03-12-10 06:00 PM
67 123
tools
  1. JRSCCivic98's Avatar
    I agree that it may be more of a Storm thing and RIM fixing some flaws that they learned about after the fact. When I got my Storm it had 4.7 on it and the battery was awful for the first couple days. Once I went to the official OS 5.0.0.419 the battery life was easily about 30% better. Not huge but enough to be noticable. I think it was just more efficient at controlling the functions of the Storm.

    I have never noticed improved battery on my Pearl or Curve with OS changes. The Storm was the first time.
    Agreed and I'm surprised AG didn't notice much change on the Bold either... that was the "first" device to gain some battery performance once they stabalized the radio issues.... however, if I remember correctly, I think that might have been done before AT&T released the Bold for sale. That might be why... prior to release, radio stability and battery along with heat issues were prevelent in that model. Maybe it was Orange that held up the release because of these issues, crap, I can't remember now. lol
    02-24-10 12:34 PM
  2. Radius's Avatar
    I'm a software engineer, let me throw out some wild guesses.

    The OS or any app for that matter can very easily affect battery life for good or bad. I can't say what is happening in the RIM OS as I've never opened it up but just about any aspect of the OS can cause more battery drain than another version.

    A simple unoptimized piece of code for a new feature for example could cause a power hungry operation to be performed constantly in the background without ever being put to sleep. Maybe something that requires a single calculation to work is actually doing the calculation 500 times. Been there, fixed that.

    It comes down to code execution and peripheral access on the board and IC's. One of the more power intensive applications is usually the erasure of the internal memory if you are doing a lot. I also suspect the processor has some functions built in that when accessed can cause pretty good drain on the battery if it's used too much.

    The OS is an application like any other, there is nothing to distinguish it from say the Slacker Radio app or Brick Breaker. It executes, it drains power.

    Having said all that, I have no idea what if any power gobbling problems have/do exist in the RIM OS and if there are any, have they been addressed in the upgrades. Possibly as there is always room for improvement but who knows if the average user will see any noticeable effect.
    02-24-10 12:50 PM
  3. amazinglygraceless's Avatar
    Agreed and I'm surprised AG didn't notice much change on the Bold either... that was the "first" device to gain some battery performance once they stabalized the radio issues.... however, if I remember correctly, I think that might have been done before AT&T released the Bold for sale. That might be why... prior to release, radio stability and battery along with heat issues were prevelent in that model. Maybe it was Orange that held up the release because of these issues, crap, I can't remember now. lol
    In all fairness Civic I should point out that I really did not have the Bold long
    before it's tragic demise. That may account for my not noticing what you are
    referencing.
    02-24-10 12:58 PM
  4. amazinglygraceless's Avatar
    I'm a software engineer, let me throw out some wild guesses.

    The OS or any app for that matter can very easily affect battery life for good or bad. I can't say what is happening in the RIM OS as I've never opened it up but just about any aspect of the OS can cause more battery drain than another version.

    A simple unoptimized piece of code for a new feature for example could cause a power hungry operation to be performed constantly in the background without ever being put to sleep.
    But that's where the theory breaks down. If this were true everyone using
    the OS would experience the same thing.

    Look at the posts on just this site alone and you will find that some users report
    better battery life with an OS and others will report degraded battery life with the
    very same OS. I just don't get this.
    02-24-10 01:01 PM
  5. Username00089's Avatar
    I think that might have been done before AT&T released the Bold for sale. That might be why... prior to release, radio stability and battery along with heat issues were prevelent in that model. Maybe it was Orange that held up the release because of these issues, crap, I can't remember now. lol
    It was AT&T. The Bold was pretty much released in September of that year
    everywhere else except the U.S. on AT&T. It wasn't released until election
    day of 2008 (November)

    Not exactly sure about Orange though. I know Rogers was for sure in September
    because I remember all the Canadians on here gloating about it.
    02-24-10 01:05 PM
  6. jgodin03's Avatar
    I too would like to see a good explanation. I always thought it was related to the radio and/or signal strength in a OS that contributed to the battery life.

    All OS run different radio file, different apps package.

    I think some app consume less energy in some version, and the radio file surelly have an effect.

    But this is only my opinion. I have no idea...
    02-24-10 03:57 PM
  7. mahen915's Avatar
    I'm a software engineer, let me throw out some wild guesses.

    The OS or any app for that matter can very easily affect battery life for good or bad. I can't say what is happening in the RIM OS as I've never opened it up but just about any aspect of the OS can cause more battery drain than another version.

    A simple unoptimized piece of code for a new feature for example could cause a power hungry operation to be performed constantly in the background without ever being put to sleep. Maybe something that requires a single calculation to work is actually doing the calculation 500 times. Been there, fixed that.

    It comes down to code execution and peripheral access on the board and IC's. One of the more power intensive applications is usually the erasure of the internal memory if you are doing a lot. I also suspect the processor has some functions built in that when accessed can cause pretty good drain on the battery if it's used too much.

    The OS is an application like any other, there is nothing to distinguish it from say the Slacker Radio app or Brick Breaker. It executes, it drains power.

    Having said all that, I have no idea what if any power gobbling problems have/do exist in the RIM OS and if there are any, have they been addressed in the upgrades. Possibly as there is always room for improvement but who knows if the average user will see any noticeable effect.
    The radio stack is one of the biggest ones, since just using the radio and sending/receiving data alone eats power. RIM apparently re-wrote the whole stack from ground up upon release of the 9700 and I hear battery life is a beast on that thing.

    I'd say the actual software and code of the OS would be next (the radio stack is very low level stuff). You're absolutely right on with the OS stuff. RIM could easily optimize memory sharing algorithms and eliminate computations that are unnecessarily consuming clock cycles. As programmers/software engineers, we know trying to optimize code by removing or rewriting as many loops as we can, we can greatly enhance performance.
    02-24-10 04:43 PM
  8. Radius's Avatar
    But that's where the theory breaks down. If this were true everyone using
    the OS would experience the same thing.

    Look at the posts on just this site alone and you will find that some users report
    better battery life with an OS and others will report degraded battery life with the
    very same OS. I just don't get this.
    It's not theory from my point of view, it's truth as I know it. Two things though, third party apps and perception. Those two will skew the results every time.
    02-24-10 09:43 PM
  9. Radius's Avatar
    The radio stack is one of the biggest ones, since just using the radio and sending/receiving data alone eats power. RIM apparently re-wrote the whole stack from ground up upon release of the 9700 and I hear battery life is a beast on that thing.

    I'd say the actual software and code of the OS would be next (the radio stack is very low level stuff). You're absolutely right on with the OS stuff. RIM could easily optimize memory sharing algorithms and eliminate computations that are unnecessarily consuming clock cycles. As programmers/software engineers, we know trying to optimize code by removing or rewriting as many loops as we can, we can greatly enhance performance.
    Radios are an interesting thing, and antennae even more so. The radio consumes a set amount of power and no amount of code will change that. At least that's been my experience with our wireless offerings. I think the big part there must be their layer between the physical radios and the OS itself, there's no telling what they have done.

    But then again if they were smart from the beginning they would just implement a simple queue and let the OS take care of it. But who knows.
    02-24-10 09:45 PM
  10. amazinglygraceless's Avatar
    It's not theory from my point of view, it's truth as I know it. Two things though, third party apps and perception. Those two will skew the results every time.
    I agree about the apps part but that is a totally separate issue. It still does
    not explain why the effect is not seen across the whole spectrum of users.

    I'm the exact opposite. When I first bought my 8900, it was running the original OS, 4.6.114 and the battery life was horrible. The phone would last about 4 or 5 hours even on standby. When was using 4.6.168, 4.6.310, and 5.0.411, the battery life was MUCH better than stock. The phone would last a whole day and I'd still have plenty of juice by midnight. Each OS brought better battery life but once I upgraded to 5.0, the battery life significantly improved.
    This again is totally anecdotal and specifically what I am trying to avoid.

    I honestly think the whole better or worse battery life thing is in the head
    of the user and has no actual basis in fact.

    All I want is for one person to definitively explain how if it is in fact true.
    Where are the hybrid makers on this one?
    02-25-10 03:02 PM
  11. KillYouWithMyMind's Avatar
    I just sorta skimmed thru the thread first of all...

    Given what I've read, the only logical hypothesis is the placebo theory. There are so many different factors to take into account that there's no possible way to narrow down any one specific factor of battery life, good or bad. IMO, the most likely culprit is memory leaks, whether it be from an app you're running or a problem with an "unofficial" 5.0 or a hybrid... who knows?
    02-25-10 04:45 PM
  12. LuvMyBB's Avatar
    AG, your question piqued my curiosity, so I did a little digging and came up with this nugget from RIM. There is definitely a cause/effect relationship between Java garbage collection and memory/battery performance. Better Java techniques and flash memory handling likely exists in later OS iterations, thereby improving battery life. (The flash memory handling sounds akin to de/fragmentation on a Wintel system.)

    I've also included the link at the bottom for your reading pleasure. (I'll also add that the jump from 4.6.0.297 to 5.0.0.411 was a *significant* increase in battery life for me, and my usage and the apps I use is pretty consistent. I know: anecdotal. )

    Causes for decreased performance

    BlackBerry smartphones based on Java technology depend on available flash memory for many tasks. Having low flash memory can lead to delays in processing information on the BlackBerry smartphone because the process for recovering blocks of flash memory, known as garbage collection, happens more frequently, which takes time. Java garbage collections automatically reclaim flash memory that is no longer being used by the Java operating system. Java garbage collections may also re-arrange blocks of flash memory that are in use to create larger contiguous blocks of available flash memory. Just as a computer uses the hard disk as a source of virtual memory, the BlackBerry smartphone writes out lesser-used portions of its operating flash memory (normally held in fast static random access memory (SRAM)) to the slower flash memory. As the BlackBerry smartphone's flash memory begins to fill up, the BlackBerry smartphone must work harder to re-arrange the flash memory to allocate the storage space it needs for normal operation. Therefore, low flash memory causes garbage collections to happen more frequently and take a longer time to complete, which also uses battery power.

    View Document
    Last edited by LuvMyBB; 02-25-10 at 05:29 PM.
    02-25-10 05:26 PM
  13. amazinglygraceless's Avatar
    My only issue with the above is that it is more of a memory management issue and
    less of an actual OS issue.

    One has to determine what is causing the Garbage Collector to kick in and most have
    nothing to do with the OS. A person who saves hundreds of emails, thousands of
    text messgaes and never closes IM apps will have memory issues to a greater
    degrees than someone who practices good "housekeeping" and has optimized
    their OS.

    While that explanation is valid, it really has nothing much to do with my question.
    02-25-10 06:13 PM
  14. elvin1983's Avatar
    Here's an issue that is right along the lines of your question AG, and it's one that's got me puzzled as well. I loaded the Sprint released 5.0.0.484 on my VZW Tour, and ever since, I've noticed a greater battery drain as opposed to the 4.7.0.61 from VZW I had loaded on there before. I read multiple posts from some people stating that since it's a Sprint release, that the radio programming is geared for the Sprint network, thus a VZW Tour will have to work harder to stay on their network, thus causing battery drain...
    I don't know the validity of this statement, but I know I'm not the only person who's had this problem with their Tour and this build.
    I've always been under the impression that aside from the vendor.xml file, an OS is the same regardless of carrier, so I don't know if this is a valid statement (I have a hard time buying it myself), but all the folks who build these "hybrid" OS's for Tours seem to believe in it... Just throwing in my two cents as it pertains to me.
    02-25-10 07:28 PM
  15. Garz's Avatar
    Agreed and I'm surprised AG didn't notice much change on the Bold either... that was the "first" device to gain some battery performance once they stabalized the radio issues.... however, if I remember correctly, I think that might have been done before AT&T released the Bold for sale. That might be why... prior to release, radio stability and battery along with heat issues were prevelent in that model. Maybe it was Orange that held up the release because of these issues, crap, I can't remember now. lol
    I ran the Bold 9000 from OS .167 on Nov 4th 08 all the way up to .266 before I sold it. I never noticed much change in battery life as it sucked period. The radio issue was fixed before Att released because of all the reboots on the network. The OS's did stabalized memory a lot as .167 would drop 20 MB's a day.
    02-25-10 07:40 PM
  16. amazinglygraceless's Avatar
    Here's an issue that is right along the lines of your question AG, and it's one that's got me puzzled as well. I loaded the Sprint released 5.0.0.484 on my VZW Tour, and ever since, I've noticed a greater battery drain as opposed to the 4.7.0.61 from VZW I had loaded on there before. I read multiple posts from some people stating that since it's a Sprint release, that the radio programming is geared for the Sprint network, thus a VZW Tour will have to work harder to stay on their network, thus causing battery drain...
    I don't know the validity of this statement, but I know I'm not the only person who's had this problem with their Tour and this build.
    I've always been under the impression that aside from the vendor.xml file, an OS is the same regardless of carrier, so I don't know if this is a valid statement (I have a hard time buying it myself), but all the folks who build these "hybrid" OS's for Tours seem to believe in it... Just throwing in my two cents as it pertains to me.
    This might be as close to a valid reason as I have seen. Gonna have to
    investigate it a bit more although I am a little suspect.
    02-25-10 07:59 PM
  17. elvin1983's Avatar
    This might be as close to a valid reason as I have seen. Gonna have to
    investigate it a bit more although I am a little suspect.
    As am I, especially since it goes against everything I've ever read about the BlackBerry OS.
    There's quite a few threads on this sort of thing in the Tour forum... Here are a couple of links if you're interested:
    http://forums.crackberry.com/f151/bb...erizon-426673/
    http://forums.crackberry.com/f151/ju...84-vzw-428002/
    It's a start...
    02-25-10 09:12 PM
  18. amazinglygraceless's Avatar
    Thanks Elvin. I'll read through them both.
    02-25-10 09:19 PM
  19. elvin1983's Avatar
    You're welcome, I'll be interested in what you think, because up until the past couple days I've never experienced an OS causing an excessive battery drain. This is the first time I've loaded an OS on one of my BB's that wasn't an official release from my carrier, so it's got me interested in finding out what causes this issue.
    For reference, when I had 4.7, I could go thru a whole day of regular use and have over 60% battery life, and that'd be on the low side, and today with the same use with 5.0.0.484, I'm down to 25% battery life. I never have any apps running in the background other than the usual ones, and I only have a handful of apps or so, none of which have any leaks.

    Posted from my CrackBerry at wapforums.crackberry.com
    02-25-10 09:39 PM
  20. amazinglygraceless's Avatar
    See Elvin that is the thing. I have six additional that are always resident and as
    I mentioned once before over 50 apps. Unless I am doing something data intensive
    like GPS navigation my battery life is absolutely consistent.

    I also have to mention that on the 8100 and 8310 it was still six additional apps
    in the background but fewer installed (memory limitation) but with the 9000 and
    8900 the apps have all been the same.

    Maybe I am just overthinking this issue.
    02-25-10 09:46 PM
  21. ragnarokx's Avatar
    I honestly think the whole better or worse battery life thing is in the head
    of the user and has no actual basis in fact.
    I can tell you that changing an OS does in fact alter battery performance, and the hybrid community (see CB's OS/Beta forum, bbhybrids.net, driphter.com) has been gathering empirical data on this subject for a long time. Thanks to Meterberry we are able to quantitatively measure our battery drainage based on percentage per hour. One of the first questions when a new hybrid comes out is "how's the battery life?" This is followed by numerous people posting their %/hr drain rate, and these numbers usually agree between the many users. With minimal usage, great battery life has drainage of <2%/hr, moderate is 2-4%/hr, and poor is >4%/hr. The biggest cause of change in battery life is usually changing OS java files, so your answer may lie in the way specific cod files are built and/or the interaction between cod files themselves.
    02-25-10 10:12 PM
  22. elvin1983's Avatar
    See Elvin that is the thing. I have six additional that are always resident and as
    I mentioned once before over 50 apps. Unless I am doing something data intensive
    like GPS navigation my battery life is absolutely consistent.

    I also have to mention that on the 8100 and 8310 it was still six additional apps
    in the background but fewer installed (memory limitation) but with the 9000 and
    8900 the apps have all been the same.

    Maybe I am just overthinking this issue.
    I don't think you're overthinking it, it's a valid question, and one that up until recently I never really thought much about, there's got to be some reason for the added memory drain!

    Even when I had an additional app running in the background, I never saw any additional battery drain, it did it's update, and then might as well have just shut down. I think if the apps that you have on your device are solid, then you shouldn't have any battery or memory issues.

    The thing that confuses me is that up until today, I had never heard about differences in radio programming based on the carrier. People have been loading operating systems released by carriers different from theirs for a long time, and I've never heard about this phenomenon prior to now, it's pretty strange.
    Also what gets me is that some people who are doing the same thing I am are experiencing better battery life than they did with 4.7!?

    Posted from my CrackBerry at wapforums.crackberry.com
    02-25-10 10:29 PM
  23. amazinglygraceless's Avatar
    I can tell you that changing an OS does in fact alter battery performance, and the hybrid community (see CB's OS/Beta forum, bbhybrids.net, driphter.com) has been gathering empirical data on this subject for a long time.
    I have read those threads and been to those forums and the data supplied
    is way short of anyones definition of empirical.

    Thanks to Meterberry we are able to quantitatively measure our battery drainage based on percentage per hour.
    Problem here is Meterberry (a total joke of an app) only tells you the
    amount of drain. what is DOES NOT DO is tell you the exact source. You
    still don't know if it is the OS or an app issue.

    One of the first questions when a new hybrid comes out is "how's the battery life?" This is followed by numerous people posting their %/hr drain rate, and these numbers usually agree between the many users. With minimal usage, great battery life has drainage of <2%/hr, moderate is 2-4%/hr, and poor is >4%/hr.
    ... and every answer in response to this question is completely subjective
    and in a great deal of cases, I suspect, from people who have no idea of
    what they are talking about.


    The biggest cause of change in battery life is usually changing OS java files, so your answer may lie in the way specific cod files are built and/or the interaction between cod files themselves.
    This falls apart by pointing out that on this site alone there are just as many
    people pointing divergent battery life issues with hybrids (all of which suck)
    as there are people reporting the same with an official release.

    If hybrids were the cure everyone would be using them and NO ONE would
    be reporting battery life issues with them. Show me a hybrid with 100%
    reporting of improved battery life.
    02-26-10 01:39 AM
  24. dchawk81's Avatar
    I get better battery life since uninstalling MeterBerry. Ironic, innit?

    As far as OS differences, I have no idea. I'm thinking some of it is perception. As long as people continue to use the terms "heavy", "moderate", and "light" to describe their usage, I won't put much stock in their opinion.
    02-26-10 02:02 AM
  25. DervMan's Avatar
    Seeing as nobody can provide you with the evidence you need, might I suggest you buy a few thousand BlackBerry devices, install several versions of the operating system and then use each one in a clearly documented way, to try to establish whatever proof you find suitable.

    Meanwhile, there is sufficient evidence to suggest to me that checking what other users report is sensible enough.

    It is good that we have that choice of course.

    Battery life is very much an individual thing as we all use our devices differently. Poo-pooing someone's volunteered information and dissing an application because you didn't notice a difference when you changed your operating system... Meh you don't have the evidence to do such a thing. Cuts both ways.

    Posted from my CrackBerry at wapforums.crackberry.com
    02-26-10 02:43 AM
67 123
LINK TO POST COPIED TO CLIPBOARD