Palemoon becoming more sluggish as the day progesses? Topic is solved

[Harvest Moon]

Hi all!

First off...thank you browser developers for the latest Pale Moon update! (28.3.0!)
From the detailed list of problems it solves, shows that Moonchild and everyone else has been hared at work fixing things to make life better using Pale Moon Browser!

DOn't know if this helps in the current sluggishness problem (just had it running the new update a few hours...no sluggishness yet....Will see what happens when others check their systems after updating them!

Webroot my new "Webroot Secure Anywhere " AV software says:
"An active process has a possible memory leak (palemoon.exe)"



Does that information help anyone tracking down the sluggishness problem?

(in case its just MY computer...I also have one other "problem" identified today, by Webroot AV:

"An active process has a possible handle leak (explorer.exe)"

Just sharing in case that has anything to do with the slowdown we experience...Does anyone else have a AV or program reporting these too?

- - - - -

.....and trying my hardest to keep my posts shorter....as another problem seems to be going around PMB Forum lately:

My problem is that I keep receiving lengthy and convoluted forum posts via Palemoon, could it be the fault of the browser?

Exactly the same here
Let us blame the browser

Harvest Moon

[PalleP]

v28.3: Fixed a potential cycle collector resource leak.

Has this got anything to do with the bug in this tread?

[Ozzed]

v28.3: Fixed a potential cycle collector resource leak.

Has this got anything to do with the bug in this tread?

I haven't dared to try Gmail yet, but it should be fairly easy to try and see if the issue persists.

[PalleP]

I haven't dared to try Gmail yet, but it should be fairly easy to try and see if the issue persists.

It does, although I haven't tried with Gmail.

[win7-7]

x86 OS still seems to have problems with UXP platform. I have managed to reproduce intermittent slowdowns in tabs switching and hangs on Windows 7 x86 virtual machine with 36 or so tabs with Pale Moon 28.3. I increased memory for this testing in VM to 2.5GB, there was memory left but when switching between tabs slowdown and hangs happened.

Also tested changing javascript.options.mem.gc_generational to false but it didn't make noticeable difference.

I also have one different hang stack attached (that is for PM 28.0.1 however as I had that at virtual machine from last time before I updated version afterwards and hangs hang stack from 28.3 are similar to last time: Browser:ForwardOrForwardDuplicate, https://aus5.mozilla.org/update/3/GMP/6 ... L%/%OS_VER and Gecko_IOThread)

Firefox ESR52.6 also has this issue with e10s force disabled. Didn't test 52.9. Firefox ESR 45 does not have this issue with e10s force disabled.

Mozregression:

Last good (mozilla-central):

19.7.2016 (50.0a1)

First bad (mozilla-central):

20.7.2016 (50.0a1)

This issue is intermittent which makes reproducing harder and I had problem to get this happen in time to notice which version regression first happened and reproducing this issue constantly is not easy.

Based to testing in thinderbox builds (mozilla-inbound) problem seems to be caused by Garbage Collection changes around 50.0a1 (20.7.2016 in Mozilla-central).

Build 1468947261 does not seems to have issue. (https://ftp.mozilla.org/pub/firefox/tinderbox-builds/mozilla-inbound-win32/1468947261/)

build 1468950678 does seems to have noticeable slowdown and hangs when switching between tabs. Closing tabs can also cause hang . (https://ftp.mozilla.org/pub/firefox/tinderbox-builds/mozilla-inbound-win32/1468950678/)

These two changes were done in build 1468950678:

Bug 1287399 - Allow the pre-barrier verifier to work in the presence of OMT parsing; r=jonco

Bug 1263771 - Sweep the InnerViewTable with WeakCache; r=sfink

[Rickkins]

Maybe it's just me, but it's even worse now with the latest release...

[PalleP]

Maybe it's just me, but it's even worse now with the latest release...

I have the same feeling, although it may just be because this bug gets more and more annoying. One thing is for sure, there is absolutely no progress in the new PM releases.
Try this: Set dom.cycle_collector.incremental to false, it helps but is not a cure...

[Rickkins]

Maybe it's just me, but it's even worse now with the latest release...

I have the same feeling, although it may just be because this bug gets more and more annoying. One thing is for sure, there is absolutely no progress in the new PM releases.
Try this: Set dom.cycle_collector.incremental to false, it helps but is not a cure...

Did that, does nothing.

[Sob__]

It's about the same for me. But it's hard to say, because if it depends on specific sites, it's almost impossible to pinpoint it for me, as I visit a lot of different ones. And it's not like the moment I visit the "right" one, there's a bang and PM is suddenly slow, it's adding up. All I know for sure is that after few hours of actively using PM, sluggishness is almost guaranteed to happen.

When I check palemoon.exe in Process Explorer, it's always the same thread that eats up CPU (start address palemoon.exe+0x2b58), but it's probably not useful info, because it's the same thread that's most active all the time, so probably some main one.

What I'm also noticing, when it happens, there's significant memory usage too. I'm seeing PM using 1.5-2.5GB, even after I close or unload (using Lull The Tabs extension) all tabs and just keep one about:blank open. I wouldn't mind, I have 16GB and half of it is free most of the time, but it seems it's related to my "something stuck in background" theory.

I'm trying to find a reliable way how to reproduce it, but no luck so far.

[cf099]

I am also encountering this problem with Palemoon 32-bit running on Windows 7 x64. I have avoided posting because the blame would be pointed to the use of NoScript.

Here are 3 suggestions which might, or might not, be relevant:

• I have always blocked quite a few cookies for specific sites via "Use custom settings for history". I've noticed that the way that this information is stored in the permissions.sqlite database has changed - before v.28 the entries would be, for example, "google.com". The entries are now stored with "http:" and "https:" stuck on the front, but omits the "www" etc. There must be an overhead processing this. I suspect that some of the privacy add-ons which manage cookies might also get caught by this.

• Is the problem caused by the JavaScript routines provided by "polyfill.io"? Palemoon releases before version 28 didn't support this functionality, so this might explain why the problem has suddenly appeared.

• One of the sites I visit frequently takes a long time to load information from a couple of linked sites - the status bar shows "Transferring data from ..." or "Connecting to ..." for some time. Is this a contributory factor?

If there is some testing or analysis I can do to help track down the cause(s) of the problem, then please let me know.

PS: I do not use Gmail, so that can be ruled out.

[Rickkins]

At this point, I have to restart the browser at least once an hour if I'm actually using it.

[Moonchild]

And I happily keep using the browser with uptimes of a week or more, without slowdowns.

If there was actually something for me to notice or test, then I could potentially improve it -- but as it is Pale Moon is performing as it always has and no slowdowns are seen on my end, at all.

How old is your profile? has it been imported from Firefox or have you used it on a mix of v27 and v28 versions (i.e. have you ever upgraded to v28, then back to v27 with the same profile?)
Maybe you want to try a profile refresh from the Help -> Troubleshooting page to solve any issues.

[PalleP]

How old is your profile? has it been imported from Firefox or have you used it on a mix of v27 and v28 versions (i.e. have you ever upgraded to v28, then back to v27 with the same profile?)
Maybe you want to try a profile refresh from the Help -> Troubleshooting page to solve any issues.

This bug is also present on a fresh portable PM, as stated several times before.

[RJARRRPCGP]

the status bar shows "Transferring data from ..." or "Connecting to ..." for some time.

It may be an internet hardware problem on your end. I would test at speedtest.net, if I were you.

[Moonchild]

This bug is also present on a fresh portable PM, as stated several times before.

Right, well I don't have any other suggestions.

But my reply as more in response to Rickkins,

[Rickkins]

And I happily keep using the browser with uptimes of a week or more, without slowdowns.

If there was actually something for me to notice or test, then I could potentially improve it -- but as it is Pale Moon is performing as it always has and no slowdowns are seen on my end, at all.

How old is your profile? has it been imported from Firefox or have you used it on a mix of v27 and v28 versions (i.e. have you ever upgraded to v28, then back to v27 with the same profile?)
Maybe you want to try a profile refresh from the Help -> Troubleshooting page to solve any issues.

Very recently I did this, with the aim of refreshing my profile....

1) Backup your bookmarks, if you need them.
2) Uninstall plugins (Flash, QuickTime or whatever).
3) Uninstall old Pale Moon version.
4) Reboot.
5) Delete these folders, if they exist (or their equivalents on other Windows versions):
C:\Program Files\Pale Moon
C:\Users\*****\AppData\Local\Moonchild Productions\Pale Moon
C:\Users\*****\AppData\Roaming\Moonchild Productions\Pale Moon
6) Delete these reg. keys (if they exist):
HKEY_LOCAL_MACHINE\SOFTWARE\Mozilla\Pale Moon
HKEY_LOCAL_MACHINE\SOFTWARE\Mozilla\Pale Moon **.*.*
7) Install new Pale Moon version (use a full installer, NOT the web installer).
8) Start Pale Moon once, then close it (for it to register properly).
9) Reinstall the plugins that you need/want.
10) Reboot.
-----------------------------------------------------------------------

System specs:

Operating System Windows 10 Enterprise LTSC 64-bit
CPU AMD Ryzen 7
RAM 16.0GB Dual-Channel Unknown @ 1066MHz (15-15-15-36)
Motherboard Gigabyte Technology Co. Ltd. X470 AORUS ULTRA GAMING-CF (AM4)
Graphics
SAMSUNG (1920x1080@59Hz)
PX2710MW (1920x1080@60Hz)
4096MB ATI Radeon RX 570 Series (Gigabyte) 46 °C
Storage
465GB Samsung SSD 960 EVO 500GB (Unknown (SSD)) <- Palemoon installed here
1863GB Seagate ST2000DL003-9VT166 (SATA ) 35 °C
1863GB Western Digital WDC WD2003FZEX-00Z4SA0 (SATA ) 30 °C
5589GB Western Digital WDC WD60EFRX-68L0BN1 (SATA ) 30 °C
1863GB Seagate ST2000DL003-9VT166 (SATA ) 44 °C
9314GB Western Digital WDC WD101KRYZ-01JPDB1 (SATA ) 43 °C
5589GB Seagate Expansion Desk SCSI Disk Device (USB (SATA) ) 29 °C
Optical Drives
HL-DT-ST BD-RE WH10LS30
Audio

[Hmm81]

Just tried 28.3.1, and after opening up gmail, having the ghost window, it seemed to work for several hours, but after that, the same issue keeps cropping up, where it just feels sluggish, and uses huge cpu resources for no reason.

[Sajadi]

Can confirm this with an alternative computer - equipped with
Intel 7 series c216 - Windows 8.1 64 Bit - 64 Bit Pale Moon version

I would say the issue is clearly in relation between certain kind of hardware and Pale Moon, no matter if 32 or 64 Bit and no normal bug - it really looks like that Pale Moon 28 works with certain kind of hardware way better as with some other combination. For example - on my main Laptop - an Asus Zenbook with an Intel I7 and Windows 8 - i am not experiencing this issue at all. The question is, can anything be done that Pale Moon is able to run smoother on "non-supported hardware variants" - or is there a way to lessen the negative impact?

Would be interesting what others who are suffering in a big way from this problem have for a kind of hardware - Intel/AMD

[Moonchild]

I'm trying to improve the cycle collector and garbage collector a bit, but it's difficult to know if it has any effect and if so if it's an improvement if I can't reproduce the problem.

[New Tobin Paradigm]

I wonder, everyone who has this issue please indicate if you are using an Intel or AMD processor.