Loose 50 fps when looking out of the window
-
I have this strange FPS problem. Normal TE is 100 fps en no problem at all. Sometimes drops to 80.
But during Campaign I’ve got 48 fps on the ground. In the air I still got 48. Sometimes when i look in a certain direction I got constant 90 fps. But when I look out of the window to the terrain it jumps back to 48. That is a lot!!!
Spec
I5 2500k
Nvidia 560ti 1gbIve applied the nvidia settings as in a topic here. Any ideas? A frame drop of 40 fps for just looking out of the window. Is the campaign that demanding that it costs 40fps?
I’ve disabled focus shadows and enabled per-vertex lighting.
-
u r ok.
Campaign uses way too much CPU and less GPU.Falcon is a combo on both.
In campaign the whole war engine must be computed thus the drop.
In TE it calculates only the units that r on the mission thus the 100fps.
-
Oké thx for your reply. But isn’t it strange that in air the fps in one view is 90fps and than when I look sideways or up it slows immediately to 48. That is a enormous amount of fps loss, but I can’t think what triggers it?
-
Oké thx for your reply. But isn’t it strange that in air the fps in one view is 90fps and than when I look sideways or up it slows immediately to 48. That is a enormous amount of fps loss, but I can’t think what triggers it?
The pit.
Demer
-
-
I have this strange FPS problem. Normal TE is 100 fps en no problem at all. Sometimes drops to 80.
But during Campaign I’ve got 48 fps on the ground. In the air I still got 48. Sometimes when i look in a certain direction I got constant 90 fps. But when I look out of the window to the terrain it jumps back to 48. That is a lot!!!
Spec
I5 2500k
Nvidia 560ti 1gbIve applied the nvidia settings as in a topic here. Any ideas? A frame drop of 40 fps for just looking out of the window. Is the campaign that demanding that it costs 40fps?
I’ve disabled focus shadows and enabled per-vertex lighting.
Hi,
No that isn’t strange. at some views Falcon will give you exaggerated FPS, for example you can try to look directly to the empty sky or directly to the empty ground when a weapon does its way towards it (before the explosion which will again suck FPS), you will get pretty high number of FPS (how much exactly depend on your card/OS/driver etc), but if you really want to test FPS quality you should do it in normal views (i.e cockpit panels, and normal outside views when both ground and air are visible).
-
but it sounds like his problem is the opposite, as in his FPS SHOULD drop when looking at models (pit, terrain, etc.) but its higher when looking around in pit, and when he goes to look out to open sky, the FPS drops. if im reading it right
-
I have found BMS is just strange when it comes to FPS. Typically I fly with vsync on and maintain 60 fps for the majority of the time with all of the shiny turned on. For fun I set up an crazy stress test with many groups of tanks engaging each other and i monitored my CPU and GPU load with hardware info. When I looked directly at the conflageration my fps dropped as low as the teens, and when I looked away the fps predictably rose a decent amount to around the 30s. Interesting, during this test my neither my CPU or my GPU hit max load, in fact both were far from it which i found a little strange. Given that this test scenario would never happen in an actual campaign or TE i just brushed it off and continued without worries.
For reference my CPU = i5 3570k OCd at 4.5 ghz
and GPU = EVGA gtx 770 very mild OC through EVGA precision at 102%I have also found that using nvidia inspector to really turn up the graphics level really hurts pit fps. This is with things like enhanced anti aliasing and higher levels of super sampling.(Not worth it in my opinion in game settings seem to work best.)
-
Try dropping a nuke on Seoul and see what happens.
-
I have found BMS is just strange when it comes to FPS. Typically I fly with vsync on and maintain 60 fps for the majority of the time with all of the shiny turned on. For fun I set up an crazy stress test with many groups of tanks engaging each other and i monitored my CPU and GPU load with hardware info. When I looked directly at the conflageration my fps dropped as low as the teens, and when I looked away the fps predictably rose a decent amount to around the 30s. Interesting, during this test my neither my CPU or my GPU hit max load, in fact both were far from it which i found a little strange. Given that this test scenario would never happen in an actual campaign or TE i just brushed it off and continued without worries.
For reference my CPU = i5 3570k OCd at 4.5 ghz
and GPU = EVGA gtx 770 very mild OC through EVGA precision at 102%I have also found that using nvidia inspector to really turn up the graphics level really hurts pit fps. This is with things like enhanced anti aliasing and higher levels of super sampling.(Not worth it in my opinion in game settings seem to work best.)
1. In “empty” TE the sim is GPU bounded, in crowded campaign/TE (lots of units, fighting etc) the sim is CPU bounded.
2. You can bring ANY GPU to its knees in Falcon by using countless number of DXT5 textures in a given scene, good example will be dropping 8 CBUs and following the bombs to ground in orbit view, you’ll get single digits FPS and your GPU will get pretty hot
3. GPU stress (and FPS affect because of it) isn’t only measured by “load” but also by texture memory. For example higher res tiles will have a serious effect on FPS but GPU “load” will be ~normal.
4. CPU load will be pretty low because multi-core usage isn’t really working in Falcon. The sim is heavily multi-threaded but it’s not optimized, so while you better not set affinity to a single core (that will give you lower FPS for sure) you will not see your CPU “stressed” to its full potential.
-
1. In “empty” TE the sim is GPU bounded, in crowded campaign/TE (lots of units, fighting etc) the sim is CPU bounded.
2. You can bring ANY GPU to its knees in Falcon by using countless number of DXT5 textures in a given scene, good example will be dropping 8 CBUs and following the bombs to ground in orbit view, you’ll get single digits FPS and your GPU will get pretty hot
3. GPU stress (and FPS affect because of it) isn’t only measured by “load” but also by texture memory. For example higher res tiles will have a serious effect on FPS but GPU “load” will be ~normal.
4. CPU load will be pretty low because multi-core usage isn’t really working in Falcon. The sim is heavily multi-threaded but it’s not optimized, so while you better not set affinity to a single core (that will give you lower FPS for sure) you will not see your CPU “stressed” to its full potential.
Fix it……LOL!!!
Demer
-
Strange! I will upload Some screenshots soon. First RL things in my house:(
-
The Pit Sucks fps, but hey, ain’t she a beautiful bird???
Huh,wait,why is that left panel not drawn??? its black……damn shaders, fix the normals for your sake…TeeHee!!!Love,
demerBTW we can reduce this to non consequence by reducing the deagg and draw level of the LOD…don’t think we need to begin to draw the pit @ 60000 ft…
-
BTW we can reduce this to non consequence by reducing the deagg and draw level of the LOD……don’t think we need to begin to draw the pit @ 60000 ft…
The pit is rendered only for the player, there is no impact from “other” pits if that what you try to say…
-
The pit is rendered only for the player, there is no impact from “other” pits if that what you try to say…
I do not know where you are coming from with that statement.
NPC code or other???
BUT, I am going to play nice nice with Falcon ATM, so riddle me this Hawkman:Slight gain, agreed, but something is better than nothing……
It still draws the pit…LOL!!!
BTW you have just ruined 6 days of work in the Guam DB… :rofl: I’ll get over it.Tired demer
-
Hi, this statement isn’t mine actually (I don’t have that kind of knowledge for such stuff anyway…) but from someone who knows what he is talking about.
Your comparison proves what exactly?? you show 2 different pictures in 2 different areas and 2 different angles, its well known that looking down into the instruments will cost you some frames, no?
(And just FYI we already checked this stuff anyway, reducing the pit LOD distance changed absolutely nothing, but believe whatever you like ;))
-
Ah that was interesting to read I-hawk. I understand better now I think, before i was simply comparing to other flight sims I have messed around with like fsx and xplane. In both of those sims at least one core is pegged all the time when using complex aircraft like the pmdg 737, even with the 4.5 ghz OC. Looking at that it seemed strange to me that not a single core was maxing out when flying in BMS. So the trick then is to get BMS to use more of the available power by potentially changing the way the processes are coded? That is way over my head lol. Cool stuff anyway though, and not a big deal since it took extreme measures to bring that behavior out.
-
Hi, this statement isn’t mine actually (I don’t have that kind of knowledge for such stuff anyway…) but from someone who knows what he is talking about.
Your comparison proves what exactly?? you show 2 different pictures in 2 different areas and 2 different angles, its well known that looking down into the instruments will cost you some frames, no?
(And just FYI we already checked this stuff anyway, reducing the pit LOD distance changed absolutely nothing, but believe whatever you like ;))
Further comparison in Sim DF module with heavy FPS hitting PR tiles:
I believe what I observe in Sim……its called testing…LOL!!! (now finish up 4.33…!!!)
Could be maybe a hardware difference.Maybe a data issue.You never know in Falcon…LOL!!!
As you know I am edgy about FPS when it comes to GUAM Theater. If we can save a $ here or there, I’m all for it.
Going forward, please post the name of the Author of included statements as I have no idea who you are referencing.
Dunc,Jam, the weasel,the vermin,Biker???FYI please check again and yes I have a HUGE AIM120 'cause you caught me in the middle of scale up……ruined a lot of good women with that one…ROTFLMAO!!!
Cheers,
Demer -
Demer what and how those pics prove?
I did the test my self of reducing the value in LE from 60000 to 100 and nothing happened. Also removed some triangles from the model again no to maybe 1-2fps difference which could be from anything…
-
Dont complain about the pits please……
This isnt the problem.
With only 1gb of GPU ram avail for your entire system, there is no wonder you have poor FPS with the newest of additions to the sim.
Lets leave the poor F16 model alone. Which BTW is 60,000 Polys, Rendered to the VIDEO RAM AT ALL TIMES!!!
So whats left? 600mb for all of Falcon.
My 2 cents.
I will add that with the Sun in view, Per pixel lighting seems to run smoother on some systems. As opposed to per-vertex.