Shadows
-
Seems like update 1 brought back a big loss of FPS with Shadows enable. Prior the update the hit was not too bad. I had a 40 FPS taxiing in the middle of Gunsan with the shadows. Now it’s in the low 20s.
Do you have the same issues?
Thanks
DISCO
On the contrary, I think the visual quality was improved in this update. check any modification in your config
-
Quality is great but performance on the ground is not as good as it was. My Config is the same.
-
@DISCODAN can you get me a save file so I can try repro it here?
-
I see exactly the same behavior - in my test TE 21 Osan Daegu I used to see up to 14000 cmd now in 4.36.1 the cmd count goes up to 22000
-
So I did a complete reinstall and I might have get an extra 10 FPS (from low 20s to low 30s taxiing in Gunsan or Osan).
I forgot to mentioned, on both 4.36 and 4.36.1 I have Janhas models installed. But still 4.36 before the update I had high 40s FPS during taxi.
How do you do a save file btw?
Thanks
-
@DISCODAN U1 is about 1Gb zipped … so, it’s not just code-fixes.
(I don’t know if new buildings / level of detail was added to Kunsan airbase. So, not sure how to do a meaningful apples-to-apples comparison.)
-
So I installed Balkans theatre and even with Janhas mods my FPS is in the 60s on the ground.
Seems to be a KTO issue. More buildings?
-
-
@Seifer I did a weird test, definitely not something supported but just to try to compare apples to apples – I placed the 4.36.0 binary in a folder next to the 4.36.1 binary (ie. within the 4.36.1 stock KTO installation) so I could try to compare frame-timings with exact same models, textures, 3d world, etc.
Scenario:
Flying TR#3, just slowly cruising 2,000 ft alt over Kunsan. All same BMS and graphics settings, ofc… triple-buffering w/o v-sync, to get clear, unimpeded frame timings.
I do have a 30fps frame rate cap, to avoid any possibility of thermal throttling. (and it’s really hot here lately so this is just how I’ve been flying past few weeks
4.36.0 => max number of draw cmds over the airbase was ~7,700, peak cpu “sim” time was +7ms over baseline (~5ms + 7 = ~12ms)
4.36.1 => max number of draw cmds over the airbase was ~13,500, peak cpu “sim” time was +15ms over baseline (~5ms + 15ms = ~20ms)
Is this expected?
-
@Mike_Bravo I also did the trick of removing the 3D trees. It helps a bit but not much.
-
@airtex2019 thanks Airtex, great testing. Can you disable all compatible changes in 3d in 4.36.1 and see if it gets back to normal? you can do that by opening the chat window and typing:
.compatiblechange none
-
@Mike_Bravo I also did the trick of removing the 3D trees. It helps a bit but not much.
How ?
-
@Mike_Bravo I also did the trick of removing the 3D trees. It helps a bit but not much.
How ?
Use 3dbuilder and extract the parent folders. Then remove parents no.
0885
0937
2739
2741
2743
2745
2747Then build the folder. 3dbuilder has very good instructions.
-
@Seifer no significant diff. maybe slightly lower num of draw calls (low 13k vs high 13k range) but I’d chalk that up to random variation.
(for this test I have head-tracking disabled and have my pov fixed forward, in the default position by double-tapping the ‘3’ key)
the chat-command did give me some feedback – showing a 0x0000… hex number, so I believe it was accepted
-
-