into finding out which apps and services may be causing your phone to wake up often, or even not sleep at all? Learn about wakelocks if you haven't already. There are good tools for finding out which ones are affecting your battery life, such as BetterBatteryStats.
Check out this thread: http://forum.xda-developers.com/showthread.php?t=1827676
start off by reading.. read... and read some more:good:
I would like to which CM12 build is best; speed, stability, and more.
spacejamnty24 said:
I would like to which CM12 build is best; speed, stability, and more.
Click to expand...
Click to collapse
(The Best topic) is on which works best for you. Here at XDA we.do not do a best topic to avoid confrontation and disrespect to devs
2SHAYNEZ
"Best" is a very relative term. To be honest, none of them are ready for daily driving. They all still seem to suffer from the Misc battery useage which can eat more power than the screen does (which is hard to believe, but true). Add in some proximity sensor issues, text issues, etc and you will quickly realize that they are all very much still beta releases. That's OK, because the developers state this and are trying really hard to fix all the issues - I'm not blaming them for the status of the ROMs. I'm just trying to give you an honest appraisal of the current situation.
Lollipop ROMs are great to load and play around with - and we definitely need people to help pin point and squash bugs - but they are not ready for daily use right now. So flash at your own risk.