urgency: Exactly what you consider it to be. I am providing information, not setting urgency! There'll still be air tomorrow.

BOINC version: 7.6.22 (x64)

OS: Windows 10, up-to-date

processor: AMD A6-4400M APU with Radeon(tm) HD Graphics 2.70 GHz (2 CPUs) No speed demon.

projects: I haven't detected this to be a project-specific condition. I process mostly SETI@Home, Einstein, and Rosetta for now.

installed as a service: no

condition: Something grabs too many of the CPU cycles in a row; I can't interact with the computer for as many as (say) five seconds (inconvenient when editing a document, especially when using the mouse).

my attempts: BOINC Manager >> Options >> Computing Preferences, then (1) limit Usage limits >> Use at most x% of the CPUs, or (2) limit Usage limits >> Use at most x% of CPU time, or (3) limit When to suspend >> Suspend when non-BOINC CPU usage is above x%, or (4) disabling BOINC :( . (That shouldn't be a consideration!)

discussion:
Only disabling BOINC was successful. Then, according to the Task Manager, CPU utilization went way down and my computer was more responsive to my inputs. (Duh!) I'm not the smartest person on the planet nor the most knowledgeable person about BOINC, of course, but I think a prime candidate for the condition behind the problem is that all of these controls limit BOINC without considering the other tasks on the computer.

Maybe it'd be nifty if we had a control for "Limit BOINC so that all tasks use at most x% of CPU time". It seems like that'd really use only excess cycles on machines. The user whose interface needs peak at 2% might set this control for 98%. Peak needs of 5%? Set this control for 95%. The machine always feels responsive; BOINC always gets lots of cycles. Happiness.

Back to the smartest person thing: Maybe it's not possible to do; I doubt I'm the first person to have this idea. Or maybe it's entirely a project thing and all project developers need to relinquish the CPU more frequently than they do (which they don't want).

On the other hand, I think something has changed. I don't remember experiencing this condition in past years.

Still a BOINC Enthusiast! Thanks for reading and for any response.

event log: blank lines between the several groups here; I'm trying to avoid at least some of the repetition
7/24/2016 10:33:49 PM | | Suspending computation - CPU is busy
7/24/2016 10:34:09 PM | | Resuming computation
7/24/2016 10:34:19 PM | | Suspending computation - CPU is busy
7/24/2016 10:34:39 PM | | Resuming computation
7/24/2016 10:34:49 PM | | Suspending computation - CPU is busy
7/24/2016 10:35:09 PM | | Resuming computation
7/24/2016 10:35:29 PM | | Suspending computation - CPU is busy
7/24/2016 10:35:38 PM | | General prefs: from http://spin.fh-bielefeld.de/ (last modified 09-Nov-2007 21:39:39)
7/24/2016 10:35:38 PM | | Host location: none
7/24/2016 10:35:38 PM | | General prefs: using your defaults
7/24/2016 10:35:38 PM | | Reading preferences override file
7/24/2016 10:35:38 PM | | Preferences:
7/24/2016 10:35:38 PM | | max memory usage when active: 2129.24MB
7/24/2016 10:35:38 PM | | max memory usage when idle: 3193.86MB
7/24/2016 10:35:38 PM | | max disk usage: 22.31GB
7/24/2016 10:35:38 PM | | suspend work if non-BOINC CPU load exceeds 15%
7/24/2016 10:35:38 PM | | (to change preferences, visit a project web site or select Preferences in the Manager)
7/24/2016 10:35:39 PM | | Resuming computation
7/24/2016 10:35:49 PM | | Suspending computation - CPU is busy
7/24/2016 10:36:10 PM | | Resuming computation
7/24/2016 10:36:20 PM | | Suspending computation - CPU is busy
7/24/2016 10:36:40 PM | | Resuming computation
7/24/2016 10:36:50 PM | | Suspending computation - CPU is busy
7/24/2016 10:37:10 PM | | Resuming computation
7/24/2016 10:37:20 PM | | Suspending computation - CPU is busy
7/24/2016 10:37:40 PM | | Resuming computation
7/24/2016 10:37:50 PM | | Suspending computation - CPU is busy

7/24/2016 11:41:19 PM | | Suspending computation - CPU is busy
7/24/2016 11:42:10 PM | | Resuming computation
7/24/2016 11:42:20 PM | | Suspending computation - CPU is busy
7/24/2016 11:43:10 PM | | Resuming computation
7/24/2016 11:43:20 PM | | Suspending computation - CPU is busy
7/24/2016 11:44:20 PM | | Resuming computation
7/24/2016 11:44:30 PM | | Windows is suspending operations
7/24/2016 11:44:30 PM | | Suspending computation - requested by operating system
7/24/2016 11:44:30 PM | | Suspending network activity - requested by operating system
7/25/2016 2:44:36 AM | | Resuming after OS suspension
7/25/2016 2:44:36 AM | | Resuming network activity
7/25/2016 9:02:12 AM | | Windows is resuming operations
7/25/2016 9:06:55 AM | | Resuming computation
7/25/2016 9:07:06 AM | | Suspending computation - CPU is busy
7/25/2016 9:07:26 AM | | Resuming computation
7/25/2016 9:07:36 AM | | Suspending computation - CPU is busy
7/25/2016 9:09:27 AM | | Resuming computation

7/24/2016 11:41:19 PM | | Suspending computation - CPU is busy
7/24/2016 11:42:10 PM | | Resuming computation
7/24/2016 11:42:20 PM | | Suspending computation - CPU is busy
7/24/2016 11:43:10 PM | | Resuming computation
7/24/2016 11:43:20 PM | | Suspending computation - CPU is busy
7/24/2016 11:44:20 PM | | Resuming computation
7/24/2016 11:44:30 PM | | Windows is suspending operations
7/24/2016 11:44:30 PM | | Suspending computation - requested by operating system
7/24/2016 11:44:30 PM | | Suspending network activity - requested by operating system
7/25/2016 2:44:36 AM | | Resuming after OS suspension
7/25/2016 2:44:36 AM | | Resuming network activity
7/25/2016 9:02:12 AM | | Windows is resuming operations
7/25/2016 9:06:55 AM | | Resuming computation
7/25/2016 9:07:06 AM | | Suspending computation - CPU is busy
7/25/2016 9:07:26 AM | | Resuming computation
7/25/2016 9:07:36 AM | | Suspending computation - CPU is busy
7/25/2016 9:09:27 AM | | Resuming computation

7/25/2016 2:33:35 PM | | Suspending computation - CPU is busy
7/25/2016 2:33:55 PM | | Resuming computation
7/25/2016 2:34:05 PM | | Suspending computation - CPU is busy
7/25/2016 2:34:55 PM | | Resuming computation
7/25/2016 2:35:06 PM | | Suspending computation - CPU is busy
7/25/2016 2:35:10 PM | | General prefs: from http://spin.fh-bielefeld.de/ (last modified 09-Nov-2007 21:39:39)
7/25/2016 2:35:10 PM | | Host location: none
7/25/2016 2:35:10 PM | | General prefs: using your defaults
7/25/2016 2:35:10 PM | | Reading preferences override file
7/25/2016 2:35:10 PM | | Preferences:
7/25/2016 2:35:10 PM | | max memory usage when active: 2129.24MB
7/25/2016 2:35:10 PM | | max memory usage when idle: 3193.86MB
7/25/2016 2:35:10 PM | | max disk usage: 22.21GB
7/25/2016 2:35:10 PM | | suspend work if non-BOINC CPU load exceeds 18%
7/25/2016 2:35:10 PM | | (to change preferences, visit a project web site or select Preferences in the Manager)
7/25/2016 2:35:26 PM | | Resuming computation
7/25/2016 2:35:36 PM | | Suspending computation - CPU is busy
7/25/2016 2:35:56 PM | | Resuming computation
7/25/2016 2:36:06 PM | | Suspending computation - CPU is busy
7/25/2016 2:36:26 PM | | Resuming computation

7/25/2016 8:42:38 PM | | Host location: none
7/25/2016 8:42:38 PM | | General prefs: using your defaults
7/25/2016 8:42:38 PM | | Reading preferences override file
7/25/2016 8:42:38 PM | | Preferences:
7/25/2016 8:42:38 PM | | max memory usage when active: 2129.24MB
7/25/2016 8:42:38 PM | | max memory usage when idle: 3193.86MB
7/25/2016 8:42:38 PM | | max disk usage: 22.21GB
7/25/2016 8:42:38 PM | | suspend work if non-BOINC CPU load exceeds 100%
7/25/2016 8:42:38 PM | | (to change preferences, visit a project web site or select Preferences in the Manager)

---

Are by any chance using GPU to process tasks? These would show up in Manager as something like "Running (x CPUs + 1 AMD GPU)". Processing tasks on GPU often results in GUI lags.

You can set BOINC to suspend GPU tasks while you are using the computer in Computing preferences. Or if it is Seti@home's app that creates the GUI lags you can go to their Number Crunching forum and ask how to configure their app to reduce the GUI lags.

With a setting this low

7/24/2016 10:35:38 PM | | suspend work if non-BOINC CPU load exceeds 15%

You will find your system constantly pausing. This may even happen if you have antivirus running. I would start with making two changes.
1. bump that % to something a bit higher. I typically set mine at 85%. Now you may think that this is counter productive. However, if the client is constantly dumping files from memory to the HDD, that can slow your system down as it offloads.

2. If your system is too sluggish leaving it at the above setting, try reducing the number of cores you allocate for BOINC to use.

---

Juha: I'm using GPU and CPU to process.

Coleslaw: I like your 85% suggestion; I've set that. (When experimenting, I repeatedly went down by 10% and saw the increase in pausing, but only slight increase in responsiveness. Then I went by 5% each time until I got to 15%. It helped, but never enough; below around 85%, it didn't seem right.)

new info:
When I finished my message last night, I set all three limits I mentioned to 100% (not limited). The sluggishness is gone. Hmmm ...

My projects now are from Einstein and Rosetta. Einstein on GPU, others on CPU.

I'm pretty sure that as I wrote, I had a project from SETI. It must have completed.

Maybe I can no longer say, "I haven't detected this to be a project-specific condition."

Thanks to both for the useful advice. I'll monitor for new tasks and see what I learn. I'll check the SETI Number Crunching forum now!

Juha: Apparently, it will never cease to amaze me that the world has grown so much smaller in my short life. I have transitioned from being entirely dependent on physical mail and weeks of delay to, in this case, getting help overnight from an expert in Finland, judging by the flag on your account. Amazing. May our smaller world also live in peace for the future!

---

Sluggishness can come from 3 parts, at least on my systems. In order of cause:

1) GPGPU computing while system is in use [duh, because if not in use... ;o]. Set GPGPU computing to only be done while system is not used.
2) Very large intermediate checkpoint file and model builds taxing the HDD to work hard (e.g. WCG's CEP2 is infamous for this, with up to 2.5 GB written). Some found solution in RAMdrives and SSD.
3) Very large result file uploading (I've set it to 1 at a time per project on the cc_config to minimize concurrent uploading). Often suffer from poor ISP/Cross continental file transfer crawling from my outback.

---

Coelum Non Animum Mutant, Qui Trans Mare Currunt

3) Very large result file uploading (I've set it to 1 at a time per project on the cc_config to minimize concurrent uploading). Often suffer from poor ISP/Cross continental file transfer crawling from my outback.

A good quality (wired) network connection should only require a trivial amount of CPU support from the host device - no matter how much buffering is being caused by a slow upload link.

But some - especially WiFi - connections are notorious for introducing latencies via badly-written wireless drivers. In my experience, USB WiFi dongles are particularly bad.

First question, about sharing CPUs: Please help educate. How does BOINC work?

Right now, I have two Rosetta tasks and one SETI task all running. The Rosettas are apparently CPU-only. The SETI task says it needs 0.0333 CPUs + 1 AMD/ATI GPU. Mine is a 2-CPU processor.

I'm being told it's obvious I'm overtasking my system because every GPU task requires a whole CPU, so I have the equivalent of 3 CPUs of demand on my system from BOINC alone.

I'm concerned that is counter to the SETI task claiming to need only 0.0333 CPUs with the GPU.

Maybe it could be argued that I'm still overtasking my system because I have 2.0333 CPUs of demand on my system from BOINC alone.

I'm concerned that is counter to the concept of BOINC. I understand BOINC to try to use otherwise unused cycles without causing any problems on the system. I further understand it runs at lower-than-normal priority so that all "normal" processing preempts it, thus giving it less than the 2 CPU capability of my system.

What's true, please?


Second question, about sharing the GPU: I get it; I'm getting lots of advice to not allow use of the GPU while I'm using the computer. I'm considering it.

On the other hand, it seems like the same techniques BOINC uses to share CPUs might be equally applied to the GPU. My experience has been very good until recently in letting tasks have the GPU (or at least a portion of it) while I'm working.

I'm reluctant to make the GPU off limits while I'm working because I put this computer to sleep then. The GPU jobs might have a tough time completing.

Again: How does BOINC work?

Thanks in advance.

---

You are partially right. When you are giving a full CPU core or thread to a work unit, BOINC will still share that core/thread with other processes. How do you think the operating system still works while running BOINC at 100%? So, when your GPU work unit says it needs .0333, it is basically saying that it doesn't need very much CPU time to process the work unit as it is basically feeding the GPU or doing a few calculations the GPU can't do on its own. So, though it says it needs .0333 CPU's, it isn't necessarily using that the entire time.

Second question... not all GPU work units/projects will use the same amount of GPU resources. So, some projects/apps will run smooth while using the PC for basic stuff. Other projects not so much. The capability of the GPU will also factor into that. But yes, BOINC will somewhat use the GPU like the CPU. In sum, BOINC isn't using 100% at any given time in reality because the system shares those resources for other things too.

Last, I don't know who is telling you that GPU work units use a full thread/core, but that isn't true of all projects/work units. It is just advised because many will see increased output of their GPU because it isn't waiting on the CPU to open up that .0333 CPU when it needs it.

---