Old 01-17-2018, 03:26 AM   #1
Cableaddict
Human being with feelings
 
Join Date: Apr 2008
Posts: 1,883
Default CPU PRIORITY SETTINGS INFO & QUESTIONS !

SETTING CPU PRIORITY IN THE TASK MANAGER:

(Important info for those that didn't know about this, plus some important questions at the end, for anyone who understands this far better than me. - Like Karbo !)
---------------------------------------------

I’ve recently become aware of the benefit of setting Reaper to a “real time” cpu priority setting, via the task manager. I thought Reaper did this automatically, when you adjust the buffering preferences, but evidently it doesn’t. Even with Reaper’s “thread priority” prefs set to “Time Critical,” the Task manager cpu priority was still set to “normal.”

Changing this setting to “real time” made a HUGE difference in how small I could set my HW buffer.
I’m now running my large “live performance” session at 96K, with a 256 buffer, whereas before this setting I was getting minor but consistent crackling. I can even run this session clean (just barely) at a 128 buffer. (Which is crazy fast for the Lynx driver, giving a true in-out throughput of about 3 ms!
(I can even set it to 64, with a little crackling, which to me is mind-blowing.)

Even though I will run this session at 256, I am using 128 or even 64 as a test bed, to see what other optimizations might be possible. (I know most of them)
================================================== ==========

NOW FOR MY QUESTIONS:

1: Should any OTHER audio processes be changed to “real time?”

I run Traktor concurrently, so for now I’ve set that to real time as well, but I have no idea if that could be a problem. Maybe it should be set to “High,” so it’s above all the Windows stuff, but below Reaper?

What about the four “Reaper Host 64…” processes listed underneath the main “Reaper” one? The article I read about this whole issue mentioned only the “main” DAW setting.
For now, I’ve changed them all, and so far it doesn’t seem to hurt, but can anyone comment on this?
——————————————————

2: Should any processes be changed from “normal” to “low” ?

Almost all of them default to “normal,” and only a few cannot be changed.
I immediately set “PACE” to low, and also Windows Explorer. (though I may completely defeat Explorer at some point, since I only need “Search / Cortana.”)

Anything else?

Perhaps such changes aren't needed at all, as long as the audio apps have a higher priority?
——————————————————

3: THIS IS MY MAIN QUESTION:

When running at 96K and a 128 buffer, again, I can play any tracks and ay VSTi’s I want, and the audio is almost 100% clean. HOWEVER, I'm clearly near the edge:

If audio is playing, and I move anything in the GUI with my trackball, this causes a terrible sustained crackling! - And if I type anything on the keyboard, the audio completely drops out!

So obviously, either the usb devices, or the graphics card / driver, are getting higher priority than my audio. (Correct?) So I immediately thought, why not set these all to “normal” or even “low” cpu priority. I surely don’t care if the screen has to wait for my audio.
Except, THESE THINGS ARE NOT LISTED IN THE TASK MANAGER > DETAILS WINDOW !

SO: Is there any way to lower the priority for these things?

Is there any other solution for this problem, or is this hard-coded into Windows 10?

Last edited by Cableaddict; 01-17-2018 at 04:32 AM.
Cableaddict is offline   Reply With Quote
Old 01-17-2018, 05:32 AM   #2
seymour22
Human being with feelings
 
seymour22's Avatar
 
Join Date: Nov 2013
Posts: 109
Default

Hi,
thanks for sharing!
Does it exist on Mac? Or windows task manager strict?

Kr
Tomek
seymour22 is offline   Reply With Quote
Old 01-17-2018, 06:15 AM   #3
ivansc
Human being with feelings
 
Join Date: Aug 2007
Location: Near Cambridge UK and Near Questembert, France
Posts: 19,159
Default

giving it a go on the studio machine, which can already run the minimum available buffer from my RME hardware. Will give it a go on my i5 laptop later to see if it gees things up there too. Anything that gees things up a bit....
__________________
"What a dick comment. I'm gonna make sure to avoid your name." Dicks other than Trump can speak????
ivansc is offline   Reply With Quote
Old 01-17-2018, 06:27 AM   #4
JHughes
Human being with feelings
 
JHughes's Avatar
 
Join Date: Aug 2007
Location: Too close to Charlotte, NC
Posts: 3,531
Default

I don't think running processes in real time is a good idea, but maybe you'll figure out how to make it work.

Rather, I would run Process Lasso in Performance mode with Reaper in the foreground.
__________________
You can only Reverse, Switch, Invert or Flip POLARITY, not "PHASE".
JHughes is offline   Reply With Quote
Old 01-17-2018, 09:19 AM   #5
ivansc
Human being with feelings
 
Join Date: Aug 2007
Location: Near Cambridge UK and Near Questembert, France
Posts: 19,159
Default

What is " Process Lasso in Performance mode"?
__________________
"What a dick comment. I'm gonna make sure to avoid your name." Dicks other than Trump can speak????
ivansc is offline   Reply With Quote
Old 01-17-2018, 09:25 AM   #6
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 25,144
Default

Quote:
Originally Posted by ivansc View Post
What is " Process Lasso in Performance mode"?
If you look in task manager, you can set affinity and priority for processes etc. via right-click. This app sort of automates such things, I can't comment on it's usefulness as I have not used it but it is sort of trying to out guess the OS and application settings priority wise, I'd use it with care either way aka here be dragons.

To the OP, can't set priority for hardware, as I mentioned in the other thread, hardware cannot wait, when it signals the OS that it has data etc., the OS must drop what it is doing for a to service that request. There are usually only two true fixes...

1. Find out what is causing the delay and reduce it if possible such as disabling a wireless NIC if it were the reason and so on.

2. The machine is just too under-scaled to handle the work it is being asked to do and whatever DPCs etc complain are simply canaries in the proverbial mineshaft.
__________________
Clever is not the same as good, clever is from the head, good is from the heart. -Benmont Tench
karbomusic is offline   Reply With Quote
Old 01-17-2018, 10:03 AM   #7
JHughes
Human being with feelings
 
JHughes's Avatar
 
Join Date: Aug 2007
Location: Too close to Charlotte, NC
Posts: 3,531
Default

Karbo always confuses this app with the old snake oil ones that just boost a priority.

READ THIS:
https://bitsum.com/how-probalance-works/

So what does it primarily accomplish?
The other day I had 100 tabs open in Firefox, with at least five youtube videos playing at the same time. I also had a 30 track Reaper project playing at the same time (not one with a lot of VSTi's though because I didn't have one handy). I had a video game running in a window to get DirectX in the mix. And I had Prime95 x64 v28.9 Build 2 running small FFTs. All at the same time and I had full mouse control.

I'm the only guy on this forum who uses it.

I'm also the only one who has ever set up his monitors in a Cardas golden ratio setup, but that's another dead thread. Folks here don't like to experiment much regardless of the upside.
__________________
You can only Reverse, Switch, Invert or Flip POLARITY, not "PHASE".

Last edited by JHughes; 01-17-2018 at 10:08 AM. Reason: Added stuff
JHughes is offline   Reply With Quote
Old 01-17-2018, 10:15 AM   #8
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 25,144
Default

Quote:
Originally Posted by JHughes View Post
Karbo always confuses this app with the old snake oil ones that just boost a priority.
I'm not knocking it, just saying there are only a handful of things that can be managed such as priority and affinity and a few other built in scheduling/process and thread management settings that can be changed through the OS native API.

I'm sure it's a fine app, especially when someone is at that edge of scale and near the cliff but IMHO, it's better to not be there to begin with if that can be helped but that choice is understandably not always available. YMMV et al.

I'm pretty lucky I suppose, I can run 100+ tracks + 200+ VSTs with a limiter on top running @32X oversampling @64 samples buffer, with room to spare; and all with a 100% stock install which also has everything I do installed on it including office stuff, programming, DAW whathaveyou. AKA it's my daily driver machine. I'm so happy with it, I'm worried it's gonna go south and inevitably lose something even with full backups.
__________________
Clever is not the same as good, clever is from the head, good is from the heart. -Benmont Tench
karbomusic is offline   Reply With Quote
Old 01-17-2018, 01:27 PM   #9
mlprod
Human being with feelings
 
Join Date: Jul 2015
Location: Stockholm, Sweden
Posts: 721
Default

Very interesting subject. As for the CPU priority, I assumed that the thead priority in Reaper perfs/buffer setting is actually setting this for us instead of going to the taskmanager, as far as I can tell the TM does not save settings.
So you are saying that this is not the case?
__________________
Magnus Lindberg Productions - VRTKL Audio - Redmount Studios
magnuslindberg.com
mlprod is offline   Reply With Quote
Old 01-17-2018, 01:32 PM   #10
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 25,144
Default

Quote:
Originally Posted by mlprod View Post
Very interesting subject. As for the CPU priority, I assumed that the thead priority in Reaper perfs/buffer setting is actually setting this for us instead of going to the taskmanager, as far as I can tell the TM does not save settings.
So you are saying that this is not the case?
In Reaper the priorities are thread priorities, in task manager it's process priorities. For the uninitiated Processes are containers for threads. When creating a thread in code, it's priority and even it's affinity can be set upon creation. In reaper settings, thread priority and behavior is how to do that from a user perspective.

Something to note, do not automatically assume the names of thread behaviors match the expected behavior, it changes based on system IME. Meaning, you might think one setting is the one, based on name, is exactly what you want, but in the real world the exact opposite may occur so experiment. I've had mine set to Automatic (Experimental) for as long as I can remember FWIW.
__________________
Clever is not the same as good, clever is from the head, good is from the heart. -Benmont Tench
karbomusic is offline   Reply With Quote
Old 01-17-2018, 02:06 PM   #11
mlprod
Human being with feelings
 
Join Date: Jul 2015
Location: Stockholm, Sweden
Posts: 721
Default

Quote:
Originally Posted by karbomusic View Post
In Reaper the priorities are thread priorities, in task manager it's process priorities. For the uninitiated Processes are containers for threads. When creating a thread in code, it's priority and even it's affinity can be set upon creation. In reaper settings, thread priority and behavior is how to do that from a user perspective.
I see, so setting this in TM is still relevent then? Is there a way to make the settings in TM permanent?
__________________
Magnus Lindberg Productions - VRTKL Audio - Redmount Studios
magnuslindberg.com
mlprod is offline   Reply With Quote
Old 01-17-2018, 02:27 PM   #12
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 25,144
Default

Quote:
Originally Posted by mlprod View Post
I see, so setting this in TM is still relevent then? Is there a way to make the settings in TM permanent?
I don't think the priority setting is permanent.

I also think it is a potential futile action via TM 'generally' because the need to do so is likely trying to get around the real issue - whatever that may be. I have seen instances in non-audio scenarios where restricting an app to a particular core for example can be worthwhile but again doesn't really apply here. I wouldn't discourage someone from changing process priority just to see what happens though - most experts become experts from trying everything.

Secondly, better to just control from within Reaper at the thread level since that is based on Justin's knowledge of threads dealing with audio and so on. The same sort of applies there too, if the machine is under-scaled or there is some piece of hardware generating unnecessary DPCs and interrupts, adjusting thread behavior is likely not going to help but in general tuning it probably will help.
__________________
Clever is not the same as good, clever is from the head, good is from the heart. -Benmont Tench
karbomusic is offline   Reply With Quote
Old 01-17-2018, 06:40 PM   #13
Dstruct
Human being with feelings
 
Join Date: Jul 2006
Posts: 12,300
Default

Quote:
Originally Posted by Cableaddict View Post
I’ve recently become aware of the benefit of setting Reaper to a “real time” cpu priority setting, via the task manager. I thought Reaper did this automatically, when you adjust the buffering preferences, but evidently it doesn’t. Even with Reaper’s “thread priority” prefs set to “Time Critical,” the Task manager cpu priority was still set to “normal.”
Quote:
Originally Posted by Justin View Post
Don't do that, instead make sure you're running with realtime permissions, then set the realtime priority either in prefs/audio/device (if using ALSA), or via JACK. You *really* do not want the GUI thread to be realtime, only the audio threads, and even then they need to be set with the correct relative priorities (which REAPER will do automatically if you configure it to).

Same applies to Windows.
Dstruct is offline   Reply With Quote
Old 01-17-2018, 06:43 PM   #14
Cableaddict
Human being with feelings
 
Join Date: Apr 2008
Posts: 1,883
Default

Quote:
Originally Posted by karbomusic View Post
If you look in task manager, you can set affinity and priority for processes etc. via right-click. This app sort of automates such things, I can't comment on it's usefulness as I have not used it but it is sort of trying to out guess the OS and application settings priority wise, I'd use it with care either way aka here be dragons.
"Process Lasso" sounds lie an app I used to use in the XP days, called "Prio."
It was a life-saver. But back then, there was no way to do it manually, AFAIK.

Using "Process Lasso" might be very helpful if a PC has to be run in different configurations all the time, and so needs to be quickly re-configured.

For a dedicated DAW, though, I thing doing it manually is best, esp since you can make one change at a time & see how things go.
Cableaddict is offline   Reply With Quote
Old 01-17-2018, 06:45 PM   #15
Cableaddict
Human being with feelings
 
Join Date: Apr 2008
Posts: 1,883
Default

Quote:
Originally Posted by JHughes View Post
I don't think running processes in real time is a good idea, but maybe you'll figure out how to make it work.

Rather, I would run Process Lasso in Performance mode with Reaper in the foreground.

Well, It's working for me right now.

- And an AMAZING difference in low-latency performance. I can't even believe it.....



Note: this is on an "AUDIO ONLY" Pc. No internet, no video, no games ..... YMMV.
Cableaddict is offline   Reply With Quote
Old 01-17-2018, 06:55 PM   #16
Cableaddict
Human being with feelings
 
Join Date: Apr 2008
Posts: 1,883
Default

Quote:
Originally Posted by karbomusic View Post

To the OP, can't set priority for hardware, as I mentioned in the other thread, hardware cannot wait, when it signals the OS that it has data etc., the OS must drop what it is doing for a to service that request. ......

Karbo, I'm learning slowly that I should only question your information with EXTREME caution !!!

But as I mentioned elsewhere, all hardware is still controlled by drivers, which is software.
Is there really no way to set a USB or video driver to lower priority?

Regarding USB hardware:

Would it possibly help if I disabled any USB ports not being used? That would take some time (experimenting to find which is which) but it's doable.

------------

Likewise: I very recently found out about USB power-management. I was having a problem with my iLOK, after updating my version of Windows10. The iLOK light kept going out after inserting it, withn maybe 10 seconds, and then my PACE-protected software wouldn't run.

I was also having intermittent failures with my external midi controller.

Then I read how you can set USB power-management to be "always on" just like we do with the monitors, drives, etc. (this must be a new thing with Windows)
I did this, setting ALL usb ports to "always on" and my iLOK and midi controller problems went away.

- But could this possibly be making my trackball hardware interrupts worse?




my head's starting to hurt ......

Last edited by Cableaddict; 01-18-2018 at 03:00 AM.
Cableaddict is offline   Reply With Quote
Old 06-12-2018, 04:58 PM   #17
godly
Human being with feelings
 
godly's Avatar
 
Join Date: Sep 2013
Location: Belgium - Gent
Posts: 702
Default

To pick up the thread here again,
I just found out that using Jbridge inside Reaper saved me from allot of troubles !!!

I run Jbridge inside Reaper because VPS Avenger gives allot of problems, and it is just perfect !
I use the x64 version to bridge inside x64 Reaper.
Before Avenger gave glitches audio cracks on my system and even sometimes crashed, because of some bugs. With Jbrdige all problems are solved, I can even run at 128 samples !

If Jbridge can secure these problems, why can't Reaper have a better bridging (or hosting) system with better options to prevent glitches and crashes?

Jbridge processes the audio async. With one buffer extra latency. So now I can run Avenger at 128 samples instead of 512 !

There's also a thread safety measure, that prevents crashing Reaper (before some plugins and Avenger could crash reaper, not funny when playing live with it)

Could you devs take a decent look into it???
godly is offline   Reply With Quote
Old 06-12-2018, 10:04 PM   #18
Jeffsounds
Human being with feelings
 
Jeffsounds's Avatar
 
Join Date: Feb 2013
Location: Northeast Michigan
Posts: 3,089
Default

Quote:
Originally Posted by Cableaddict View Post
What about the four “Reaper Host 64…” processes listed underneath the main “Reaper” one?
I got a bit curious and decided to check this out. I'm using 64-bit Reaper on a Win7, 64-bit machine but I don't have Reaper Host 64 in my task manager, I have reaper_host32.exe *32. Why is it different?
__________________
"Music is a fundamental human desire." - Mike Butera
Jeffsounds is offline   Reply With Quote
Old 09-03-2018, 11:24 AM   #19
Pictus
Human being with feelings
 
Pictus's Avatar
 
Join Date: Sep 2017
Location: Brasil
Posts: 5
Default

The FREE Process Hacker BETA can change the "Processes Affinity", "Priority" and "I/O Priority".
In the Process Hacker "Processes" TAB, right click to change things.

Process Hacker must be running for the "Processes Affinity" settings(when saved) to be
applied, but not for the Priority, Process Hacker can start at logon.
Pictus is offline   Reply With Quote
Old 10-22-2019, 10:47 AM   #20
Tim Rideout
Human being with feelings
 
Tim Rideout's Avatar
 
Join Date: Jan 2013
Location: Montreal, Canada
Posts: 200
Default

This video explains everything you need to know about your misbehaving CPU:

https://www.youtube.com/watch?v=GUsLLEkswzE
__________________
---
www.TimRideout.com
Tim Rideout is offline   Reply With Quote
Old 10-22-2019, 01:33 PM   #21
azslow3
Human being with feelings
 
Join Date: Nov 2017
Location: Heidelberg, Germany
Posts: 747
Default

Quote:
Originally Posted by Tim Rideout View Post
This video explains everything you need to know about your misbehaving CPU:

https://www.youtube.com/watch?v=GUsLLEkswzE
Almost everything... But it is not up to date and not about REAPER:
* LatencyMon in its current version is good (he mention DPC latency checker does not work on modern Windows). Another tool is Microsoft Windows Performance Toolkit (also free). Using both it is possible to identify exact driver which introduce latency or identify that something wrong with CPU/Chipset in general. So, no more "guesses" which device to disable (what is suggested in the video), till the same driver is used for many devices (f.e. ACPI.sys).
* it seems like all modern systems can not deliver top latency till all power saving features are disabled. On notebooks that is a big problem, they can not work in full power all the time. On desktop that is less problematic, it just requires better cooling (so can be a bit more noisy).
* finally, REAPER is processing all not record armed tracks in "offline" (not real-time) mode. So only live track and recording is affected by system latency (till it is extra high). Demonstrated playback problems in the video normally do not exist in REAPER.
azslow3 is offline   Reply With Quote
Old 10-24-2019, 09:31 AM   #22
Tim Rideout
Human being with feelings
 
Tim Rideout's Avatar
 
Join Date: Jan 2013
Location: Montreal, Canada
Posts: 200
Default

Thanks for your response - I enjoyed the video because it goes in-depth about the fact that your CPU has very little to do with real time audio these days - the bottleneck is most often badly-behaving peripherals. I was not aware of this.

To this end, I have been using LatencyMon, but have not been able to identify the driver. It's not specific enough. The new DPC Latency Checker has been working on my system, and shows a constant 1000mu latency - right on the edge of being ok for audio, so I dont understand why . I'm not getting any spikes - just constant high DPC. I've disabled every device I can, except the graphics cards - but no change, even with all power saving off. I'm not sure where to go from here, and I can't access the windows tool you speak of...
__________________
---
www.TimRideout.com

Last edited by Tim Rideout; 11-15-2019 at 10:34 AM. Reason: spelling
Tim Rideout is offline   Reply With Quote
Old 10-24-2019, 10:34 AM   #23
azslow3
Human being with feelings
 
Join Date: Nov 2017
Location: Heidelberg, Germany
Posts: 747
Default

http://www.azslow.com/index.php/topi...8.html#msg2318
Open "Use Latency Monitor...." spoiler.
DPC checker shows what it shows for you on Win10, just because it is not designed for modern Windows. So what it shows is more or less useless.

I suggest you go in sequence I propose in the linked post. At some point you will see spikes and from which concrete driver they come (can be bad luck, f.e. with ACPI, then you will have to go throw all ACPI devices and there are many and some are important...).
But if "in depth" test give you 0.5ms strait away, there is no reason to search drivers (till they are much worse). For desktop that means temperature throttling, not fixes voltage/frequency, not matching RAM, etc. For notebook that is just "game over" for low latency (the only mode I have not checked for my XPS is fixed at some low frequency clock and C/P states disabled, since without lowering the frequency it sounds like a helicopter when power management is disabled).
Hopefully tomorrow I will assemble i9 based computer, I will publish all measurements for it in several modes (so other can compare/comment).
azslow3 is offline   Reply With Quote
Old 11-15-2019, 10:50 AM   #24
Tim Rideout
Human being with feelings
 
Tim Rideout's Avatar
 
Join Date: Jan 2013
Location: Montreal, Canada
Posts: 200
Default

Thanks AZ!

You are right about DPC Latency Checker - on their own site, they say it reports incorrect values - thank you for that info.

Great advice on your site, too: you are saying that a 0.5 MILLIsecond latency in LatencyMon is perfectly acceptable? I get 827 MICROseconds reported as highest from my Nvidia graphics driver

Thanks!

T


Quote:
Originally Posted by azslow3 View Post
http://www.azslow.com/index.php/topi...8.html#msg2318
Open "Use Latency Monitor...." spoiler.
DPC checker shows what it shows for you on Win10, just because it is not designed for modern Windows. So what it shows is more or less useless.

I suggest you go in sequence I propose in the linked post. At some point you will see spikes and from which concrete driver they come (can be bad luck, f.e. with ACPI, then you will have to go throw all ACPI devices and there are many and some are important...).
But if "in depth" test give you 0.5ms strait away, there is no reason to search drivers (till they are much worse). For desktop that means temperature throttling, not fixes voltage/frequency, not matching RAM, etc. For notebook that is just "game over" for low latency (the only mode I have not checked for my XPS is fixed at some low frequency clock and C/P states disabled, since without lowering the frequency it sounds like a helicopter when power management is disabled).
Hopefully tomorrow I will assemble i9 based computer, I will publish all measurements for it in several modes (so other can compare/comment).
__________________
---
www.TimRideout.com
Tim Rideout is offline   Reply With Quote
Old 11-15-2019, 05:42 PM   #25
azslow3
Human being with feelings
 
Join Date: Nov 2017
Location: Heidelberg, Germany
Posts: 747
Default

Quote:
Originally Posted by Tim Rideout View Post
Thanks AZ!
Great advice on your site, too: you are saying that a 0.5 MILLIsecond latency in LatencyMon is perfectly acceptable? I get 827 MICROseconds reported as highest from my Nvidia graphics driver
0.5 ms (500 uS) is something possible work with. Not perfect, especially if you try to work with small RTL. For my taste 800uS is a bit too much in case it is not "one time" and you get it every test without something big running in the background.

Intel i9/MSI z390 for me:
* in depth test shows spikes up to 150 uS. C states and frequency lock independent.
* latency monitor shows around 80 uS max delay, with DPC of NVidia and other up to 300uS.

REAPER does not show 300uS delay even when it is reported in Latency Monitor. I have the feeling more or less proper working drivers with real hardware can work "better" then Latency Monitor.

All 3 interfaces I have are able to work with minimal settings they support without glitches. I have also decided to leave BIOS in default state, without disabling power options. The difference is not worse extra power continuously consumed for me.

I can not drive RT too high, after 50-60% RT CPU glitches can happened. Fortunately I do not need that.

I still wonder how some people get under 100uS everywhere and can work with 90% RT CPU load. I am more skeptical about such reports now.
azslow3 is offline   Reply With Quote
Old 11-18-2019, 03:00 PM   #26
sightlessness
Human being with feelings
 
sightlessness's Avatar
 
Join Date: Mar 2010
Posts: 937
Default

@ azlow3

I get this using In Depth Latency Tests (running Reaper using 50% CPU)
Half the values when not running Reaper or anything), am I fcked? :

PER CPU MEASURINGS
__________________________________________________ __________________________________________________ _____
Highest latency on CPU 0: 537.60 µs
Highest latency on CPU 1: 538.80 µs
Highest latency on CPU 2: 537.90 µs
Highest latency on CPU 3: 539.10 µs
Highest latency on CPU 4: 538.20 µs
Highest latency on CPU 5: 539.50 µs
Highest latency on CPU 6: 547.40 µs
Highest latency on CPU 7: 526.50 µs


__________________________________________________ __________________________________________________ _____
SYSTEM INFORMATION
__________________________________________________ __________________________________________________ _____
Computer name: LAPTOP-AU123PRE
OS version: Windows 10 , 10.0, build: 17763Highest latency on CPU 7: 526.50 µs
Hardware: Nitro AN515-51, Acer, KBL, Freed_KLS
CPU: GenuineIntel Intel(R) Core(TM) i7-7700HQ CPU @ 2.80GHz
Logical processors: 8
Processor groups: 1
RAM: 16267 MB total

Running this test while Reaper running almost stalls my laptop, ouch. So much stuttering I barely could push stop on Reaper, lol.
__________________
I want to live PEACEFULLY PLEASE WORLD "LEADERS" GET THIS DONE/LET IT BE FOR GOOD AND MAKE HISTORYYYYYYY! Thanks.
sightlessness is offline   Reply With Quote
Old 11-18-2019, 03:05 PM   #27
sightlessness
Human being with feelings
 
sightlessness's Avatar
 
Join Date: Mar 2010
Posts: 937
Default

And using LatencyMon, I got this:

__________________________________________________ __________________________________________________ _____
CONCLUSION
__________________________________________________ __________________________________________________ _____
Your system seems to be having difficulty handling real-time audio and other tasks. You may experience drop outs, clicks or pops due to buffer underruns. One or more DPC routines that belong to a driver running in your system appear to be executing for too long. One problem may be related to power management, disable CPU throttling settings in Control Panel and BIOS setup. Check for BIOS updates.
LatencyMon has been analyzing your system for 0:00:19 (h:mm:ss) on all processors.


__________________________________________________ __________________________________________________ _____
SYSTEM INFORMATION
__________________________________________________ __________________________________________________ _____
Computer name: LAPTOP-AU123PRE
OS version: Windows 10 , 10.0, version 1809, build: 17763 (x64)
Hardware: Nitro AN515-51, Acer, KBL, Freed_KLS
CPU: GenuineIntel Intel(R) Core(TM) i7-7700HQ CPU @ 2.80GHz
Logical processors: 8
Processor groups: 1
RAM: 16267 MB total


__________________________________________________ __________________________________________________ _____
CPU SPEED
__________________________________________________ __________________________________________________ _____
Reported CPU speed: 2808 MHz

Note: reported execution times may be calculated based on a fixed reported CPU speed. Disable variable speed settings like Intel Speed Step and AMD Cool N Quiet in the BIOS setup for more accurate results.

WARNING: the CPU speed that was measured is only a fraction of the CPU speed reported. Your CPUs may be throttled back due to variable speed settings and thermal issues. It is suggested that you run a utility which reports your actual CPU frequency and temperature.



__________________________________________________ __________________________________________________ _____
MEASURED INTERRUPT TO USER PROCESS LATENCIES
__________________________________________________ __________________________________________________ _____
The interrupt to process latency reflects the measured interval that a usermode process needed to respond to a hardware request from the moment the interrupt service routine started execution. This includes the scheduling and execution of a DPC routine, the signaling of an event and the waking up of a usermode thread from an idle wait state in response to that event.

Highest measured interrupt to process latency (µs): 1009.50
Average measured interrupt to process latency (µs): 7.278018

Highest measured interrupt to DPC latency (µs): 1004.40
Average measured interrupt to DPC latency (µs): 2.226081


__________________________________________________ __________________________________________________ _____
REPORTED ISRs
__________________________________________________ __________________________________________________ _____
Interrupt service routines are routines installed by the OS and device drivers that execute in response to a hardware interrupt signal.

Highest ISR routine execution time (µs): 61.094017
Driver with highest ISR routine execution time: ACPI.sys - ACPI Driver for NT, Microsoft Corporation

Highest reported total ISR routine time (%): 0.006291
Driver with highest ISR total time: Wdf01000.sys - Kernel Mode Driver Framework Runtime, Microsoft Corporation

Total time spent in ISRs (%) 0.006520

ISR count (execution time <250 µs): 3462
ISR count (execution time 250-500 µs): 0
ISR count (execution time 500-999 µs): 0
ISR count (execution time 1000-1999 µs): 0
ISR count (execution time 2000-3999 µs): 0
ISR count (execution time >=4000 µs): 0


__________________________________________________ __________________________________________________ _____
REPORTED DPCs
__________________________________________________ __________________________________________________ _____
DPC routines are part of the interrupt servicing dispatch mechanism and disable the possibility for a process to utilize the CPU while it is interrupted until the DPC has finished execution.

Highest DPC routine execution time (µs): 1295.408120
Driver with highest DPC routine execution time: Wdf01000.sys - Kernel Mode Driver Framework Runtime, Microsoft Corporation

Highest reported total DPC routine time (%): 0.385919
Driver with highest DPC total execution time: Wdf01000.sys - Kernel Mode Driver Framework Runtime, Microsoft Corporation

Total time spent in DPCs (%) 0.457927

DPC count (execution time <250 µs): 40821
DPC count (execution time 250-500 µs): 0
DPC count (execution time 500-999 µs): 354
DPC count (execution time 1000-1999 µs): 7
DPC count (execution time 2000-3999 µs): 0
DPC count (execution time >=4000 µs): 0


__________________________________________________ __________________________________________________ _____
REPORTED HARD PAGEFAULTS
__________________________________________________ __________________________________________________ _____
Hard pagefaults are events that get triggered by making use of virtual memory that is not resident in RAM but backed by a memory mapped file on disk. The process of resolving the hard pagefault requires reading in the memory from disk while the process is interrupted and blocked from execution.

NOTE: some processes were hit by hard pagefaults. If these were programs producing audio, they are likely to interrupt the audio stream resulting in dropouts, clicks and pops. Check the Processes tab to see which programs were hit.

Process with highest pagefault count: svchost.exe

Total number of hard pagefaults 1
Hard pagefault count of hardest hit process: 1
Number of processes hit: 1
__________________
I want to live PEACEFULLY PLEASE WORLD "LEADERS" GET THIS DONE/LET IT BE FOR GOOD AND MAKE HISTORYYYYYYY! Thanks.
sightlessness is offline   Reply With Quote
Old 11-18-2019, 03:06 PM   #28
sightlessness
Human being with feelings
 
sightlessness's Avatar
 
Join Date: Mar 2010
Posts: 937
Default

And this:


__________________________________________________ __________________________________________________ _____
PER CPU DATA
__________________________________________________ __________________________________________________ _____
CPU 0 Interrupt cycle time (s): 1.047280
CPU 0 ISR highest execution time (µs): 15.269943
CPU 0 ISR total execution time (s): 0.009571
CPU 0 ISR count: 3456
CPU 0 DPC highest execution time (µs): 1295.408120
CPU 0 DPC total execution time (s): 0.628825
CPU 0 DPC count: 35884
__________________________________________________ __________________________________________________ _____
CPU 1 Interrupt cycle time (s): 0.543614
CPU 1 ISR highest execution time (µs): 61.094017
CPU 1 ISR total execution time (s): 0.000348
CPU 1 ISR count: 6
CPU 1 DPC highest execution time (µs): 982.378561
CPU 1 DPC total execution time (s): 0.058796
CPU 1 DPC count: 3726
__________________________________________________ __________________________________________________ _____
CPU 2 Interrupt cycle time (s): 0.098045
CPU 2 ISR highest execution time (µs): 0.0
CPU 2 ISR total execution time (s): 0.0
CPU 2 ISR count: 0
CPU 2 DPC highest execution time (µs): 43.920940
CPU 2 DPC total execution time (s): 0.000499
CPU 2 DPC count: 215
__________________________________________________ __________________________________________________ _____
CPU 3 Interrupt cycle time (s): 0.425249
CPU 3 ISR highest execution time (µs): 0.0
CPU 3 ISR total execution time (s): 0.0
CPU 3 ISR count: 0
CPU 3 DPC highest execution time (µs): 33.245726
CPU 3 DPC total execution time (s): 0.004035
CPU 3 DPC count: 449
__________________________________________________ __________________________________________________ _____
CPU 4 Interrupt cycle time (s): 0.115346
CPU 4 ISR highest execution time (µs): 0.0
CPU 4 ISR total execution time (s): 0.0
CPU 4 ISR count: 0
CPU 4 DPC highest execution time (µs): 60.799501
CPU 4 DPC total execution time (s): 0.000629
CPU 4 DPC count: 179
__________________________________________________ __________________________________________________ _____
CPU 5 Interrupt cycle time (s): 0.108929
CPU 5 ISR highest execution time (µs): 0.0
CPU 5 ISR total execution time (s): 0.0
CPU 5 ISR count: 0
CPU 5 DPC highest execution time (µs): 32.906695
CPU 5 DPC total execution time (s): 0.000455
CPU 5 DPC count: 156
__________________________________________________ __________________________________________________ _____
CPU 6 Interrupt cycle time (s): 0.115109
CPU 6 ISR highest execution time (µs): 0.0
CPU 6 ISR total execution time (s): 0.0
CPU 6 ISR count: 0
CPU 6 DPC highest execution time (µs): 42.189103
CPU 6 DPC total execution time (s): 0.000510
CPU 6 DPC count: 92
__________________________________________________ __________________________________________________ _____
CPU 7 Interrupt cycle time (s): 0.111567
CPU 7 ISR highest execution time (µs): 0.0
CPU 7 ISR total execution time (s): 0.0
CPU 7 ISR count: 0
CPU 7 DPC highest execution time (µs): 40.904915
CPU 7 DPC total execution time (s): 0.002886
CPU 7 DPC count: 481
__________________________________________________ __________________________________________________ _____
__________________
I want to live PEACEFULLY PLEASE WORLD "LEADERS" GET THIS DONE/LET IT BE FOR GOOD AND MAKE HISTORYYYYYYY! Thanks.
sightlessness is offline   Reply With Quote
Old 11-18-2019, 03:18 PM   #29
sightlessness
Human being with feelings
 
sightlessness's Avatar
 
Join Date: Mar 2010
Posts: 937
Default

So, in the CPUs section, ACPI.sys had a high DPC execution, and DPC count of 3726
but Wdf01000.sys was higher (35884) and ISR count of 3456

In the Drivers' one, rspLLL64.sys (REsplendence Latency monitoring) had a DPC count of 30875
Wdf01000.sys a DPC count of 3454 (Total execution of 596 ms)
ntoskrnl 2335
nvlddmkm.sys 1756
dxgkrnl.sys 1355
porcls.sys 539
storpot.sys 105

I don't know what to do to tap the other 50,40% of my cpu... I open one more plugin and starts stuttering bad... and I am using the highest buffer settings I can: 4096 samples.
__________________
I want to live PEACEFULLY PLEASE WORLD "LEADERS" GET THIS DONE/LET IT BE FOR GOOD AND MAKE HISTORYYYYYYY! Thanks.
sightlessness is offline   Reply With Quote
Old 11-19-2019, 03:04 AM   #30
azslow3
Human being with feelings
 
Join Date: Nov 2017
Location: Heidelberg, Germany
Posts: 747
Default

Depth Latency Tests is for idle system. It puts about 80% load on all cores, running REAPER in parallel with some plug-ins logically put CPU into impossible to solve condition.
Also except on desktop with extra cooling CPU can not work in turbo mode for long without throttling.

From my observations 250-300 uS in idle there is ok. At least without tweaking. There are posts which claim it should be under 100uS, but these posts are quite old and not for Win10. I am also not sure which timer is used to get reported numbers.

ISR 60uS is fine.

1ms max DPC from Wdf01000.sys is the only questionable measurement, may be worse to check. Wdf is a framework, LatencyMon can not resolve it down to real driver. Use https://docs.microsoft.com/en-us/win...ware/test/wpt/
During ADK installation select WP tools only, there is no separate WP download.
The interface can be confusing at first, i have not found good tutorial video. In short, you run recorder for a while (default settings are fine, if I remember correctly). At the end you will be asked to run analyzer. Find DPC graphs on the left panel and check which real driver has big value. I bet USB...

Note that 1ms can still be fine. Check REAPER RT performance monitor, if you do not see underruns there (with buffer setting 128) everything is ok.

The math is simple: 48kHZ/64samples is 1.5ms. If something periodically delay by 1ms, there is just 0.5ms for processing job. Depending from plug-ins that can be sufficient or not sufficient.
I guess (note, just IMHO) the difference between RT CPU in % and RT buffer in % comes from all delays. F.e. when I have RT CPU 10%, I have 0.26ms/1ms for the buffer, so 26%. In fact I start to get cracks/pops when RT CPU is around 60-70%, far below 100%.

Note that some interfaces have huge safety buffers, that can be recognized by total latency RTL. F.e. my Phonic with 96kHz/64 samples and 1ms(lowest) safety setting has ~7ms RTL. So in REAPER monitoring it is more forgiving then 96kHz/96 samples for BFP with 3ms RTL.

If RTL is toward 10ms, +-1ms system latency is almost not an issue.

Processes priority (original subj of this thread) can play significant role once (RT) CPU load from the DAW is high and/or interface drivers are not perfect.

I think these days it is simpler to get more powerful system in this case then attempt to get last 10-15% from existing. Till the system already is most powerful from available (or the next level is several times more expensive). Or check which plug-ins are loaded and may be remove some or lower there settings.

Note that RT REAPER monitoring is affected by in-plugin waiting times. F.e. if you use "bad written" sampler (f.e. my FluidSynth VST wrapper...), the time not cached sample is loaded from the disk is counted as RT (while it has nothing to do with CPU power).
azslow3 is offline   Reply With Quote
Old 11-19-2019, 09:01 AM   #31
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 25,144
Default

Quote:
Originally Posted by azslow3 View Post
Use https://docs.microsoft.com/en-us/win...ware/test/wpt/
During ADK installation select WP tools only, there is no separate WP download.
The interface can be confusing at first, i have not found good tutorial video. In short, you run recorder for a while (default settings are fine, if I remember correctly). At the end you will be asked to run analyzer. Find DPC graphs on the left panel and check which real driver has big value. I bet USB...
You might try windows media experience analyzer which is also WPR/WPA (all are actually XPerf under the covers) - I'm pretty familiar with XPerf but haven't played with MXA yet...

https://www.microsoft.com/en-us/down....aspx?id=43105

Video...

https://channel9.msdn.com/Blogs/WinH...zer-MXA-Part-1
__________________
Clever is not the same as good, clever is from the head, good is from the heart. -Benmont Tench
karbomusic is offline   Reply With Quote
Old 11-19-2019, 03:13 PM   #32
godly
Human being with feelings
 
godly's Avatar
 
Join Date: Sep 2013
Location: Belgium - Gent
Posts: 702
Default

hm,

I think I've installed a wrong version
https://wj32.org/wp/software/process-hacker/

This one is version 3.0.105

All the others have 2.39 version.
https://processhacker.sourceforge.io/
I hope I'm not hacked :-/
godly is offline   Reply With Quote
Old 11-19-2019, 03:14 PM   #33
godly
Human being with feelings
 
godly's Avatar
 
Join Date: Sep 2013
Location: Belgium - Gent
Posts: 702
Default

Quote:
Originally Posted by Pictus View Post
The FREE Process Hacker BETA can change the "Processes Affinity", "Priority" and "I/O Priority".
In the Process Hacker "Processes" TAB, right click to change things.

Process Hacker must be running for the "Processes Affinity" settings(when saved) to be
applied, but not for the Priority, Process Hacker can start at logon.
are you sure this is the correct version?
I also wasn't able to shut down my windows after a while. Hard resetting didn't solve it either, only removing the software worked.

Last edited by godly; 11-20-2019 at 06:57 AM.
godly is offline   Reply With Quote
Old 11-19-2019, 04:45 PM   #34
JHughes
Human being with feelings
 
JHughes's Avatar
 
Join Date: Aug 2007
Location: Too close to Charlotte, NC
Posts: 3,531
Default

In reply to my post about Process Lasso:

Quote:
Originally Posted by Cableaddict View Post
Well, It's working for me right now.

- And an AMAZING difference in low-latency performance. I can't even believe it.....

Note: this is on an "AUDIO ONLY" Pc. No internet, no video, no games ..... YMMV.
Thank you for trying it. Have you tried it yet on a laptop too? I don't think it's necessary to go into the BIOS and adjust C-States with this software.

AND are you in Performance Mode?

Windows 10's Game Mode is a subset of what this program does with Performance Mode which initially was called Game Mode before Windows came out with theirs. https://bitsum.com/docs/pl/gaming/
So Process Lasso has implemented certain things for a long time that Windows eventually adopted.

And I don't know why anyone, especially with a laptop, wouldn't try the free ParkControl they (Jeremy Collake that is) make. https://bitsum.com/parkcontrol/

I got a lifetime license real cheap on some site they do specials on once in a while.

Quote:
Bitsum Highest Performance Mode
Process Lasso’s Performance Mode induces the ‘Bitsum Highest Performance’ power plan that keeps your CPU ready to execute code at all times. This eliminates latency otherwise encountered while bringing the CPU out of a low power state. This especially benefits bursting CPU loads, which are the most common real-world CPU use pattern.

With ParkControl, we revealed hidden CPU settings that control core parking, and wrote about how CPU core parking and frequency scaling can affect performance of real-world CPU loads. Put simply, these power saving technologies come with a performance trade-off, so they should be disabled when maximum performance is desired.

Both ParkControl and Process Lasso offer a power profile, Bitsum Highest Performance, that is pre-configured for ultimate performance. In this power plan, your CPU always remains ready to execute new code. Core parking is disabled and the CPU never drops below its nominal (base) frequency.

Since you probably don’t want to be in this power plan all the time, we include automation to switch the active power plan when specific applications or games are running (Performance Mode), or only when the user is active (IdleSaver).

Process Lasso also allows for specific power profiles to be associated with an application in case you want to use different power plans.

Finally, the IdleSaver feature of Process Lasso will switch to a more conservative power plan when you go idle. Similarly, ParkControl has a function called Dynamic Boost that is essentially the opposite of IdleSaver – it raises to a more aggressive power plan when the system is active.

BHP vs AMD Ryzen High Performance
Some users have asked about the new AMD High Performance power plan on Ryzen systems with AMD’s Ryzen Master software installed.

First, you can opt to not use Bitsum Highest Performance (BHP), and instead use AMD’s power plan(s). To do this, associate a different power plan with Performance Mode via ‘Options / Power / Performance Mode settings / Select power plan to use …‘

On analysis, we found the only difference between BHP and AMD Ryzen High Performance in AC mode (not battery) is that Processor Performance Autonomous Mode is enabled in the AMD power plan. Note that this is already the default setting of BHP for Intel systems, and potentially for some Ryzen systems since Autonomous Mode may already be enabled in the system High Performance power plan that BHP starts a clone from, thus inheriting the value from it.

This single difference matters less in a performance power plan than it does in a Balanced power plan. When you trying to both conserve power *and* get good performance, there is a lot more tuning that is necessary to achieve that balance. When you are running ‘all out’, there is less to tune.

Power Setting GUID: 8baa4a8a-14c6-4451-8e8b-14bdbd197537 (Processor performance autonomous mode)
GUID Alias: PERFAUTONOMOUS
Possible Setting Index: 000
Possible Setting Friendly Name: Disabled
Possible Setting Index: 001
Possible Setting Friendly Name: Enabled
Current AC Power Setting Index: 0x00000001
Current DC Power Setting Index: 0x00000000

Enabling this attribute in BHP will make it identical with AMD’s HP power plan. You can make this alteration by switching to the BHP power plan and running the command:

powercfg.exe -setacvalueindex scheme_current sub_processor 8baa4a8a-14c6-4451-8e8b-14bdbd197537 1
__________________
You can only Reverse, Switch, Invert or Flip POLARITY, not "PHASE".
JHughes is offline   Reply With Quote
Old 11-19-2019, 05:37 PM   #35
JHughes
Human being with feelings
 
JHughes's Avatar
 
Join Date: Aug 2007
Location: Too close to Charlotte, NC
Posts: 3,531
Default

Reading through the thread a bit more, I see a need to address a common and well deserved concern:

Process Lasso does adjust priorities on the fly but it's damn smart. You can in no way compare it to, say, permanently boosting the priority of your favored exe. I don't dig into it because you can screw things up as easily as any other app that messes with processes. All it does is transparently keep anything from hogging your CPU's time.

I think of it as: A million people are smoothly entering the highway at high speed. Then some old grandpa wants to take a bit too long and cause a clog on the ramp. So, Process Lasso makes him slow down just enough to keep everyone moving, and nobody feels the difference, not even grandpa!
Get the CPUEater test at https://bitsum.com/how-probalance-works/#skeptics

Quote:
Misconception #3: A user should set default (persistent) priorities on their processes to rank their importance
NO. You should NOT re-prioritize all your processes, ranking them in importance to you. Doing that is risky and detrimental to PC performance. Just let Process Lasso's ProBalance algorithm do its job.

Misconception #4: Other optimization software increases the priority of the foreground process. Does that achieve the same effect?
Absolutely not! These authors only want to sell their wares. They don't care about efficacy. If they did, they'd know that foreground boosting is already done by Windows and additional foreground boosting is ineffective and detrimental to PC performance. Do NOT trust any application that does this. They do it because it's the most simple thing to implement, and it sounds good, even though it is terrible in practice. Also, note that raising a process's priority class does *not* make it run faster.

Process Lasso's ProBalance has been carefully tuned for over a decade in the real-world. Our algorithm works, and we've proven that with our CPUEater Demo and real-world examples like Sony Vegas. Best of all, the conservative nature of our algorithm, along with it's very marginal adjustments, ensure it's safe, as well as effective. Don't trust your PC to bad imitators!
__________________
You can only Reverse, Switch, Invert or Flip POLARITY, not "PHASE".
JHughes is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 08:49 PM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2019, vBulletin Solutions Inc.