Low CPU utilization & Low FPS

[ghostrobobh 1]

do you have all of your cpu's un-parked?

[EXpMiNi 1]

You are not very forgiving, I'm not on a console, for my case I would like to have 50fps min, especially with the config I have. The Arma's engine is technically not very impressive and I'm pretty sure the polycount it's not so high.

Edited March 14, 2013 by EXpMiNi

[ginger mcale 11]

... unless everyone expected us to play nothing but Blitkrieg PVP with no AI. :p

Ha yeah, but even without AI the we got the issue. Ive monitored the same low fps behaviour (as reported in this thread) also while playing Blitzkrieg. I hope this issue will get its needed Attention. Thank you!

[MaverickFerran 1]

Well, They have defintly improved something with the new Develeoper Build.

As i wrote before, i got 22 fps in the Helicopter Showcase on my Q6600@ 3Ghz,

Now i am getting at least 30, dropped one Moment to 28. So thats an 36% increase.

And very well playable on a 6 year old Quad Core.

People with newer Sandy and Ivy Bridge, that still get low FPS should maybe adjust their Settings, it should be easy for them to reach 40+ fps at all Times, which is enough for ARMA.

even in MP my fps are never lower than 30 fps on good servers.

[white 1]

Well, They have defintly improved something with the new Develeoper Build.

As i wrote before, i got 22 fps in the Helicopter Showcase on my Q6600@ 3Ghz,

Now i am getting at least 30, dropped one Moment to 28. So thats an 36% increase.

And very well playable on a 6 year old Quad Core.

People with newer Sandy and Ivy Bridge, that still get low FPS should maybe adjust their Settings, it should be easy for them to reach 40+ fps at all Times, which is enough for ARMA.

even in MP my fps are never lower than 30 fps on good servers.

20-22 fps when alpha came out, 20-22 fps in the test i made 1min ago.

[EmirSc 10]

Well, They have defintly improved something with the new Develeoper Build.

As i wrote before, i got 22 fps in the Helicopter Showcase on my Q6600@ 3Ghz,

Now i am getting at least 30, dropped one Moment to 28. So thats an 36% increase.

And very well playable on a 6 year old Quad Core.

People with newer Sandy and Ivy Bridge, that still get low FPS should maybe adjust their Settings, it should be easy for them to reach 40+ fps at all Times, which is enough for ARMA.

even in MP my fps are never lower than 30 fps on good servers.

new Develeoper Build?, was an update or something?

[aop 1]

Well, They have defintly improved something with the new Develeoper Build.

As i wrote before, i got 22 fps in the Helicopter Showcase on my Q6600@ 3Ghz,

Now i am getting at least 30, dropped one Moment to 28. So thats an 36% increase.

And very well playable on a 6 year old Quad Core.

People with newer Sandy and Ivy Bridge, that still get low FPS should maybe adjust their Settings, it should be easy for them to reach 40+ fps at all Times, which is enough for ARMA.

even in MP my fps are never lower than 30 fps on good servers.

Most likely mission specific change, since this setup provides exactly the same fps as in the last build:

http://cloud-2.steampowered.com/ugc/882976114553809006/3CFBA4DB4C83B2CD576067036DF13442E7B69A3F/

[wasabi 1]

-FPS still that low in General in both MP and SP, doesnt matter if on High or Standart

-FPS still that low as soon as a few AI´s are active

-FPS still as horrible when looking at a pixelated washed up little village with ~10 Houses(Infrantry Showcase)

-Heli showcase maybe 3 FPS better yet all i saw that they changed is that a few AI´s go idle so the FPS wont go numnuts

-GPU/CPU Utilization still as bad

Thats it for me, ill be heading to other Games till i see an improvement

Also here is a good Video;

Edited March 14, 2013 by wasabi

[MavericK96 0]

What we really need is a legit benchmark mission like in ArmA 2 so we have consistent results between users. It would be ideal if it somehow tracked CPU and GPU utilization as well, so we could send this information to BIS directly.

[tremanarch 6]

If your CPU is at 20% doesnt mean it has headroom. Because that % is for a time intervall. In this intervall of 1 second maybe the CPU was needed only for some nanoseconds and was at 100%.

read:

For systems where power consumption and heat are not an issue, but maximum performance is, it may be desirable to disable CPU Parking. The fact is that even if unparking is near instantaneous, CPU utilization usually occurs in very brief bursts, something I mention frequently. If you look at your CPU utilization and see 15%, that is actually the percentage of time the CPU was active within a relatively large interval (usually 1 second). Actual CPU utilization most often occurs in micro-bursts where the CPU is fully consumed for a few nanoseconds or microseconds. You want those bursts to execute as fast as possible, and this is why all power saving technologies, including frequency scaling, incur *some* performance hit. This is also why a faster CPU does matter, even if you typically don't utilize 100% of available CPU time over a larger interval.

I often watch in the Resource Monitor as it parks cores while system load is nearing 50% of total CPU time. This may be why they provided the newer, and even less documented, option "Processor performance core parking over utilization history decrease factor". This option has *something* to do with how aggressively the Windows scheduler will park CPUs based on over-utilization of parking in the past. In other words, if it has been parking them too often, it will 'ease up'. This *additional* value related to core parking has the description: "Specify the threshold above which a core is considered to have had significant affinitized work scheduled to it while parked". Sadly, since there is no documentation on this power option, I am not yet going to comment more on it - yet. I will say it defaults to 2 and has a maximum of 1000. There are also additional new values such as "Processor performance core parking increase time". I'll leave it to the reader to explore these undocumented advanced values at their own risk!

source: http://bitsum.com/about_cpu_core_parking.php

[daze23 1]

Also here is a good Video;

that's kinda what I noticed. I could bolster my gpu usage by turning Post Processing up to 'ultra', but my framerate wouldn't change at all

[white 1]

If your CPU is at 20% doesnt mean it has headroom. Because that % is for a time intervall. In this intervall of 1 second maybe the CPU was needed only for some nanoseconds and was at 100%.

read:

For systems where power consumption and heat are not an issue, but maximum performance is, it may be desirable to disable CPU Parking. The fact is that even if unparking is near instantaneous, CPU utilization usually occurs in very brief bursts, something I mention frequently. If you look at your CPU utilization and see 15%, that is actually the percentage of time the CPU was active within a relatively large interval (usually 1 second). Actual CPU utilization most often occurs in micro-bursts where the CPU is fully consumed for a few nanoseconds or microseconds. You want those bursts to execute as fast as possible, and this is why all power saving technologies, including frequency scaling, incur *some* performance hit. This is also why a faster CPU does matter, even if you typically don't utilize 100% of available CPU time over a larger interval.

I often watch in the Resource Monitor as it parks cores while system load is nearing 50% of total CPU time. This may be why they provided the newer, and even less documented, option "Processor performance core parking over utilization history decrease factor". This option has *something* to do with how aggressively the Windows scheduler will park CPUs based on over-utilization of parking in the past. In other words, if it has been parking them too often, it will 'ease up'. This *additional* value related to core parking has the description: "Specify the threshold above which a core is considered to have had significant affinitized work scheduled to it while parked". Sadly, since there is no documentation on this power option, I am not yet going to comment more on it - yet. I will say it defaults to 2 and has a maximum of 1000. There are also additional new values such as "Processor performance core parking increase time". I'll leave it to the reader to explore these undocumented advanced values at their own risk!

source: http://bitsum.com/about_cpu_core_parking.php

or the game running mostrly in 1 core and bottlenecking everything else that happens in any other core do make then underutilized, because theres too much shit going on in the first core to begin with. solution? takign all the crap in the first core and spreading over to diff cores. iver only ever seen the cores parking when they were doing nothing for a while, and i had never seen cores park on other multithreaded games/software.

but like was already posted in this same topic i bet the answer would be something like the answer Soma (lead programmer) gave to a bug report before rejecting it about fps and AI 2 years ago:

"I am sorry to disappoint you, but some things are really too hard and we are not willing to spend the time on them." - Soma

dev-heaven.net/issues/6963

[k3lt 3]

"I am sorry to disappoint you, but some things are really too hard and we are not willing to spend the time on them." - Soma

dev-heaven.net/issues/6963

This is disturbing... :butbut:

[Raklodder 1]

20-22 fps when alpha came out, 20-22 fps in the test i made 1min ago.

Sir, I concur, sir, still ZERO performance improvements or CPU-optimization, sir!

[stk2008 14]

See this comment

"I am sorry to disappoint you, but some things are really too hard and we are not willing to spend the time on them." - Soma

I could understand for arma2 it was out befor quad cores etc where mainstream but there is no ascuse if arma 3 is the same.

[WalkerDown 1]

Did they fired Soma after that sentence? lol

[JuggernautOfWar 1]

See this comment

"I am sorry to disappoint you, but some things are really too hard and we are not willing to spend the time on them." - Soma

I could understand for arma2 it was out befor quad cores etc where mainstream but there is no ascuse if arma 3 is the same.

Where did he say this?

[k3lt 3]

Where did he say this?

https://dev-heaven.net/issues/6963

#28

[wasabi 1]

Reporting in ... Getting insane micro stuttering here since the patch .. hope for a fix!

[JuggernautOfWar 1]

So what Suma said was about ArmA 2 two years ago. How does that apply here?

[white 1]

So what Suma said was about ArmA 2 two years ago. How does that apply here?

AI/FPS issues on a bug report that got rejected

didnt go into detail as to why, but basically said they wouldnt fix it because it was too hard.

and well, its the same engine.

[Dwarden 1125]

let me spread some 'light' into 'darkness' :) the low cpu utilization most likely (nearly in fact) has nothing to do with low FPS

so while some of the concerns and complains about MT (multithreading) are valid they not exactly linked to actual low FPS in Arma 3 Alpha

[cenkcnk 0]

i get good fps but heavy stutters

[Reuben5150 2]

let me spread some 'light' into 'darkness' :) the low cpu utilization most likely (nearly in fact) has nothing to do with low FPS

so while some of the concerns and complains about MT (multithreading) are valid they not exactly linked to actual low FPS in Arma 3 Alpha

Example-

Open editor and add 1 unit = 70fps, add 20- 30 AI = 30 fps or less.

What else can i assume if this is not the issue ?

[Giganteus 1]

let me spread some 'light' into 'darkness' :) the low cpu utilization most likely (nearly in fact) has nothing to do with low FPS

so while some of the concerns and complains about MT (multithreading) are valid they not exactly linked to actual low FPS in Arma 3 Alpha

Hi, Dwarden. Can you shed any light on what may be the cause?