MSI afterburner bug report & suggestion

I know and i don't say that my RAM is faulty. I don't know how you came to your conclusion! And no. THIS is not normal (look at CPU time and pagefaults... 105.000 each second on the CPU:3eyes. Check your skill*.

Edit: *reading and understanding my text above :eyes:

 

So your seeing 105,000 page faults for MSI AB.exe , over how long of time ?
I see 490 on my system right now

Are you logging history or something.

 

I'm curious about this too. My PC has been on for about 15 minutes and shows over 2 million and growing. Firefox is doing the same thing.

Not a very clear answer here

http://windows.microsoft.com/en-us/windows/what-task-manager-memory-columns-mean#1TC=windows-7

Internet talks about hard and soft page faults. Will need to investigate.

Alex, any comment?

BTW, not logging any data.

Edit: increasing AB polling time from 1000 to 10000 slows this down. So it must be link to that.

 

Good enough for me, thanks.

 

I meant the skill of reading and understanding a text, because i wasn't talking about defects on my DRAM. I must admit that my sentence was a bit confusing and never meant to hurt everyones skills in programming etc.

So, Unwinder, you say this is the normal behaviour. Than everything is cleared. I had some issues in the past regarding page-faults (e.g. false database API-calls) and it was sometimes an unwanted "feature", so i thought i might report it.

@EdKiefer: see CPU-Time

 

I got the impression that you thought something was wrong with high number of page faults. This is not the case, and I said so. I didn't impy anything else - I'm sorry if you got that from my post.

 

I don't have problem with cpu time but i have set polling rate to 3000 and limit how many data points I monitor .

 

HWinfo now supports VRM FURY temps monitoring.

It would be nice if AB could do the same

Cheers.

 

I've just disabled my Windows Page File in Win7. However, in MSI Afterburner v4.2.0 it says that I'm still using a considerable amount of space for my page file. Is there a way to confirm whether or not my PC is using a page file when I've told it not to in the Windows system settings?

Screenshot:

 

http://forums.guru3d.com/showthread.php?t=405208

 

To add another way , go to C;\ or where root OS is, folder options>view>disable "hide protected operation files" and enable "show hidden files, folders an drives"

Then you should see a file "pagefile.sys , note its size that "should" match what you set for pagefile size in windows setting.

 

Thanks for the link. I looked up "commit charge" and found that in my task manager, it says Commit (GB): 4/31 or 6/31 etc. I have 32GB of RAM, 31 of which I assume is usable. The number over 31 changes alongside the Pagefile Usage graph in MSI AB v4.2.0. When the Commit (GB) says 4/31, AB shows ~4000MB of Pagefile Usage. When the Commit (GB) says 6/31, AB shows ~6000MB of Pagefile Usage. Does this mean that Windows is using a pagefile even though I've told it not to?

Screenshot:

Thanks for the tip. I've done so and there appears to no longer be a pagefile.sys file in the C:\ drive. There was one there before I disabled the Pagefile in Win7, so it appears that disabling it in Windows has done something, but I'm still not certain if all is operating correctly, since MSI AB is reporting high Pagefile usage and Windows Task Manager has a nonzero commit amount (4/31, 6/31, etc)

 

Page about "Commit charge" on Wikipedia (https://en.wikipedia.org/wiki/Commit_charge) says:

(Retrieved 2016-03-01 19:15 UTC)

So if you have 32 GB of physical RAM and no pagefile, the commit charge limit should be a bit under 32 GB. (A small amount of memory will be used by hardware drivers. This can't be paged so it isn't counted towards commit charge.)

 

Ah. I guess I'm still confused as to why MSI AfterBurner is reporting varying pagefile usage, though. Is it because, although Windows is not using a pagefile, the total amount of pageable virtual address space or "commit charge" is still being calculated, even though all of that pageable data is staying in RAM and not going into an actual pagefile?

 

@RudyH246: 100% correct

 

Can someone with a similar graphics card (Fiji / R9 Fury) and operating system (Windows 10 x64 1511) as mine, to run an i2cdump on your card?

When I try, it gets stuck in an infinite loop, never finishing and topping out CPU utilization on one core.

EDIT: Never mind... Takes like 5 minutes to finish, but will eventually complete.

 

Well, with newer Crimson 16.2.1 and W10 cumulative patch 10586.122 this problem seem gone.

Edit: Nope. What solves it is another trick: Enabling unofficial OC mode with powerplay, and then disabling it with an overclock applied. 3D clocks are then badly detected (pressing reset will keep the card at the same OC'd 3D clocks, but with power at 0% instead os 20%).
When AB's "default" clocks are the as the applied OC, the clocks alteration problem when running multiple users is gone.

 

Hi everyone!

I've noticed that there are 5 GPUs listed under Master Graphics Processor selection when using a AMD A10 7870K and its integrated GPU.

Is this behaviour normal or is it a bug?

Crimson 15.12 installed.