Major Performance Issues

[AfterDune]

I just (soft) deleted a whole bunch of posts, because they derailed the thread too much. Let's stay on-topic and try to find the cause(s) of the performance issues.

I'll post this here yet another time, so everyone is happy (hopefully). It's nothing new, but here goes anyway:

We are working on the performance issues too, just like some of you are. We go through pretty much every little change we made ever since we started building v1.0, which takes a sickening amount of time. Know that we're reading a LOT on these forums, especially regarding performance issues. We don't always reply, because there's simply no need to. Just know that we read stuff and talk about your ideas and findings in our internal team and dev forums.

If you think we may have missed something, feel free to poke one of us about it (though we'll probably let you know we already know about it ).

Let's get on with it...

[Murphy]

I'm pretty sure my post was a move to get this thread back on track and yet it was deleted without merit. I am sad to see my pertinent post lost in the wave of flames.

It was about the billows of smoke that run through their animation way to fast to appear "realistic" by any stretch of the imagination. I was wondering if there had been any investigating of the smoke animations being a little intensive thanks to the rate at which they play and restart. I had the idea to test it myself using both Fallujah and a map entirely void of anything excluding said smoke effects. I would like to know if I would be wasting my time because this has been looked into, and if not I would gladly post my results after appropriate tests are run.

[Michael Z Freeman]

I just (soft) deleted a whole bunch of posts, because they derailed the thread too much. Let's stay on-topic and try to find the cause(s) of the performance issues.

I'll post this here yet another time, so everyone is happy (hopefully). It's nothing new, but here goes anyway:

We are working on the performance issues too, just like some of you are. We go through pretty much every little change we made ever since we started building v1.0, which takes a sickening amount of time. Know that we're reading a LOT on these forums, especially regarding performance issues. We don't always reply, because there's simply no need to. Just know that we read stuff and talk about your ideas and findings in our internal team and dev forums.

If you think we may have missed something, feel free to poke one of us about it (though we'll probably let you know we already know about it ).

Let's get on with it...

Thanks for popping in and thanks for giving so much time to PR ! Its an important game for me. So, I've been left a little confused though. You say ...

Let's stay on-topic and try to find the cause(s) of the performance issues.

Boris has done some very detailed testing. I noticed a lack of threads compared to other mods for the PR exe. There are so many variables here and I'm not into regularly developing for Battlefield, but a thread problem would make sense. This is because in Coop/SP AI activity takes my frame rate right down. It can be at "normal" at 60fps on the edge of the map with not much AI activity going on, but goes right down to between 16 and 20 frames when the AI are doing their thing (48 bots standard Coop started from the menu). Now this is normal for the AI started from within the client. I get the same thing with other mods, but not to the crippling extent that it happens in PR 1.0. I need to get hold of PR 0.97 or whatever it was and see if it does the same thing. So have you, as PR devs, looked into the coding of the launcher and the PR Battlefield exe ? Or have you discounted that and are looking at some other issue ?

[AfterDune]

I'm pretty sure my post was a move to get this thread back on track and yet it was deleted without merit. I am sad to see my pertinent post lost in the wave of flames.

Sorry, Murphy, that wasn't supposed to happen. The posts were "soft" deleted, which means they're not visible to the public. Your post has been restored.

So have you, as PR devs, looked into the coding of the launcher and the PR Battlefield exe ? Or have you discounted that and are looking at some other issue?

Yes, we have. We have made changes to the launcher with each patch, to increase performance. Unfortunately the major performance issues still exist. We're still looking into the launcher, but so far the code looks normal and changes we made haven't done much to performance.

We haven't found it has to do with the background processes, like BUIS, compass marker, etc. Every change we made has had no impact and some of us (devs/testers) have pretty old systems too. There used to be issues with that, but now it doesn't seem to be apparent anymore. If you think otherwise, please share (and we'll read).

We've also looked into PRBF2.exe, but so far haven't found anything odd. For those who wonder what PRBF2.exe is, it has some tweaks to show 8-player squads and some other small things. But changing that or replacing it with BF2.exe doesn't seem to have any effect.

Also changing the way Nvidia (for example) looks at PRBF2.exe doesn't seem to have any effect either, so if you ask me, it seems the issues are not with that one. But it never hurts to just try your own thing.

I wonder what others experience when they turn off settings (or set it to the lowest value) like shadows, dynamic lighting, etc. Just shut down one thing at a time, see what effect it has. If it has no effect, turn it back on and turn something else off. If that has no effect, try to shut off multiple things and see what that does.

[Prevtzer]

AfterDune;1958995']I wonder what others experience when they turn off settings (or set it to the lowest value) like shadows, dynamic lighting, etc. Just shut down one thing at a time, see what effect it has. If it has no effect, turn it back on and turn something else off. If that has no effect, try to shut off multiple things and see what that does.

I've tried turning off shadows and setting everything else light related to as low as possible- it made no difference whatsoever- the shadows were still there. The difference between setting everything to max vs. setting everything as low as possible is 5, maybe 10fps. I was averaging 7 fps in 1.0, while I had 40 fps in 0.98.

[-1-Gabe-1-]

I know this might not be the place to say this,,but Im watching the number of players go down every day,,every weekend,,and it really hurts,,,
Playing this new release even tho i had to set most of my settings to medium or low is great,,I still enjoy it,,and I thought it might be a good idea to set back to 64 the number of players,,since,, 99.9% of the players Ive talked to have no problems regards to fps when the server hosts that amount of players or less,,every1 agrees about the higher the number of players,,the the higer the frames drop to unplayable levels...
So,,having said that,,,wouldnt it be wise to set that back untill a solution is found?
I know some of you disagree about being the number of players the issue,,I've tracked all ur tests Boris wich I found to be great and with a great level of expertise...

Just my two cents guys,,something has to be done,,and fast,,

As Im typing this post this is what PRSPY reports Number of servers:4 Number of players:130
Pretty conclusive to me..

Best to all!!

[Heavy Death]

Gabe, we played on 200 slot servers in 0.98 and it was ok with fps.

[-1-Gabe-1-]

Gabe, we played on 200 slot servers in 0.98 and it was ok with fps.

Exactly buddy,,Ive played on 128 servers and it was fine too!! But now on 1.x it doesnt anymore!!

[SShadowFox]

We've also looked into PRBF2.exe, but so far haven't found anything odd. For those who wonder what PRBF2.exe is, it has some tweaks to show 8-player squads and some other small things. But changing that or replacing it with BF2.exe doesn't seem to have any effect.

I wonder what others experience when they turn off settings (or set it to the lowest value) like shadows, dynamic lighting, etc. Just shut down one thing at a time, see what effect it has. If it has no effect, turn it back on and turn something else off. If that has no effect, try to shut off multiple things and see what that does.

And why not just let us run with Lighting on "Low" again? People that are finding hard to play with good FPS even on Low may get some advantage, or do you Devs only plan to do that as last resort?

I play FH2 with Dynamic Lighting and Dynamic Shadows activated (Low) and it runs quite fine, I've noted some FPS gain with they on Medium in PR, but still impossible to play.

[tankninja1]

Maybe some new and strange information here but a few days ago I was playing and was getting massive amounts of lag and was surprised my computer didn't crash and I got this message for the first time "PR Launcher has crashed" causing mumble to exit and not restart. After this launcher crash I noticed a 8fps increase (I went from 10ish before the launcher crash to 19). Also BUIS seems to cause lag when I switch modes there is a 1/2second to 2second pause.

[Inspektura43]

And why not just let us run with Lighting on "Low" again? People that are finding hard to play with good FPS even on Low may get some advantage, or do you Devs only plan to do that as last resort?

I play FH2 with Dynamic Lighting and Dynamic Shadows activated (Low) and it runs quite fine, I've noted some FPS gain with they on Medium in PR, but still impossible to play.

And why not let us run everything on low, because I can still see peoples heads behind hills even if they are not showing up when you get closer.Running on medium makes the game look even worse.You either run high or low

[Boris]

Any thoughts on why load times have increased so much? Short of a reasonable explanation for this (as in a known cost of some new feature since added to the mod), then this to me would surely indicate an issue somewhere, and one that could quite easily lead to further problems down the line if, for example, it turned out to be being caused by some errant background process eating up CPU cycles. I feel like the results I found from the testing I've done lend support to this theory.

I've tried turning off shadows... - it made no difference whatsoever- the shadows were still there.

Bear in mind that static objects like buildings have shadows pre-rendered for the map and are drawn onto the ground permanently. Vehicle/player shadows should be off though.

[SShadowFox]

And why not let us run everything on low, because I can still see peoples heads behind hills even if they are not showing up when you get closer.Running on medium makes the game look even worse.You either run high or low

Indeed, not to mention the fact that people lived with that for almost 8 years with only a limited amount of people complaining about that, I don't think it would be a problem to let us people with bad PCs to run it on Low again just while they find a better solution for the FPS problems...

[Prevtzer]

Bear in mind that static objects like buildings have shadows pre-rendered for the map and are drawn onto the ground permanently. Vehicle/player shadows should be off though.

Could be that. In 0.98 the visual difference between no shadows and shadows was pretty big.

Anyway, it didn't help with the FPS at all.

[rodrigoma]

Any thoughts on why load times have increased so much? Short of a reasonable explanation for this (as in a known cost of some new feature since added to the mod), then this to me would surely indicate an issue somewhere, and one that could quite easily lead to further problems down the line if, for example, it turned out to be being caused by some errant background process eating up CPU cycles. I feel like the results I found from the testing I've done lend support to this theory.

It seems to be some problems with compatibility. Remember just because it affects you it doesn't mean it affects everyone.


And SShadowFox remember no one normally complained because it is basically an undetectable unbanable cheat.

[Boris]

It seems to be some problems with compatibility. Remember just because it affects you it doesn't mean it affects everyone.

I never said it did affect everyone, but it seems likely it would. Plenty of others mentioned the same thing when v1.0 came out, and I've seen the same on 3 different systems here, different hdds, video cards, ram, cpu, mobos, and OS's, etc, so I doubt it's particular to only my setup, whatever that means. Bear in mind that you're not going to notice this so much either if you run a decent system to start with, so ask those who run lesser systems, and are familiar with how pre v1.0 PR ran, or run actual tests as I did.

[Boris]

Indeed, not to mention the fact that people lived with that for almost 8 years with only a limited amount of people complaining about that, I don't think it would be a problem to let us people with bad PCs to run it on all Low again just while they find a better solution for the FPS problems...

Agreed, if only to help redress some of the imbalance, as folk with slower systems are at a considerable disadvantage in many areas of PR anyway.

[SShadowFox]

And SShadowFox remember no one normally complained because it is basically an undetectable unbanable cheat.

But, who with an average PC would run PR on low just to get some almost not notable advantage? I mean, this is PR, not CoD, a really limited number of people reduced their settings to get some sort of advantage.

Lighting settings on Low don't make the guy inside the building be black while the building itself is white, in fact, the guy, just like the building, would look like as if it was facing the sun directly, which doesn't make a really big advantage unless the other faction uses a dark camouflage.

[Raic]

But, who with an average PC would run PR on low just to get some almost not notable advantage? I mean, this is PR, not CoD, a really limited number of people reduced their settings to get some sort of advantage.

Lighting settings on Low don't make the guy inside the building be black while the building itself is white, in fact, the guy, just like the building, would look like as if it was facing the sun directly, which doesn't make a really big advantage unless the other faction uses a dark camouflage.

But, who with and average PC would run PR on higher settings just to make the game look better? I mean, this is PR, not BF4, a really limited number of people would raise their settings to the limit just to make the game look marginally better.

But honestly, I run most multiplayer games on lowest settings simply because it gives most stable playing and I don't care much for how the game looks. Problem wasn't really that its possible to see people behind ridge or stupid camouflage, but that you can see trough trees and grass or other objects, even at very short distance. Making firefights annoying as you keep getting killed by unseen enemies if you run higher settings.

[SShadowFox]

I'm talking about Lighting settings, I don't care about running geometries on High, it doesn't affect the FPS as much as Lightning on Medium, and really gives people some advantage, but I don't know why would anyone take cover at grass anyway, it should be known to everyone that grass does nothing at range.