Author Topic: Disk Control has never been implemented?  (Read 4221 times)

Offline Gianluigi Salvi

  • New
  • *
  • Posts: 1
  • Gender: Male
Disk Control has never been implemented?
« on: February 07, 2019, 02:23:46 PM »
Around version 7

@BenYeeHua made a lot of very interesting proposal:

Quote
3.Monitoring the Disk and/or SATA controller usage/active time of the Processes
I found that most people getting better CPU performance then the Hard Disk, and they always care about Disk usage/active time.
If Process Lasso support Disk usage, the Windows user that is not using Windows 8+ Task Manager will love to see this feature.

4.Support restraint of the Processes that monopolize the Disk/controller.
As about, and I found that most stutter is caused by the Hard disk is monopolized by some heavy read/write process, like game patch/validate.
If Process Lasso support restraint of I/O monopolize Processes, it will be a good feature too. :)

5.Better Responsiveness of Process Lasso/GUI while I/O is getting monopolized
I always found that Process Lasso GUI is laggy while opening it from Tray Icon, while I/O is monopolized by other processes.
Maybe stop trimming Process Lasso and Process Governor, and lock the pages in RAM should be enough?
http://msdn.microsoft.com/en-us/library/windows/desktop/aa366895%28v=vs.85%29.aspx

You can find them here:
https://bitsum.com/forum/index.php/topic,4399.msg15542.html#msg15542

Did they were taken into account?



Offline edkiefer

  • Volunteer User Moderator
  • Member#
  • *****
  • Posts: 1766
  • Gender: Male
Re: Disk Control has never been implemented?
« Reply #1 on: February 07, 2019, 06:32:55 PM »
Around version 7

@BenYeeHua made a lot of very interesting proposal:

You can find them here:
https://bitsum.com/forum/index.php/topic,4399.msg15542.html#msg15542

Did they were taken into account?
Hi, Thanks for bringing that thread up, it shows how much PL has improved since then.
On the topic of Hrad disk/I/O restraint, PL does have now I/O priority and it can lower I/O priority when Probalance restrains a process (Under advanced settings).

That said, in my experience, it is a tough call to restraint on I/O because a lot of times you want I/O to go as fast as it can, cause it will probably be the slowest area of the system.

For example, say you got a program scanning HD (like AV) unless the CPU usage goes way high most times it be best to let it finish its scan and if the process does go high on CPU usage Probalance will handle that part.

So it can be tricky for the best thing to do here IMO.
« Last Edit: February 18, 2019, 09:30:56 AM by edkiefer »
Bitsum QA Engineer

Offline Coldblackice

  • Member
  • ***
  • Posts: 86
  • Gender: Female
Re: Disk Control has never been implemented?
« Reply #2 on: September 28, 2019, 11:26:47 PM »
Hi, Thanks for bringing that thread up, it shows how much PL has improved since then.
On the topic of Hrad disk/I/O restraint, PL does have now I/O priority and it can lower I/O priority when Probalance restrains a process (Under advanced settings).

That said, in my experience, it is a tough call to restraint on I/O because a lot of times you want I/O to go as fast as it can, cause it will probably be the slowest area of the system.

For example, say you got a program scanning HD (like AV) unless the CPU usage goes way high most times it be best to let it finish its scan and if the process does go high on CPU usage Probalance will handle that part.

So it can be tricky for the best thing to do here IMO.
Around version 7

@BenYeeHua made a lot of very interesting proposal:

You can find them here:
https://bitsum.com/forum/index.php/topic,4399.msg15542.html#msg15542

Did they were taken into account?

Hi, Thanks for bringing that thread up, it shows how much PL has improved since then.
On the topic of Hrad disk/I/O restraint, PL does have now I/O priority and it can lower I/O priority when Probalance restrains a process (Under advanced settings).

That said, in my experience, it is a tough call to restraint on I/O because a lot of times you want I/O to go as fast as it can, cause it will probably be the slowest area of the system.

For example, say you got a program scanning HD (like AV) unless the CPU usage goes way high most times it be best to let it finish its scan and if the process does go high on CPU usage Probalance will handle that part.

So it can be tricky for the best thing to do here IMO.

Bumping this, as well: is it possible to monitor (and possibly control) for this? The shared usage of I/O, namely, where the bottleneck lies -- SATA/storage controllers access -- with programs fighting over each other for their share of the pie. My machine will sometimes get bogged down where it feels like responsiveness is 20-40%, but Process Lasso reports 100%. It makes sense that this is because it's likely an I/O/disk/controller bottle-necking.

If not, any suggestions on how better to analyze this + control for it would be great. And analyzing better than the crude way I know to do it, which is opening a task manager and sorting by disk read/write bytes to see who's at the top. This won't take into account many brief/small/momentary accesses that could be causing the issue, I believe (correct me if I'm wrong).

Offline edkiefer

  • Volunteer User Moderator
  • Member#
  • *****
  • Posts: 1766
  • Gender: Male
Re: Disk Control has never been implemented?
« Reply #3 on: September 29, 2019, 07:08:43 AM »
Process Explorer is the best that I can think of, it should be able to give you performance graphs of the process.

One thing to remember, many AV type processes PL can not alter these process so keep that in mind. Any I/O from them is best to handle it another way.
The best thing I can think is to first try lower CPU priority of the process and see how results go, then try I/O priority.
You will just need to test as I can't think of any other way.

Here a possible example: (I have not tried this yet). I found searchindex.exe was running when I didn't want it to (I always thought these maintenance type processes are supposed to run in background only, idle times). So this process uses I/O to index files that get changed, it is a background type process and might work well with lower I/O but again I would try CPU priority first.

Above is just example of a process that might work, but as I mentioned above always try other ways, in this example for me I removed the games folders that had indexing set in indexing advanced settings, that stopped indexing running while playing those games.
Bitsum QA Engineer

Offline Coldblackice

  • Member
  • ***
  • Posts: 86
  • Gender: Female
Re: Disk Control has never been implemented?
« Reply #4 on: October 17, 2019, 06:35:40 PM »
Thanks, I can try that. The issue, however, is when there are lots of separate/individual/little programs that are all accessing disk I/O at the same time, so it's not possible to find or target just one main offender, but rather, cumulatively altogether. I wonder if there's a better way or software to see this, if your disk controller is choking on all the different I/O requests.

Offline edkiefer

  • Volunteer User Moderator
  • Member#
  • *****
  • Posts: 1766
  • Gender: Male
Re: Disk Control has never been implemented?
« Reply #5 on: October 17, 2019, 06:54:02 PM »
Under those conditions, I would make sure the OS is on the fastest hardware I can afford. So, in this case, a good SSD for OS.
Other than that CPU priority probably would work but it depends on the type of app.
Bitsum QA Engineer