CPU core use graph does not show, windows 64 pro [FIXED]

[MickeyG]

I am not seeing the check mark and the graph does not show up

[BenYeeHua]

It can be Operation System(OS) issues.
Can you tell me which language of the os?

PS:Moved to BugTrack

[Jeremy Collake]

Have you run any system tweaking or optimization software recently? Any changes to your service configuration?

[Hotrod]

This was the closest existing thread to the item I am seeing. Move my post if you wish. On my single core XP machines the per core usage graph is not shown. This is precisely as it should be, however, in the view menu the option still exists but does nothing when chosen. I propose that this option should not even appear as a choice or alternately be greyed out on single core machines. This would prevent people from reporting the feature as broken when it is actually performing as designed.

[smartbit]

Hi,

I've noticed on the BitSum homepage that there is a per-core CPU Usage/Park graph shown in between the usual performance graph and the RAM graph (http://bitsum.com/images/screenshots/1.png).

But for somewhat reason, I can't find a way to enable it.
I tried the "View>Show per-core CPU utilization", I'm not sure if this is the setting, but it never gets toggled anyway (probably a bug).

So, how can I enable this nice thing?

Thanks for help
(and sorry if it is already has been discussed, I tried to search the forums unsuccessfully..)

[BenYeeHua]

I think there are nothing I can help you here...
It look like another bug for can't getting the information from system.

Can you tell which language and version the Windows that you are using?
And did you try disable anti-virus to check it up?

[smartbit]

Hi,
I tried disabling security software (installed: Avast A/V, Comodo FW + HIPS), and relaunch the Process Lasso as an Administrator - still, the "show per-core..." won't toggle.

Process Lasso Pro x64 v.6.7.0.0
System Info - Lenovo ThinkPad 430s:
Windows 7 Ultimate x64 English (US), CPU: i7 3520M (2 cores/4 threads), 16GB RAM, SSD & HDD storage.
Tell me if you need more info on SW/HW.

Should I submit it as a bug somewhere, or posting here is enough?

[edkiefer]

Do you have aero on in desktop or you use classic  mode ?
Classic has issues with CPU % graph .

[smartbit]

Aero enabled...

[Jeremy Collake]

For a very small minority of systems, the performance counters used as the source for this information are not available, or not properly operating. I still have not determined why. I need to relaunch a thorough investigation, but since this affects so few people, it's not been a huge priority.

My best guess so far is that changes to the system or service configuration are the culprit, but I have no further guidance at this time.

[smartbit]

Hmmm..

Well, sorry to hear that.

Regarding the configuration issues, I did not disabled or messed much with system services. Though, the PC is a ThinkPad laptop with "Enhanced Experience 3" preinstalled - which means, Lenovo did some registry/config. tweaks to boost the performance (or/and to optimize it for the hardware). Maybe it has something to do with it.

It is weird though, because all of Process Monitors I used so far had no trouble detecting core utilization or parked cores.
But, I guess they are mostly based on the build in windows components for measurements reporting, maybe you do stuff differently.

Thanks for the reply anyway

[Jeremy Collake]

All the utilities, including Process Lasso, should get their data from the same source.

My assumption on the cause could be wrong. It definitely bothers me to go around saying 'yea, it's broken sometimes', so I will surely address this at some point.

One curiosity is that another user who had this problem reported that the same display in Bitsum's ParkControl works fine (Options / Configure CPU Core Parking). The same code module is used for both ParkControl and Process Lasso's core utilization display. This caused me to think maybe the application compatibility assistant is going awry, but that didn't seem to be the case. I still consider that a possibility though, and you might want to right click on 'ProcessLasso.exe' and check the 'compatibility' tab, just to make sure.

I will take a new look at this ASAP, perhaps a fresh perspective will yield some results.

[smartbit]

No compatibility options are enabled, ProcessLasso.exe is running in a usual manner, so nothing there.

Regarding the core parking control module:
It is working, tested it several times in a disable/enable parking manner - and the Windows Resource Monitor reported park status according to expectations.

But, nevertheless, the Utilization display won't either show itself or toggle in Menu>View>Show Per-Core Utilization.

[BenYeeHua]

For the compatibility, I has a story about it, but I think it might not helps to find out why.

http://vvvv.org/forum/windows-8-fullscreen-issue (http://vvvv.org/forum/windows-8-fullscreen-issue)
http://msdn.microsoft.com/en-us/library/windows/desktop/hh994464%28v=vs.85%29.aspx (http://msdn.microsoft.com/en-us/library/windows/desktop/hh994464%28v=vs.85%29.aspx)

As you might know, for DirectX 9 games, DWM that using after Windows Vista can be a performance killer, and a bad coding for DirectX 9E games like WoT, a windowed or sync DWM present with full screen is a big performance killer for WoT(50% GPU usage!).
(SHIM_RENDERING_OPTIONS_ENABLE_DWM_ASYNC_PRESENT if you are using Nvidia, you will see this for WoT profile)

I meet a guy that having this issues, when he is playing full screen for a game, the FPS are normal, but if he start using the IME and that game don't support IME, windows 8 will switch to windowed borderless mode for the game to support showing the IME forever(means the FPS will be reduced forever), until he delete the value of HKEY_CURRENT_USER\Software\Microsoft\Direct3D\Shims\MaximizedWindowedMode .
PS:search Application fails due to graphics and display issues on the MSDN link for more information about this.

And there are no ways to know this happen, as it only showing the windows for 1 time, you will easy to forgot it, I wonder that did Windows 7 that having this compatibility system will affecting PL getting the information too.

[geloxo]

I have the same issue with Win7 64bit. Everything was working fine till a couple of weeks ago. I don´t know what update came to my system (if Windows update, java, nvidia,..) but I can´t see now the graph and the option in the menu to show it is not responding 

[BenYeeHua]

En...
I think you can checking which thing has been installed by using the "Control Panel\All Control Panel Items\Programs and Features", click on the "Installed On" to arrange it , you can also check the installed update by clicking on "View installed updates".

For now, we are really don't has an idea for what's going on.

[geloxo]

Yeap, I know how to see the installed things but no idea to know which update caused lasso graphic to stop working... The most recent updates which are not usual from my previous installations where lasso worked fine are:
- Adobe flash player 11.9.900.117
- Net Framework 4.5.50709
- HWinfo64 4.26
- Java 7.0.450
- NVIDIA 331.65

Hope it helps..

[Hotrod]

I have an Inspiron 1540 laptop with Win7X64 and my graph works as it should including the menu option switching.

I have the adobe flash player 11.9.900.117
I only have .NET up to 4.0.30319
I have Java 7.0.400
I do not have the others installed.

Hopefully this will help you track the issue.

[BenYeeHua]

En...

I also having the newest .net 4.5.1, HWiNFO64, Nvidia graphic driver installed, but I don't facing this issues.

It can be registry, .net or the graphic driver causing it, and somehow this don't happen with Windows 8.

[Jeremy Collake]

This has moved back to the top of my agenda, after having seen it out in the field myself. I hope to get to the bottom of it soon.

[BenYeeHua]

Welcome back.

And I think you might asking them for Remote control, like using Team Viewer, then test on their side?
I has seen one of the guy at Guru3d doing this, and fix many issues for the driver.

[geloxo]

This has moved back to the top of my agenda, after having seen it out in the field myself. I hope to get to the bottom of it soon.

Great. Many thanks. I hope it´s solved easily 

[Jeremy Collake]

Welcome back.

And I think you might asking them for Remote control, like using Team Viewer, then test on their side?
I has seen one of the guy at Guru3d doing this, and fix many issues for the driver.

Yes, if all else fails, I may need to Remote into a machine - we'll see.

And I was out of town for a week, work related - apparently Bitsum supporters aren't used to this, I think some people thought I died .

[BenYeeHua]

And G+ telling me that you are still posting thing.

PS:smartbit need you there. https://bitsum.com/forum/index.php/topic,3596.0.html

[smartbit]

This has moved back to the top of my agenda, after having seen it out in the field myself. I hope to get to the bottom of it soon.

Nice to hear this issue coming up to the To-Do lists  :-)
As it turns out, I wasn't the only one with this issue. (Glad for me, sorry for others...)

(though, for now I have another nasty issue going on as BenYeeHua mentioned...)

[Jeremy Collake]

I believe I finally determined the probable cause of the failure to display core-utilization graphs for both Process Lasso and ParkControl.

Unfortunately, I noticed it right after the upload of v6.7.0.14, during some code review. I have therefore uploaded a silent update to v6.7.0.14. In the About Box, it will say 'Engine Version: 810600' . If you already downloaded v6.7.0.14, you'll need to run the installer again (or wait a few days at least).

I can't say for sure that this addresses affected cases, but I can say it fixes a similar cause of a failure to display these per-core utilization and parking graphs.

If this is an effective fix, I'll push it out to the users who managed to update to the prior release build.

[smartbit]

Hi,
Thanks for the update

I've just updated PL via the Auto update, and turns out that the version is the right one (engine version shows 810600).

But, unfortunately, no changes on the Core Utilization Graph subject.
The behavior is still exactly the same - the menu item "View -> Show per-core CPU utilization" is frozen, shows as checked, but still no graph displayed.
I tried to disable/enable by editing the prolasso.ini, and relaunching PL:
The UI menu state responds (the checkmark) to the ini modification, but still no graph shown, no mater whether the flag "ShowPerCoreUtilGraph" is true or false in ini file.

Bottom line - the fix is not effective.
The problem cause is still somewhere inside...

[Jeremy Collake]

Ok, thanks for letting me know. The fix did at least improve the behavior on non-English systems, so it wasn't entirely in vain.

I've built a new framework for debugging failures in that code, so the next beta will give us some clue as to what's going on.

[Jeremy Collake]

For v6.7.0.15+ , there is debug output we can use to diagnose this. The 15 build only has debug output applicable to this issue. 17 has a lot of governor debug output that isn't applicable. So, either one.

See https://bitsum.com/forum/index.php/topic,3642.0.html for information on how to collect and submit this debug output using DebugView.

Thanks for your help! Hopefully this will yield the missing clue, or will certainly put us one step closer.

[edkiefer]

will 17b be slower because of debugging or its not much of a hit ?

If so is there easy way to disable it, if not testing/debugging .

[Jeremy Collake]

Not any big difference in performance. If you don't have DebugView running, it pretty much is disabled. The *only* reason there is a *very negligible* decrease in efficiency is because the debug output results in less efficient compiled code, a little extra computational work, and a bit more memory for the strings. We're talking levels that nobody would ever notice.

(to the point the effective answer is 'essentially NO difference in performance')

[smartbit]

Thanks, I'll collect the data once build 15 is released.

[Jeremy Collake]

Thanks! 15 was released yesterday. 17 today, just now.

[edkiefer]

For v6.7.0.15+ , there is debug output we can use to diagnose this. The 15 build only has debug output applicable to this issue. 17 has a lot of governor debug output that isn't applicable. So, either one.

See https://bitsum.com/forum/index.php/topic,3642.0.html for information on how to collect and submit this debug output using DebugView.

Thanks for your help! Hopefully this will yield the missing clue, or will certainly put us one step closer.

Ok , I read to fast so your using debugview.exe (SysInternals) to catch data .

[BenYeeHua]

And, any knowledge that I can learn from that tools?

[edkiefer]

http://technet.microsoft.com/en-us/sysinternals/bb896647.aspx

Don't know if there are any doc's for it , I have used it few times in past for gathering data (mainly crashes ,if memory is right) .

there is help file with it .

[BenYeeHua]

Yup, I guess it must having many video/blog about it, as it is one of the tools from Mark Russinovich.
I will find for it, a little busy for a few days, thank for the link.

[edkiefer]

yes, it kind of depends on the app your using it on, as what the dev have left open to debug .

Here link I found, but again it might not be apply to PL .Some tips, FAQ

http://www.sophos.com/en-us/support/knowledgebase/119577.aspx

The help file has some info too as mentioned before, its more than normal .

But the instructions Jeremy posted is pretty simple, straight forward .
https://bitsum.com/forum/index.php/topic,3642.0.html

[BenYeeHua]

Ok, thanks again for the link.

And it look like I somehow get Firefox.exe don't monitor again, I wonder I can reproduce now or not.
As this issues look like only happen while boot?
----
En...
I am not facing this case(CPU graph not showing), but I see this log.

Code Select Expand

00000001 0.00000000 [6804] Process Lasso: adding \Processor(0)\% Processor Time
00000002 0.07840922 [6804] Process Lasso: adding \Processor(1)\% Processor Time
00000003 0.07845151 [6804] Process Lasso: adding \Processor(2)\% Processor Time
00000004 0.07849216 [6804] Process Lasso: adding \Processor(3)\% Processor Time
00000005 0.07853321 [6804] Process Lasso: InitializeCoreParkingCounters_Free
00000006 0.07993928 [6804] Process Lasso: Failed to load brd value
00000007 0.09502749 [6804] Process Lasso: Excluding our modules from application compatibility assistant
00000008 0.09506650 [6804] Process Lasso: Responsiveness running in priority class 32768
00000009 0.09509031 [6804] Process Lasso: Responsiveness running at thread priority -2
00000010 0.09511042 [6804] Process Lasso: Est effective priority is 32766
00000011 0.17897396 [8252] Process Lasso: Excluding our modules from application compatibility assistant
00000012 0.19453305 [8252] Process Lasso: Initializing keep running vectors ...
00000013 0.19529089 [8252] Process Lasso: pg. Standard context
00000014 0.19610168 [8252] Process Lasso: Starting main loop
I wonder what's the meaning of Failed to load brd value...

I guess it is safe to put part of the log on forum, without the processes name, right?

[smartbit]

Hi,

I didn't noticed the update, it's beta (and I've watched the stable channel)

So, I tried today with the build 19 (though shown as 17 in about dialog  ), the graph didn't show up as usual.
And here is the PL's part of the DebugView log:

Code Select Expand

...
00000003 4.90707302 [6456] Process Lasso: adding \Processor(0)\% Processor Time
00000004 5.19769478 [6456] Process Lasso: dph failed
...
It's only 2 lines. I let it run for a couple of minutes, but nothing added.
Anyways, I expect the core graph related error appear at the initialization of the PL.
So, I hope that line 4 gives some clue...
(Edit: The CPU is 2 phys./4 logic cores, so I would expect to see the first 4 lines as in BenYeeHua's log)

I also tried the DebugView over the previous stable build (v.14), and saw this:

Code Select Expand

...
00000024 155.51936340 [27156] bitsum: Failed to load brd value
...
(concluding from the BenYeeHua's post, it's not related to the issue, but I decided to notice it anyways)

[BenYeeHua]

So lucky that I posted the first part of my log, so you can compare it.
I guess we need to do is wait, let us see that did this information is enough or not.

[Jeremy Collake]

Thanks guys!

Yes, I know exactly at what API call it's failing now, but not why. I need to add a bit more debugging logging to the next build to show extended error information at this location. That should give an important clue. So, one more time.. maybe

In .19 and above I've disabled all the debug logging except for a few spots that we're looking at here.. may change back.

p.s. Just ignore 'failed to load brd value', it's not important, and not any critical error, or even a real error. Just a notification.

[BenYeeHua]

ok, thank for telling that.
I hope we are not finding a windows bug here.

[Jeremy Collake]

We will find out soon. New build coming today that reports the error code, then we'll go from there.

I doubt it is a Windows bug, but may be an OS configuration issue. Or maybe just a bug of my creation, though it's hard to imagine what.

[Jeremy Collake]

Misc diagnostic stuff ... I've long suspected that maybe the applicable perf counters are disabled on these systems. This will help you check.

Download and install - http://download.microsoft.com/download/win2000platform/exctrlst/1.00.0.1/nt5/en-us/exctrlst_setup.exe  (info)

Then go find 'C:\Program Files (x86)\Resource Kit' (where exctrlst.exe is located).  Right-click exctrlst.exe and select 'Run as administrator'.

It will let you go through various performance counter sources. As you select each, you'll see 'Performance Counters Enabled' checked (or not) at the bottom.

Anyway, no need to waste lots of time, we'll have more detailed error info soon.

[BenYeeHua]

And it is still nice that Microsoft website/server support hot-linking or direct link, so that even you are giving the direct link to other people, you can still download it.

[Jeremy Collake]

v6.7.0.21 beta is now available. Please submit the debug output from it, as you did before, and *hopefully* it will tell us what's going on.

[edkiefer]

v6.7.0.21 beta is now available. Please submit the debug output from it, as you did before, and *hopefully* it will tell us what's going on.

Just updated to .21 and i see changes not listed in changelog .
1) log got cleared
2) no search or log-view button above bottom panel (logs list) .
3) don't see any way to open logviewer now, You used to have it in menu too .

[BenYeeHua]

And also, the log folder only left this.
1.prolasso.log (new one)
2.prolasso.log.1
3.prolasso.log.3
4.prolasso.log.5
5.prolasso.log.7
6.prolasso.log.9

1, 3, 5, 7, 9, I wonder what's their means.
Maybe PL just hate 2, 4, 6, 8.

[smartbit]

Hi,

Just tested with beta build 21.

Sadly, the output i see is exactly the same as before:

Code Select Expand

00000002 1.47741389 [20244] Process Lasso: adding \Processor(0)\% Processor Time
00000003 1.86653686 [20244] Process Lasso: dph failed
(The Core Graph is not present)