[REPORTED]VR Performance OB 2.5.6

[imacken]

I finally got round to installing Maple Flag's F18 Aggressor campaign. In the first (Familiarisation) mission, I get:

 

2.5.5 last OB 40 fps (ASW kicking in)

2.5.6 latest OB 5 fps !!!

 

This is on entering the mission, sitting in the cockpit of the powered down aircraft and looking around.

 

Settings are as per my previous test. ( https://forums.eagle.ru/showpost.php?p=4221938&postcount=73 ).

 

This is the first complex (i.e. containing many objects), and Nevada, mission I have tried - and have to agree with many other posters that the frame rate decrease is severe in this case.

 

Yes, a perfect example! Or, look at the C-130 intro flight.

 

There are many examples all over.

[imacken]

Yes, a perfect example! Or, look at the C-130 intro flight.

 

There are many examples all over.

 

It's not just VR. I tested the same scenario on a monitor, and I have gone from a frame limited 58 fps to single figures! (12 in the attachment!)

[Gryzor]

I reverted back to the latest openbeta 2.5.5.41962. I hope some fix in the near future. Having such performance loses its a serious situation (even taking account its a beta branch).

 

Seriously ED, maybe it would wise to hire some expert in code development - software optimization, according reddit post and "spaguetti" code, please consider it the hightest priority now.

 

In my case, I invest more than 700 € since 2011, high end pc in all modules / scenary / campaign etc... I don´t remember having such huge downgrade in performance (appart from graphic corruption in VR like no L eye light rendering) in DCS history. Even alpha 2.0 version were quite similar in performance terms compared with 1.5 3-4 years ago (having now a reduction of 30-40% in performance... in other situations with many unit, to single fps digits, even with my pc high end VR-PC).

 

I hope 2.5.6 would be at least the same performance of 2.5.5 (now, after many years of waiting, I´m not very optimistic in performance optimization). Sincerely, I hope solutions, focused primaly in the core sim.

Edited February 27, 2020 by Gryzor

[hawk4me]

I hear so many issues on VR performance and it's really odd. My FPS went up from 2.5.5 to 2.5.6. I can hold 70-90fps pretty easily in more SP mission's I don't really do big server's in VR due to still getting used to it. It does seem like there are a lot of massive FPS drops depending on where you are looking though I can get 90fps and I look down at my gauges barely any head movement at all and it drops to 45-50fps. I have noticed though that the HUD in the 16 and 18 is pretty much unreadable in VR atm though no matter what AA or MSAA option you turn on. If i run at 16X FXAA and 2X MSAA it's better but then i get like 40 frames most of the time. Still hard to make stuff out on the HUD though.

[Canada_Moose]

It's not just VR. I tested the same scenario on a monitor, and I have gone from a frame limited 58 fps to single figures! (12 in the attachment!)

 

 

 

 

According to other threads / topics it appears that its the CPU thats taking a hammering in 2.5.6, especially when lots of units are present. Far more so than in 2.5.5

 

 

This issue would be magnified in VR of course.

 

 

 

Lets hope they figure it out.

[imacken]

According to other threads / topics it appears that its the CPU thats taking a hammering in 2.5.6, especially when lots of units are present. Far more so than in 2.5.5

 

Yes, exactly that. The CPU usage has gone ballistic in certain situations in 2.5.6, hence the huge drop in fps.

I’m sure ED are on the case.

[Cornelius]

Everyone's always complaining about the DCS not using enough CPU. So here you go.

And now they're all complaining again...

ED really don't has it easy!

:D

Edited February 28, 2020 by Cornelius

[imacken]

Everyone's always complaining about the DCS not using enough CPU. So here you go.

And now they're all complaining again...

ED really don't has it easy!

 

Really? I don’t think so. People complain about DCS being CPU bound on (more or less) one core all the time, hence the need for Vulkan. This current build, however takes that situation to a new level under certain circumstances.

[Cornelius]

Really? I don’t think so. People complain about DCS being CPU bound on (more or less) one core all the time, hence the need for Vulkan. This current build, however takes that situation to a new level under certain circumstances.

 

Sorry, forgot the smiley. Corrected my post.

[Hippo]

OB 2.5.6.44266 - performance remains unchanged (i.e. from worse to a lot worse than 2.5.5).

[Fang333333]

OB 2.5.6.44266 - performance remains unchanged (i.e. from worse to a lot worse than 2.5.5).

 

I don't mean to be dramatic here but..

 

PLEASE BRING BACK DCS!

[speed-of-heat]

I don't want to state the blindingly obvious but roll back to 2.5.5

[sirrah]

OB 2.5.6.44266 - performance remains unchanged (i.e. from worse to a lot worse than 2.5.5).

 

I can confirm

 

Don't want to beat a dead horse, but hey, reporting and confirming issues is what the OB is for :). Performance most certainly didn't get any better for me. First mission I tested it on it appeared to actually got a few fps worse.

 

 

My results on mission "Instant action F-16 cold and dark PG", directly after loading, while looking around in the cockpit:

2.5.5 = somewhere between 45 and 90fps

2.5.6 last week = between 23 and 30fps

2.5.6 today = didn't get above 25fps

 

All tested with same settings, except for water. Had to reduce water setting from medium to low, due to the medium water setting being broken

[Hippo]

I don't want to state the blindingly obvious but roll back to 2.5.5

 

It is not immediately obvious to me what the point is in telling people to roll back to 2.5.5 in a thread where people are reporting issues relating to the latest 2.5.6 open beta.

 

I have just gone to the trouble of running a couple of, admittedly fairly simple, tests, and have found that I get the same reduction (from 2.5.5) in performance that I did in the previous OB. This in a thread relating to performance in 2.5.6, in a forum dedicated to reporting issues with game performance. I don't think "roll back to 2.5.5" is the appropriate response, suggest instead "thanks for beta testing for free".

[Fang333333]

maybe it would wise to hire some expert in code development - software optimization, according reddit post and "spaguetti" code, please consider it the hightest priority now.

 

.

 

agree

[Fang333333]

Hi all

 

Can I ask you to attach track replays to show us which missions are creating problems for you. Please make them as short as possible, but long enough to show the FPS hit you see.

 

thanks

 

 

Bignewy, Sorry no tracks here but specifically anything with AI and objects. I.E Every FA 18 mission for the F-18 Red Flag AGRESSORS campaign.

 

I starts off at Nellis and the FPS is Crippled on tarmac. Go into a new one on one dogfight out in the middle of nowhere and fps is "meh" ok.

 

Strangely when I sit on Tarmac during the Red Flag campaign and I press escape Fps bumps up. Unpause and DCS is broke

[ataribaby]

BTW what happened with once announced VR optimization that should deliver about 50% more FPS in VR? From that time there is silence about that. I remember terrain maps was changes because of this but I now have a lot worse FPS in VR so it is nearly unplayable.

[SkateZilla]

Please Post Settings, if Possible Options.lua in a Zip.

[speed-of-heat]

It is not immediately obvious to me what the point is in telling people to roll back to 2.5.5 in a thread where people are reporting issues relating to the latest 2.5.6 open beta.

 

I have just gone to the trouble of running a couple of, admittedly fairly simple, tests, and have found that I get the same reduction (from 2.5.5) in performance that I did in the previous OB. This in a thread relating to performance in 2.5.6, in a forum dedicated to reporting issues with game performance. I don't think "roll back to 2.5.5" is the appropriate response, suggest instead "thanks for beta testing for free".

 

Wasn't aimed at you

[Gryzor]

BTW what happened with once announced VR optimization that should deliver about 50% more FPS in VR? From that time there is silence about that. I remember terrain maps was changes because of this but I now have a lot worse FPS in VR so it is nearly unplayable.

 

Maybe missed adding a minus sign just beforte the "50%", just to reflect at least the beta reality and 2.5.6 disaster in performance test.

[ataribaby]

Please Post Settings, if Possible Options.lua in a Zip.

 

Here it is.

options.zip

[Caliboy42]

In terms of this thread, I’m not sure I understand what the problem is? I see a lot of posts about massive frame rate loss. 80 to 60? 60-40? 40-20? It’s all different. And a lot of people that complain don’t have their specs posted. Entire first page, no specs. Are you meeting the minimum requirements for DCS? How’s your computer’s memory? Is your hard drive full? Background programs? Are you in VR? Are you recording? Is everything set on high settings? Are you doubling up on Nvidia CP, Oculus tool tray or steam settings? List your specs people so ED can see if the problem is, well, lack of power. Personally? I’m still on high settings (no Nvidia CP) with constant 40 FPS (I can’t tell the difference between 40 and 80 FPS anyway in VR) and it looks fantastic!! Are we asking too much from ED without checking our own systems?

 

Just saying...

[Canada_Moose]

In terms of this thread, I’m not sure I understand what the problem is? I see a lot of posts about massive frame rate loss. 80 to 60? 60-40? 40-20? It’s all different. And a lot of people that complain don’t have their specs posted. Entire first page, no specs. Are you meeting the minimum requirements for DCS? How’s your computer’s memory? Is your hard drive full? Background programs? Are you in VR? Are you recording? Is everything set on high settings? Are you doubling up on Nvidia CP, Oculus tool tray or steam settings? List your specs people so ED can see if the problem is, well, lack of power. Personally? I’m still on high settings (no Nvidia CP) with constant 40 FPS (I can’t tell the difference between 40 and 80 FPS anyway in VR) and it looks fantastic!! Are we asking too much from ED without checking our own systems?

 

Just saying...

 

 

You mean the people that had perfectly acceptable performance on 2.5.5? Nothing has changed on their end except for the update to 2.5.6

Why would anybody be 'doubling up' on any settings all of a sudden?

 

 

 

There has already been evidence posted in other areas of this forum that the CPU is being hit a lot harder in 2.5.6 than in 2.5.5

 

 

Considering that currently DCS is pretty much single threaded and hits one core to 100% all the time, extra drain on the same thread of the CPU would not be a good situation, especially in VR.

 

 

You will find most people on here are not at their first rodeo.......

[Caliboy42]

You mean the people that had perfectly acceptable performance on 2.5.5? Nothing has changed on their end except for the update to 2.5.6

Why would anybody be 'doubling up' on any settings all of a sudden?

 

 

 

There has already been evidence posted in other areas of this forum that the CPU is being hit a lot harder in 2.5.6 than in 2.5.5

 

 

Considering that currently DCS is pretty much single threaded and hits one core to 100% all the time, extra drain on the same thread of the CPU would not be a good situation, especially in VR.

 

 

You will find most people on here are not at their first rodeo.......

I understand that there are a lot of people that have been here a lot longer than I have. What I don’t understand is why I, and others, am not having any FPS issues? 2.5.5 or 2.5.6. I do not see any hits on FPS or a change in single core usage. I also understand that I am fortunate enough to have a powerful system. But in testing, I see no more than 45% cpu usage on MP. I’m currently flying with 8 other pilots on PG map. The other pilots aren’t having an issue either.

 

With that said, I have to question other people’s setups. Saying that this is all ED’s fault is rather hypocritical when there is proof in the forums and more that people are having the same or improved performance since 2.5.6. Looking at it logically, ED is the common denominator in this equation. It’s the consumers that are the variables. Hence all of the differences in performance... so if you can tell me what I’m doing right, I will gladly pass it along to others to try and help.

[JumpinK]

I understand that there are a lot of people that have been here a lot longer than I have. What I don’t understand is why I, and others, am not having any FPS issues? 2.5.5 or 2.5.6. I do not see any hits on FPS or a change in single core usage. I also understand that I am fortunate enough to have a powerful system. But in testing, I see no more than 45% cpu usage on MP. I’m currently flying with 8 other pilots on PG map. The other pilots aren’t having an issue either.

 

With that said, I have to question other people’s setups. Saying that this is all ED’s fault is rather hypocritical when there is proof in the forums and more that people are having the same or improved performance since 2.5.6. Looking at it logically, ED is the common denominator in this equation. It’s the consumers that are the variables. Hence all of the differences in performance... so if you can tell me what I’m doing right, I will gladly pass it along to others to try and help.

 

well... my settings didn't change from 2.5.5 to 2.5.6 but the performance did.

Half of the FPS are gone in VR and ~30% less performance in 2D.

 

(and I run almost the same system as you do.)