Jump to content
Sign in to follow this  
Game__On

6 million "page faults" in 5 minutes. What does this mean ?

Recommended Posts

So i was running "process explorer" , to try and find out why arma craps out on me the longer i play multiplayer.

I saw the "page fault" number increasing with an insane rate, now gonna test other programs but what does it mean ?

Share this post


Link to post
Share on other sites
Generally, making more physical memory available also reduces page faults.

Sounds interesting enough. Why don´t you try slapping double the ammount of RAM you have now in your PC and see if the number of page faults goes down by 50%?

Still struggling to understand the magic behind all this, but from a logical standpoint 3 million operations (of whatever kind) less should leave more room to breathe for ArmA :confused:

PS: pls don´t take this as one of those "buy better hardware you fool" posts that became the lousy norm on these forums lately.

Share this post


Link to post
Share on other sites
Sounds interesting enough. Why don´t you try slapping double the ammount of RAM you have now in your PC and see if the number of page faults goes down by 50%?

Still struggling to understand the magic behind all this, but from a logical standpoint 3 million operations (of whatever kind) less should leave more room to breathe for ArmA :confused:

PS: pls don´t take this as one of those "buy better hardware you fool" posts that became the lousy norm on these forums lately.

Beyond 4gb it wouldn't matter since ArmA 3 is 32bit. Even if you only had 4gb of RAM, ArmA 3 only uses up to 2-2.2gb of physical RAM and never more, so 4gb should be plenty unless you are running lots of programs along side it.

6 million page faults in 5 minutes seems pretty excessive to me, be it hard or soft page faults. A soft page fault is when something is in virtual memory and needs to be paged into physical memory. A hard fault is when something does not reside in virtual memory and needs to then be loaded or paged into virtual memory/physical memory. At least that's been my understanding of it. If the addressing space or working space of the program has been filled, for instance it's at the limit of 32 bit addressing for a program, adding more memory won't help in the least as the working space is already full. The only way to fix it would be to increase the addressing space, for instance 64 bit addressing.

Something to think about...

Major page faults on conventional (hard disk) computers can have a significant impact on performance. An average hard disk has an average rotational latency of 3ms, a seek-time of 5ms, and a transfer-time of 0.05 ms/page. So the total time for paging is near 8ms (8 000 μs). If the memory access time is 0.2 μs, then the page fault would make the operation about 40,000 times slower.

Share this post


Link to post
Share on other sites

Already have 8 gigs .

And arma can only use 2 GB of it .

ROFL , that's just sad.

Share this post


Link to post
Share on other sites

... 6 million per 5 minutes (300sec), are 20.000 per second and this means 20 faults per ms (average) ...

Lets hope this are soft page faults, because hard page faults means i/o reads writes occure.

But soft page faults, even if very much faster (page transition, in memory), they need cpu resources too.

Are there other applications running in the background while you observe this?

Share this post


Link to post
Share on other sites

Uhm, those are the page faults of ARMA 3 only .

Needless to say it seems a bit excessive, as other programs seem to have about 10% of that number.

Share this post


Link to post
Share on other sites

Did a quick test and was averaging 10,000-25,000 page faults per second while playing. I also had spikes up to 60,000 pf/sec.

Share this post


Link to post
Share on other sites

... i see something like 1-3.000 per second, but i use a custom memory allocator, with most pages locked to phys. RAM.

I will compare with default allocator later, 60.000/s seems to be a bit to much.

Share this post


Link to post
Share on other sites

For reference, I tested in multiplayer on a random server playing a CTI or domination mission. I tested in single player as well and I got closer to 8,000 page faults per second on average with spikes up to 35,000.

Also suggest using performance monitor instead of resource monitor. It's much easier to read and much more configurable.

Share this post


Link to post
Share on other sites

how to explain, my system commit ram usage gose up to 8gb from 12gb, when playing arma 3, when it's usualy 3,3gb at idle, when i browsing inet? Btw it's pointless to say "more then 4gb ram won't change anything, cause arma 3 using 2 gb of ram" Not in my case, when my system uses 3gb at idle. So at least 1 gb goes in to page file, slowing down your system and the game itself.

Share this post


Link to post
Share on other sites
For reference, I tested in multiplayer on a random server playing a CTI or domination mission. I tested in single player as well and I got closer to 8,000 page faults per second on average with spikes up to 35,000.

Also suggest using performance monitor instead of resource monitor. It's much easier to read and much more configurable.

... i tried with performance monitor now and got 7-8000 PF/s average, much more as process explorer told me (1-3000). Not sure what tool tell us the truth ...

I think the source for that soft pagefaults, is the frequently mapping/transfering (via MapViewOfFile) of small (64kB) areas, from this big "addressless" system memory area (~ 1.6GB) to arma process address space.

Share this post


Link to post
Share on other sites
... i tried with performance monitor now and got 7-8000 PF/s average, much more as process explorer told me (1-3000). Not sure what tool tell us the truth ...

I think the source for that soft pagefaults, is the frequently mapping/transfering (via MapViewOfFile) of small (64kB) areas, from this big "addressless" system memory area (~ 1.6GB) to arma process address space.

That's why I suggested using performance monitor as ArmA makes heavy use of file mapping which is outside the scope of the process and thus won't be counted by resource monitor or process explorer.

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
Sign in to follow this  

×