Odd performance behaviour around wheels

I know you can't have an entire vehicle being processed by more than one core, but could it be possible to have just the wheels of the T75 processed by another core?

Or is the low FPS with wheels just a bug or optimization issue?

 

We need your hardware specs.

 

Really late reply I know, but setting it to 2x realtime my performance sky rockets to 61fps (same as what it is with no wheels) over the usual 20fps I get in realtime. I am really confused as to why this is happening as removing the self collision has made no difference to the performance for me at all in realtime.

Realtime:


2x:


Also did a better bench mark:

physics on, all wheels. 19.4 fps


Physics off, all wheels. 70.4 fps


Physics on, no wheels. 63.8 fps


Physics off no wheels. 72.5fps


I also noticed when I remove one set of rear wheels I gain around 3-4 fps pushing it up to around 24 fps with physics enabled. When removing the second set of rear wheels it goes all the way up to exactly the same as what it was with no wheels, meaning that the front wheels have no noticeable performance impact whatsoever on my system.

Specs (also in sig but meh)

AMD FX-8350 @ 4GHz with Coolermaster Hyper 212 EVO| Sapphire HD7870 2GB | GIGABYTE 970A-DS3P| 8GB DDR3 @1333MHz | HDD 1TB SATA II 3.0GB/s | Corsair CX750M modular PSU|

This is really the difference between unplayable and playable, I will do some tests with standard cars to see if there is any similar, albeit less noticeable issue.

Edit: Well I've discovered something quite unusual with the frame-rate on all of the official vehicles that I have tested.

For example you'd naturally expect the D-15 or the Grand Marshal to have overall lower performance than the Covet or the unfinished Sunburst... However when I spawn them I end up (with physics enabled and set to realtime) the frame rate hits 64 fps and stops it wont go above it unless the physics are paused. This time though similar to the T-45, slowing the simulation speed although very small there is a 1-4 fps increase in any of the default vehicles pushing the fps slightly above 64fps but this only happens when not in realtime. Removing wheels on these vehicles has no noticeable impact.
I have checked several times and v-sync is indeed disabled completely and enabling it drops the fps to exactly 59-60 as it should so it is defiantly disabled during these tests. I can also rule out any hardware issues as this is a newly built PC with all of the parts no older than around 2 months and testing and bench marking have shown no problems.

 

OK here I go... I've been troubleshooting my PC's stuttering for over a week... Today I turned my BIOS power's setting from 'normal' to 'performance' and holy crap what an interesting change of things that I have noticed happen... Not all necessarily for the better though there are improvements. It's barely been an hour so we'll see what else I discover throughout the rest of this day and beyond.

I'm still getting the stuttering in BNGDrive though it is now less pronounced. In Minecraft the stuttering is now apparently non-existent but we'll see about that once I play for more than 3 minutes. Furthermore while playing around with BNGD's graphic settings, when I switched everything from 'Lowest' to 'Highest', almost everything on-screen turned black; as if the GPU was about to die! I troubleshot my way through that and when I set the lighting option to 'Lowest' everything worked once again. I increased the lighting setting to 'Low' and the game crashed. I restarted the game, graphic options were all pre-set to 'Highest', loaded a map and all was fine...

As for the T75's lag, things have taken a turn for the better but I think the improvement occurred yesterday when I optimized my RAM in my BIOS settings. I had never properly set the BIOS when I upgraded my RAM about half a year ago; I know, I'm an idiot newb at times. So anyways, since my BIOS RAM setting were optimized to the RAM's manufacturer's test values (9-10-9-27) the lag with the T75 has almost entirely stopped. However this is untrue on the East Coast map. As soon as I drive on the large bridge straight ahead of the default spawn my framerate plummets from above the 80's (this is with full graphic settings on 'High') to the mid 20's. Once I've crossed the bridge the framerate slowly climbs back up.

Here are my gridmap results:

Physics: ON
Wheels: ~98FPS
No Wheels: ~104

Physics: OFF
Wheels: ~110
No Wheels: ~117

In comparison with the Ibishu Covet:

Physics: ON
Wheels: ~99
No Wheels: ~112

Physics: OFF
Wheels: ~99
No Wheels: ~113



However I now get a strange bug when I spawn the T75 without the rear 'rear axle'.


Note that the stretched material casts a shadow:



My PC specs:
Asus Sabertooth P67 with most up-to-date BIOS and most recent Intel Chipset driver
Intel i7 2600K (@stock 3.4GHz) w/ Coolermaster FreezerPro HSF
16GB Kit Corsair Vengeance Pro (2x8GB) 1866MHz 9-10-9-27 as per factory test
Asus ROG Xonar Phoebus (sound card)
Asus GTX 770 DirectCU II OC 2GB GPU
Corsair HX650 PSU (potential culprit to my performance issues)

I am under the impression that if a computer system has not been optimized properly then it will not run BeamNGDrive very efficiently. I wish I knew exactly what I did during my stutter troubleshooting but I no longer have the severe drop in framerate that I originally experienced with the T75 unless of course I drive over that bridge on East Coast USA. I haven't gone as far as to note if the other bridges also cause this down spike in performance.

Hopefully all this feedback will help.

 

Smeowkey, you are only supposed to be removing the wheels, not the entire axle (otherwise the driveshaft between the 2 axles loves to spaz out like that, or I think that is what it is)

 

Hmm, I don't have any problems with the T75.

Physics ON and all wheels: 78 fps
Physics ON and no wheels: 82 fps.

Everything on max details and full hd

CPU: i5-2500k overclocked to 4,5 GHz
Graphics card: GTX 670

when vsync enabled of course both cases 60 fps.

 

I made sure to remove only the wheels. My post is accurate, no worries, I guarantee it!

Earlier, those of us who had had the performance issues, would remove the rear 'rear axle' to get an acceptable and playable level of performance.

In the past I had successfully spawned the T75 without the rear 'rear axle' without having the materials stretching out vertically into infinity. This is a problem that developed yesterday.

 

Sorry, excessively late reply. I thought the thread died. Thats damn interesting, and I have to say I get a similar experience upon further experimentation. so I'm not crazy, this does exist.

 

i was actually browsing the facepunch beamng thread, and noticed this post by gabester:



so the devs are looking into it

besides, it's not called experimental for no reason

(have a link to the post as well if you want to see it for yourself: http://facepunch.com/showthread.php?t=1295724&p=45983825&viewfull=1#post45983825)