[WINDOWS] Performance degradation over time?

[WINDOWS] Performance degradation over time? - Printable Version

+- Kodi Community Forum (https://forum.kodi.tv)
+-- Forum: Support (https://forum.kodi.tv/forumdisplay.php?fid=33)
+--- Forum: General Support (https://forum.kodi.tv/forumdisplay.php?fid=111)
+---- Forum: Windows (https://forum.kodi.tv/forumdisplay.php?fid=59)
+---- Thread: [WINDOWS] Performance degradation over time? (/showthread.php?tid=87804)

Pages: 1 2

[WINDOWS] Performance degradation over time? - Viald - 2010-12-17

I noticed that my system performance is degraded over time.
I never reboot my system, each time I put it in suspend mode for quick start.

For example, after a reboot, playing 1080p video file uses 50 to 60% of CPU. After few days playing music, videos, showing pictures and multiple go back from suspend state. Playing the same video uses 90% to 100% CPU Shocked

which of course causes dropped frames.

After a reboot all comes back as before.
How can I diagnose this problem ?

Thank you.

- Tobas - 2010-12-17

Switch to linux

This problem is common with Windows

- Viald - 2010-12-17

Tobas Wrote:Switch to linux

This problem is common with Windows

I'm not sure it's a Windows issue Big Grin

Next time I have degraded performance, I'll only restart XBMC to compare.

- a_user - 2010-12-17

Viald Wrote:I'm not sure it's a Windows issue

it is. in fact even without suspend running windows for too long time results in slow performance regression. oh, and to make it even worse, even rebooting from time to time won't stop the regression. it will only slow down the speed of regression, except you don't change/install/uninstall software.

the 3 types of regression have different reasons btw. Wink

- Mindzai - 2010-12-17

I used to use Windows and noticed the same thing. I now use Linux and never see it. For example, my linux server currently has an uptime of 86 days and it is running as fast as ever.

- SpectreX - 2010-12-18

Next time this happens, download MPC-HC : http://mpc-hc.sourceforge.net/ and try to playback the same files with it and see how it handles them.

- bodhi - 2010-12-18

a_user Wrote:it is. in fact even without suspend running windows for too long time results in slow performance regression. oh, and to make it even worse, even rebooting from time to time won't stop the regression. it will only slow down the speed of regression, except you don't change/install/uninstall software.

the 3 types of regression have different reasons btw.

Would you mind telling us what are the reasons?

- SpectreX - 2010-12-18

Ignore the Linux trolls, my HTPC is up and running for 4 days and works like a charm. Same goes for my desktop, uptime about 2 weeks, no performance degradation or anything. Let`s not get into the statistics on netcraft, Windows Sever 2003/2008 machines running for years as an uptime. On XP there was some performance degradation due to the obsolete memory management, but he has Windows 7, not XP.

- cold_realms - 2010-12-18

There are several threads and a bug already in tracker about this issue. There has been some work tracking this down to a timer going crazy after being in sleep for a long period (over a few hours). Just do a search for sleep and you will find them.

For the trolls: windows does not have lingering performance issues, not since the days of win95 (even then I had none). I have 6 machines in the house , 5 win7 and 1 server 2008 R2. The only time these machines ever get a reboot is on patch day IF a reboot is required. Zero slowdowns in anything (except the sleep bug that is in XBMC ONLY and all it takes to clear that is restart XBMC).

- bodhi - 2010-12-18

cold_realms Wrote:There are several threads and a bug already in tracker about this issue. There has been some work tracking this down to a timer going crazy after being in sleep for a long period (over a few hours). Just do a search for sleep and you will find them.

For the trolls: windows does not have lingering performance issues, not since the days of win95 (even then I had none). I have 6 machines in the house , 5 win7 and 1 server 2008 R2. The only time these machines ever get a reboot is on patch day IF a reboot is required. Zero slowdowns in anything (except the sleep bug that is in XBMC ONLY and all it takes to clear that is restart XBMC).

I'm just curious to know whether the statement about "3 types of regressions" has any basis? (since I did not have any performance issue, and I always put the HTPC to sleep, hardly ever shut it down). I noticed that XBMC sometime has problem responding when I exit it to desktop, and as you said, restart XBMC it would clear all problem. I think perhaps because I'm running Dharma beta4. Hope the stable version will be better.

- SpectreX - 2010-12-18

bodhi Wrote:I'm just curious to know whether the statement about "3 types of regressions" has any basis? (since I did not have any performance issue, and I always put the HTPC to sleep, hardly ever shut it down). I noticed that XBMC sometime has problem responding when I exit it to desktop, and as you said, restart XBMC it would clear all problem. I think perhaps because I'm running Dharma beta4. Hope the stable version will be better.

I had similar issues with Beta 3. Upgraded to RC2 and has been working like a charm. Do yourself a favor and upgrade to RC2, it`s been rock solid for 2 weeks.
As for basis, as i explained above, on XP and below operating sistems, it had "some" basis due to the ineficient memory managament. But trolls usually forget to update their responses Big Grin

- bodhi - 2010-12-18

SpectreX Wrote:I had similar issues with Beta 3. Upgraded to RC2 and has been working like a charm. Do yourself a favor and upgrade to RC2, it`s been rock solid for 2 weeks.
As for basis, as i explained above, on XP and below operating sistems, it had "some" basis due to the ineficient memory managament. But trolls usually forget to update their responses

Thanks for the RC2 recommendation Big Grin

will do that shortly.

- BlueScreenJunky - 2010-12-18

a_user Wrote:it is. in fact even without suspend running windows for too long time results in slow performance regression. oh, and to make it even worse, even rebooting from time to time won't stop the regression. it will only slow down the speed of regression, except you don't change/install/uninstall software.

the 3 types of regression have different reasons btw.

Not sure if you're joking, but in case you're not, I'd like to point out that the last time I've reinstalled Windows 7 x64 was when it was released (got a great price by pre-ordering btw ^^), and I've never experienced any kind of performance regression with any application, even though I only reboot once a month or something like that. And it's not like I'm not installing any applications either (although I do uninstall them carefully when I know I won't use them anymore).

So I'd highly recommend that you run an anti-malware / anti-virus because there's obviously something wrong with your system.

- Viald - 2010-12-18

Thank you all for your feedback. The goal of my question was not to wake up conflicts between linux and windows Wink

I'm sure that it is not windows issue, simply because, before XBMC, I used Mediaportal without performance degradation.

SpectreX Wrote:Next time this happens, download MPC-HC : http://mpc-hc.sourceforge.net/ and try to playback the same files with it and see how it handles them.

Very good idea. I'll try.

cold_realms Wrote:For the trolls: windows does not have lingering performance issues, not since the days of win95 (even then I had none). I have 6 machines in the house , 5 win7 and 1 server 2008 R2. The only time these machines ever get a reboot is on patch day IF a reboot is required. Zero slowdowns in anything (except the sleep bug that is in XBMC ONLY and all it takes to clear that is restart XBMC).

Even, I like linux, I absolutely agree with you.

bodhi Wrote:I noticed that XBMC sometime has problem responding when I exit it to desktop, and as you said, restart XBMC it would clear all problem. I think perhaps because I'm running Dharma beta4. Hope the stable version will be better.

I'm have the same issue with RC2

SpectreX Wrote:I had similar issues with Beta 3. Upgraded to RC2 and has been working like a charm. Do yourself a favor and upgrade to RC2, it`s been rock solid for 2 weeks.
As for basis, as i explained above, on XP and below operating sistems, it had "some" basis due to the ineficient memory managament. But trolls usually forget to update their responses

My degradation over time is with the currently RC2.

- Viald - 2010-12-18

cold_realms Wrote:There are several threads and a bug already in tracker about this issue. There has been some work tracking this down to a timer going crazy after being in sleep for a long period (over a few hours). Just do a search for sleep and you will find them.

For the trolls: windows does not have lingering performance issues, not since the days of win95 (even then I had none). I have 6 machines in the house , 5 win7 and 1 server 2008 R2. The only time these machines ever get a reboot is on patch day IF a reboot is required. Zero slowdowns in anything (except the sleep bug that is in XBMC ONLY and all it takes to clear that is restart XBMC).

I found nothing very interesting with 'sleep', do you have a link ?
Thank you.