Page 59 of 74 FirstFirst ... 949545556575859606162636469 ... LastLast
Results 1,161 to 1,180 of 1472
Like Tree4Likes
  1. #1161  
    I was having very high load averages as well. They seemed to be related to data use for me, something was waking up the phone a lot to do stuff. I don't need data much so I set up a mode switcher mode to turn it off when the screen is disabled. This has done wonders for battery.

    I suspect this issue is webos related, not kernel, as I was having it with uberkernel as well.
  2. #1162  
    I get the high load averages everywhere even though I don't turn on Sprint data, only Wi-Fi when I'm home.
  3. #1163  
    Quote Originally Posted by gollyzila View Post
    I get the high load averages everywhere even though I don't turn on Sprint data, only Wi-Fi when I'm home.
    I primarily use wifi, so i can't vouch for what effect 3G has on the load averages. Try turning it off and see what happens to your load averages.
  4. #1164  
    Quote Originally Posted by NickVTPre View Post
    I primarily use wifi, so i can't vouch for what effect 3G has on the load averages. Try turning it off and see what happens to your load averages.
    WiFi off: Load Averages ~30 ~20 ~11
  5. #1165  
    Quote Originally Posted by gollyzila View Post
    WiFi off: Load Averages ~30 ~20 ~11
    Maybe this is snake oil, but have you tried:
    Govnah -> profiles -> app menu -> disable profile -> reboot?
    Then when it boots select your profile, make sure 3G and wifi are off and see what they do.
  6. #1166  
    Quote Originally Posted by NickVTPre View Post
    Maybe this is snake oil, but have you tried:
    Govnah -> profiles -> app menu -> disable profile -> reboot?
    Then when it boots select your profile, make sure 3G and wifi are off and see what they do.
    Hey I think it worked! With no data on, LA: ~2 ~2 ~2. So what really happened in the background?

    UPDATE: With WiFi on, LA stay below 3. I'm very curious how did disabling the profile fixed it.
  7. #1167  
    Quote Originally Posted by gollyzila View Post
    Hey I think it worked! With no data on, LA: ~2 ~2 ~2. So what really happened in the background?

    UPDATE: With WiFi on, LA stay below 3. I'm very curious how did disabling the profile fixed it.
    I doubt it fixed it, it'll likely come back. Wifi off is the only thing that keeps my load averages down... The disable profile clears out any profile things which may have stuck when switching profiles in Govnah. that can cause unexpected hiccups sometimes. It did for me with compcache.
  8. #1168  
    You're right, after I switched to my custom profile and back to the screenstate profile my phone was having high LA.
  9. #1169  
    I see there's a new update for F105 Thunderchief. Does this possibly fix the high LA while screen is off?
  10. #1170  
    Quote Originally Posted by gollyzila View Post
    I see there's a new update for F105 Thunderchief. Does this possibly fix the high LA while screen is off?
    What does the Changelog say?


    M.
  11.    #1171  
    Quote Originally Posted by gollyzila View Post
    I see there's a new update for F105 Thunderchief. Does this possibly fix the high LA while screen is off?
    <flame>
    For the love of Gawd and all that is Holy!

    There is no problem.
    </flame>

    I posted about this in detail recently. Please refrain from repeatedly asking something if you do not want to do the effort in finding out why your phone has a high load average.
    Live free or DIE!
  12. #1172  
    Quote Originally Posted by unixpsycho View Post
    I have just spent a few hours on this testing with a Pre2 and a Pre-. Pre2 using SR71 and Pre- using F105, also a stock test run for a control.

    All kernels using Screenstate-v2 are using the same code base regardless of platform (exactly the same SSv2 module and override), except for F14 (SSv1), F102A/B (stripped SSv1/SSv2) and Uberkernel (SSv1).

    Pre-:
    SSv2 1005/500: screenoff 15 mins, back on. LA peak 7.
    Performance 500/500: screenoff 15 mins. back on. LA peak 12.
    Stock 500/500: screenoff 15 mins. back on. LA peak 8

    Pre2:
    SSv2 1200/300: screenoff 15 mins, back on. LA peak 14.
    Performance 300/300: screenoff 15 mins. back on. LA peak 22.
    Performance 600/600: screenoff 15 mins. back on. LA peak 10.
    Stock (ondemandtcl) 1000/300: screenoff 15 mins. back on. LA peak 7

    Test Results: Inconclusive

    It is useless to do these tests as the phone goes into suspend aggressively and gets woken up alot, which drives up load. Also not knowing what is running in the background during these times, it is impossible to get any definitive baseline/benchmark.

    Screenstate v2 has a little overhead due to polling charger status but not that much.
    As a data point my Pre + running F102B (Screenstate v2) routinely gets to a 12 or so LA in a few minutes or less while idle. The load rapidly vanishes when I wake it up. I am using this with Vdemand set to medium. This number doesn't seem to climb higher than this after even 30 mins of idle time. These are with WiFi and with WiFi set to sleep.

    Now that I see your results I'm more confident that my results are not out of the ordinary.
  13. #1173  
    Odd behavior from the Camera application...

    This past week I found that running F102B (1.4.5-172) with Vdemand (default 500/800 screenstate v2) I had trouble with the camera app when the phone went to sleep. The app seemed to hang frequently. If the app was on when the phone went to sleep the screen would wake up in the app but gray and blank (no picture no controls). I could flick it off and then restart and it would still be "dead". No other problems were noticed. A Luna restart fixed the issue every time and sometimes if I ran jstop it and forced a GC it would allow the app to work again.

    When I turned Vdemand profiles off this did not happen. Normally my Pre is very stable with Vdemand and I have been able to run with Vdemand set high reliably with all other applications. This was only set to Medium and still gave me trouble. I did not try Low yet but I will test further.

    Full device restarts did not change this behavior.

    I have not doctored.

    Is anyone else able to replicate this issue?
  14.    #1174  
    Quote Originally Posted by Unclevanya View Post
    Odd behavior from the Camera application...

    This past week I found that running F102B (1.4.5-172) with Vdemand (default 500/800 screenstate v2) I had trouble with the camera app when the phone went to sleep. The app seemed to hang frequently. If the app was on when the phone went to sleep the screen would wake up in the app but gray and blank (no picture no controls). I could flick it off and then restart and it would still be "dead". No other problems were noticed. A Luna restart fixed the issue every time and sometimes if I ran jstop it and forced a GC it would allow the app to work again.

    When I turned Vdemand profiles off this did not happen. Normally my Pre is very stable with Vdemand and I have been able to run with Vdemand set high reliably with all other applications. This was only set to Medium and still gave me trouble. I did not try Low yet but I will test further.

    Full device restarts did not change this behavior.

    I have not doctored.

    Is anyone else able to replicate this issue?
    hmm. I havent looked at the F102 code for 1.4.5 in a while. i'm sure i made some fixes in the 2.1.0 branch for it that never made it in the 1.4.5 branch. i'll take a look.
    Live free or DIE!
  15. #1175  
    Quote Originally Posted by unixpsycho View Post
    hmm. I havent looked at the F102 code for 1.4.5 in a while. i'm sure i made some fixes in the 2.1.0 branch for it that never made it in the 1.4.5 branch. i'll take a look.
    I noticed you released a new version 1.4.5-174 which I installed last night. I have a few questions:

    1) VDEMAND seems to no longer have an option for how aggressive you want to set it - it's TRUE or FALSE with no more selection of LOW/MED/HIGH/OMG - is this intended/correct?

    2) When I updated from 1.4.5-172 I did not first downgrade to the recovery kernel - my assuption was that this was OK for an update - was this incorrect?

    3) My default F102B profile was active at the time of the upgrade. The VDEMAND Poll rate was set for 100 before the upgrade. After upgrading I noticed this was set to "10". I realized "10" wasn't even in the choices available so I tried disabling the profile and then immediately restarting the phone but that didn't change anything. Finally I gave up and switched to 100 manually. Was this the right set of steps to take?

    4) While off charger I did see some creep up in the Load Average - but nothing unusual compare with the results posted earlier in the thread. However while on the charger overnight I noticed that my LA was virtually zero when I resumed - is this expected? I realize that the charger sets the speed to 500 but I'm not clear if the VDEMAND profile is active during charger operation and not clear if the phone "sleeps" while charging or not. Ignorance on my part - but I'm hoping someone can fill me in on this so I understand it better.

    I haven't had a chance to play with the camera yet - the problem is a bit intermittent so I'll have to tinker for a couple of days before I'm sure it's gone if it doesn't happen right away.
  16.    #1176  
    Quote Originally Posted by Unclevanya View Post
    I noticed you released a new version 1.4.5-174 which I installed last night. I have a few questions:

    1) VDEMAND seems to no longer have an option for how aggressive you want to set it - it's TRUE or FALSE with no more selection of LOW/MED/HIGH/OMG - is this intended/correct?

    2) When I updated from 1.4.5-172 I did not first downgrade to the recovery kernel - my assuption was that this was OK for an update - was this incorrect?

    3) My default F102B profile was active at the time of the upgrade. The VDEMAND Poll rate was set for 100 before the upgrade. After upgrading I noticed this was set to "10". I realized "10" wasn't even in the choices available so I tried disabling the profile and then immediately restarting the phone but that didn't change anything. Finally I gave up and switched to 100 manually. Was this the right set of steps to take?

    4) While off charger I did see some creep up in the Load Average - but nothing unusual compare with the results posted earlier in the thread. However while on the charger overnight I noticed that my LA was virtually zero when I resumed - is this expected? I realize that the charger sets the speed to 500 but I'm not clear if the VDEMAND profile is active during charger operation and not clear if the phone "sleeps" while charging or not. Ignorance on my part - but I'm hoping someone can fill me in on this so I understand it better.

    I haven't had a chance to play with the camera yet - the problem is a bit intermittent so I'll have to tinker for a couple of days before I'm sure it's gone if it doesn't happen right away.
    i screwed it up. i'll have a new build out today. I dont have a dual-booted Pre- to test it thoroughly. i wanted to remove the vdemand poll since it shouldnt be changed too high, if it is then a RSOD will happen. Some people like to play with settings, then screw it up, then complain.

    vdemand still operates during charger scaleback.
    Live free or DIE!
  17. #1177  
    Just want to pop in and say thanks for all the hard work. I'm running F104A-35 on a PreMinus with a resting temperature between 18-27c with an intensive workload temp of 30-35. Very impressed. This is the best my Pre has run in ages.
  18. #1178  
    +1 F104 on my pre- running 2.1 runs amazing. nice work.
  19. #1179  
    Quote Originally Posted by unixpsycho View Post
    i screwed it up. i'll have a new build out today. I dont have a dual-booted Pre- to test it thoroughly. i wanted to remove the vdemand poll since it shouldnt be changed too high, if it is then a RSOD will happen. Some people like to play with settings, then screw it up, then complain.

    vdemand still operates during charger scaleback.
    OK - I'll look for the update. Should I try reverting? Is there a way to get back to older versions of the testing kernels?

    BTW - Thanks for all of the work you do.
  20.    #1180  
    Quote Originally Posted by Unclevanya View Post
    OK - I'll look for the update. Should I try reverting? Is there a way to get back to older versions of the testing kernels?

    BTW - Thanks for all of the work you do.
    It should be in the feeds now.

    You can go back to older versions but it is a manual process.
    http://ipkg.preware.org/feeds/webos-...testing/1.4.5/
    Live free or DIE!

Tags for this Thread

Posting Permissions