Yep, but does that make a difference compared to turning it off in Nvidia profile inspector?
I see a few posts mentioning that folks are doing it both ways, so it probably should work.
I can confirm this is true with my 3070ti… this card was behind the door when VRAM was passed out and this makes a difference. Thanks.
Turned in off in TUF BIOS searching on BAR.
Tried this today and saw very positive results after disabling ReBAR via Nvidia Profile Inspector.
Test scenario is IniBuilds Heathrow 2024 (all scenery settings on + HD textures), gate 508, 12pm, clear skies, in the Fenix A320 — chosen purely for simplicity and consistency.
Specs:
- 5700X3D
- 4070 Super (driver 572.83 + DLSS 310.2.1)
- 32GB RAM
- 3440x1440
DLSS frame generation is enabled, using TAA. Most settings are on high, with TLOD at 50 and OLOD at 100 (adjusted in-flight, but not using MSFS dynamic settings ). Raytraced shadows were off initially due to their heavy FPS/VRAM impact.
With ReBAR on: ~80 FPS and ~9GB VRAM when stationary with an external view of Terminal 5. Camera panning was juddery. During taxi, FPS could drop to ~50 with VRAM exceeding 10GB, and sometimes even lower for no clear reason.
With ReBAR off: Still ~80 FPS, but VRAM drops to ~7.5GB in the same static view. Panning is much smoother, as is switching between internal and external cameras. During taxi, FPS held ~60 with a max VRAM usage of 8.5GB. I’ve also been able to re-enable raytraced shadows thanks to the headroom gained.
Currently on my second full flight to test further, but very glad I found this thread.
Yes, I can confirm all this with my 3070ti. I have been using CapFrameX to get a precise memory usage in the card and does not top out until 7.7GB out of 8.0… so the system is using very little … 0.3.
This allows me to run at TLOD 250 until I get close to the terminal at KSFO with the FlightBeam enhancements.
Remember to also turn off Chrome and Mozilla. They are good for 1.1 GB in my hands.
I wonder how much VRAM would be used at KSFO and FlightBeam using a 16GB card. This number would tell us if SU2 is making better VRAM adjustments based upon card capacity.
Anyway, running better for me… Hey I see the new 5060ti is coming out cheap with 16GB… my testing days may be over!
I turned off ReBar, and not only has my VRAM usage decreased, but it’s also helped with stutters in 2024.
My setup is a 9800X3D, 5080, and 64GB of RAM.
Are you having any VRAM issues with the 5080? Looks like the 5060ti will also have 16GB. I was originally planning on the 5080 but hard to find etc…
I would be interested if you get the chance to see what happens to your VRAM at KSFO the standard ILS approaches are to 28L or R. I have a custom terminal as I am a local but the standard terminal is very complex.
I’m using under 12GB at Flightbeam SFO with ultra settings, SU2, TLOD at 250, and MFG at 3x. With ReBar enabled, usage was around 13.5GB.
Looks good, need that new card.
I see your tag you must be a 787 fan. It is a nice plane. I do miss the gull-like wings from MSFS 2020:
since i disabled the r bar i no longer had stuttering and performance drops in airports.
i can run fs24 with all ultra and tlod 400 in 4k with an rtx 4080 super and the simulation is smooth as butter
I can confirm that in my case, too, disabling ReBar noticeably eliminated stuttering and reduced VRAM usage by gigabytes.
5800x3D, 64 GB RAM, 5070 Ti (16 GB VRAM)
i repeat myself . i can see a placebo effect with rbar OFF on loss of FOV in VR . anyone with same experience?
Seems like I have the problem.
Sim shows 14.6gb of vram available in the menu, which seems about right with everything else I have open.
But when i load into an airport, and the vram usage starts to climb, the available vram drop to around 12gb (sometimes even less). Thus causing me to hit vram limit when there should be some to spare. ReBar and HAGS off has no effect on this.
7800x3d, 64 GB RAM, 9070xt
Exactly this. I never actually go above 13GB but whenever it needs the memory, the available memory shows 12GB. When I’m comfortably under, the limit shows as 14.6.
Tried ReBar OFF here but doesn’t work for me. VRAM is a little less (4070 TI Super) but FPS are a lot less than when Rebar is ON. On my tests on a small airport with a GA aircraft I have 148 FPS with ReBar ON and 105 with ReBar OFF. Perhaps ReBar OFF works better for airliners in big airports, but that’s not my kind of flight, so I enabled it again.
So yeah, same scenario:
- Fenix A320
- Dallas Airport - default
- 9070 XT, 16GB VRAM
- 5120 x 1440 with FSR
- High preset, some options off
With Rebar:
Available: 12.6GB VRAM
Usage: 13-14GB
Behavior: smooth with large stutters
Without Rebar:
Available: 13.9-14.6 GB VRAM
Usage: 10.something-12GB VRAM
Behavior: lots of small stutters
Not enjoyable in either scenario.
Use of “Dynamic settings ON” can reduce little stutters. It automatically adjust TLOD when the game needs it, in complex areas to be rendered.
I read somewhere that setting ReBar OFF has a positive effect for those who use TAA anti-aliasing but a negative effect for those who use DLSS.
I couldn’t check it by myself because I have an AMD card…
Yeah, seems to have the biggest impact for those that are more-frequently GPU-limited and using TAA. The GPU-limited part makes sense because it seems to actually slow the speed at which data makes it from the CPU to the GPU to be processed, so the GPU is less likely to go over its VRAM limits.
I saw some comments on a YT video that they had disabled ReBar in Profile Inspector to save about 1GB in VRAM. My impression was ReBar gave the CPU direct access to the frame buffer to speed things up. Anyone out there know about this?