Performance Guide
If you’re tuning your PC for sim racing (especially with overlays or triple screens), the goal isn’t “more FPS”, it’s stable frame times, predictable input timing, and a system that behaves consistently under load.
If that sounds harder than it should be: you’re not imagining it. Getting a sim to run smoothly can be genuinely frustrating, and the internet is full of performance tips that are either outdated, copied from other genres, or only “work” by hiding the real problem.
So yes, this is yet another guide on the internet. The difference is we’ll tell you why each setting matters, and we’ll avoid the “magic tweaks” that only work on someone else’s PC.
This guide is built from what we’ve learned supporting thousands of RaceLab users across a wide range of hardware. It’s meant to be a reliable baseline: settings that are broadly applicable, easy to validate, and unlikely to create new issues.
We also try to explain the reasoning behind every recommendation. If a tweak only fixes a narrow edge case (or trades one symptom for another), it doesn’t belong in a general performance guide, so we don’t recommend it here.
This guide is written specifically for the RaceLab overlay setup (RaceLab running on top of a sim in windowed/borderless mode via the Windows desktop compositor).
Technologies like G-SYNC/FreeSync/VSync can work great in other setups, but in this specific scenario they often interact poorly with DWM + overlays and can make frame pacing feel less consistent.
If you just want a reliable starting point:
- Run the sim in Borderless (or Windowed) so RaceLab overlays can render.
- Disable VSync and VRR (G-SYNC/FreeSync) for the sim.
- Disable driver-side FPS caps and latency/sync features.
- If you need an FPS cap, use the sim’s in-game limiter only.
- Use a High Performance Windows power plan and disable USB power saving.
- Disable other overlays (Xbox Game Bar, Discord, GeForce/AMD overlays) while testing.
Overview
Sim racing titles behave very differently from typical games.
They run:
- High-frequency physics
- Real-time netcode
- Continuous input sampling
- Telemetry and overlays
- Often triple-screen rendering
Because of this, frame pacing and CPU scheduling matter more than raw FPS.
Many popular “gaming” optimizations actively make sim racing worse.
This guide explains what each major performance setting does and what you should use instead.
Frame Synchronization and FPS Limits
What these settings do
G-SYNC, FreeSync, VSync and FPS limiters change when frames are allowed to be presented.
They reshape frame presentation timing so irregular frame delivery looks smoother on the display.
They do not improve physics timing or CPU scheduling — they only change how the output is presented, often at the cost of extra buffering and latency.
With RaceLab overlays this is especially problematic. The CPU, GPU and physics engine benefit from running without extra timing gates imposed by the driver/OS display pipeline. These sync features can:
- Add input latency
- Hide CPU bottlenecks
- Create stutter when physics load changes
In-game settings
Disable:
- VSync
- Frame rate limiters
- Dynamic resolution
RaceLab recommendation
Sim engines need full, unrestricted timing control. Any frame control layer interferes with physics and input sampling.
NVIDIA Control Panel
For both Global Settings and the sim’s Program Profile:
| Setting | Value |
|---|---|
| Max Frame Rate | Off |
| Vertical Sync | Off |
| G-SYNC | Disabled |
| Low Latency Mode | Off |
| Triple Buffering | Off |
| NVIDIA Reflex | Off |
Why Driver-side synchronization and frame caps override the game's own timing and cause uneven frame delivery under CPU load.
NVIDIA Reflex note Some sims (e.g., ACC) support Reflex natively. If available in-game, it's safe to enable there. Do not enable it in the NVIDIA Control Panel globally — the driver-side toggle doesn't understand sim timing and can interfere with frame pacing, similar to other driver-level sync features.
AMD Radeon Settings
| Setting | Value |
|---|---|
| FreeSync | Disabled |
| Enhanced Sync | Disabled |
| Radeon Chill | Disabled |
| Frame Rate Target Control | Disabled |
Why
These features manipulate the render queue and GPU pacing, which breaks the tight CPU-GPU-physics loop that sims rely on.
RaceLab overlays and windowed mode
RaceLab renders its overlays in a separate transparent window on top of the sim.
Your sim must therefore run in windowed or borderless mode.
There is no inherent performance penalty for this on modern Windows (10/11).
Borderless windowed and exclusive fullscreen typically use the same modern presentation path (flip model / direct scan-out when possible). That means:
- No extra “frame copying” just because it’s borderless
- No built-in FPS advantage from fullscreen
- No built-in input-latency advantage from fullscreen
The main difference today is how synchronization features behave (G-SYNC / FreeSync / VSync / HDR), not raw performance.
RaceLab requires windowed/borderless for overlays, but that does not make the sim slower. It only makes sync technologies more likely to interfere with frame pacing, which is why we recommend disabling them.
Why sync behaves differently with overlays
In windowed mode Windows uses the Desktop Window Manager (DWM) to combine:
Game → RaceLab overlay → Windows → Display
When G-SYNC, FreeSync or VSync are active, the display synchronizes to the composited output, not the game itself. Any jitter in the overlay now affects the game.
Because RaceLab updates telemetry, timing and gap data continuously, the display must wait for both the game and the overlay on every frame. Any small delay in the overlay stalls the entire composited frame and indirectly couples physics timing to overlay rendering.
This causes microstutter and inconsistent driving feel even when GPU usage looks low.
If you need to limit FPS
If you want to cap FPS (to reduce GPU load, heat or fan noise), only use the sim's built-in frame limiter.
The sim's internal limiter understands when physics ticks run and how the render queue is managed. Driver and external limiters (NVIDIA Max Frame Rate, AMD FRTC, RivaTuner) do not — they inject delays into the GPU pipeline without understanding simulation timing, which creates jitter and input lag.
If you can’t stand tearing (last resort)
If tearing is unacceptable, treat this as a compromise and test carefully (because you’re reintroducing timing gates into a composited, overlay-heavy presentation path).
Recommended order to try:
- Try in-game VSync (driver VSync remains Off), then verify frame pacing.
- If you use a cap, keep using the in-game limiter (avoid driver/external caps).
- If you enable VRR anyway, understand that in windowed + overlays the display will react to the composited output, and any overlay jitter can become visible as microstutter.
If any of this worsens frame-time stability, revert to the baseline (sync Off).
Measure and Validate Changes
Performance tuning is only useful if you can confirm it helped.
What "good" looks like
- Stable frame times (flat frame-time graph, no periodic spikes)
- Consistent driving feel (no random "heavy" steering/input moments)
- No rhythmic hitching (every few seconds) and no stutter during traffic/starts
How to test with CapFrameX
CapFrameX (built on PresentMon) is the best free tool for measuring frame-time stability. Use it with a replay to get perfectly repeatable results:
- Save a replay in your sim — pick a track and car you use often.
- Play the replay and choose a specific lap as your test segment.
- Start a CapFrameX capture at the same point in that lap every time (e.g. start/finish line) and record for a fixed duration (20–30 seconds is enough).
- Change one setting, restart the replay, and capture the same lap and duration again.
- Compare the frame-time graphs and 1% / 0.1% lows between captures.
Using a replay eliminates variables like AI behavior, traffic and network conditions. As long as you capture the same lap for the same duration, the workload is identical and the comparison is fair.
MSI Afterburner (or similar) is useful alongside CapFrameX to monitor CPU/GPU usage and clock speeds during the capture.
Bottleneck and Stutter Diagnosis
Different symptoms point to different fixes.
If GPU usage is ~95–99% most of the time
You’re GPU-limited.
Try:
- Lower resolution or render scale
- Reduce expensive settings (shadows, reflections, mirrors, AA)
If GPU usage is low/moderate but stutters happen in traffic/starts
You’re likely CPU/physics-thread limited or you have a scheduling/latency problem.
Try:
- Ensure all cores/SMT are enabled (avoid BIOS “gaming modes”)
- Use a High Performance power plan (avoid downclocking)
- Disable sync/driver caps (they can mask bottlenecks and add jitter)
If stutters are periodic (e.g. every few seconds)
This is commonly background work interrupting the sim.
Check:
- Windows updates/downloads, cloud sync (OneDrive), antivirus scans
- Other overlays and capture features
- USB power saving (wheel/pedals disconnect/reconnect can look like “stutter”)
AMD BIOS Gaming Modes
Modern AMD boards include features that try to “optimize” games by changing core and CCD usage.
These are harmful for sim racing.
Turbo Game Mode
Disables one CCD and often SMT, forcing the game onto fewer cores.
Recommendation: Disabled
Gaming Adaptive CCD Parker
Parks cores and moves threads between CCDs dynamically, causing cache misses and jitter.
Recommendation: Disabled
Recommended AMD BIOS Configuration
| Setting | Value |
|---|---|
| Turbo Game Mode | Disabled |
| Gaming Adaptive CCD Parker | Disabled |
| SMT | Enabled |
| All CCDs active | Yes |
| CPPC / Preferred Cores | Enabled |
| XMP / EXPO | Enabled |
Why XMP / EXPO matters Sim racing is latency-sensitive — faster RAM reduces memory access latency for physics and input processing. On AMD systems, memory speed is directly linked to the Infinity Fabric clock, which affects inter-core communication and cache performance. XMP / EXPO simply enables the speed you already paid for, without it, your RAM runs at slower default JEDEC speeds. You don't need to overclock beyond your kit's rated speed; just make sure the profile is enabled.
Intel BIOS Gaming and Hybrid-Core Settings
Intel CPUs use fast P-cores and efficient E-cores.
Sim racing benefits from using both.
Disable motherboard “gaming” modes and let Intel manage boosting and scheduling.
Recommended Intel BIOS Configuration
| Setting | Value |
|---|---|
| Intel Turbo Boost | Enabled |
| SpeedStep / Speed Shift | Enabled |
| Hyper-Threading | Enabled |
| P-cores | Enabled |
| E-cores | Enabled |
| Thread Director | Enabled |
| AI / Gaming Boost modes | Disabled |
| Multi-Core Enhancement (MCE) | Disabled |
| XMP | Enabled |
Why XMP matters Without XMP enabled, your RAM runs at slower default JEDEC speeds regardless of what you paid for. Faster memory reduces access latency for physics and input processing — just enable the profile in BIOS.
AMD X3D vs Intel Hybrid Cores in Sim Racing
AMD X3D CPUs use:
- One CCD with massive cache
- One CCD with higher clocks
Intel CPUs use:
- P-cores for latency
- E-cores for background work
Windows already knows how to schedule both correctly.
What you should do
- Keep vendor scheduling features enabled (AMD CPPC/Preferred Cores, Intel Thread Director)
- Avoid motherboard “game boost” presets that disable cores or move threads aggressively
- Keep BIOS and chipset drivers up to date (especially on multi-CCD AMD systems and Intel hybrid-core systems)
The goal is simple: stable cache locality and predictable scheduling, not fewer cores. BIOS "gaming" modes break this by disabling cores, moving threads aggressively and destroying cache locality.
Let the OS and CPU vendor handle scheduling.
Windows Power and CPU Scheduling
With RaceLab overlays, short stalls are much more noticeable. You want stable clocks and fewer power-saving transitions.
Power plan
- Use High performance (or AMD Ryzen High Performance if available)
- Avoid aggressive power saving plans while driving
If you see clock dropouts or periodic hitching, also check:
- PCI Express → Link State Power Management: Off
- Processor power management settings (let High Performance do the work; avoid custom “tweak” plans unless you can validate them)
USB and device power saving (important for wheels/pedals)
- Disable USB selective suspend in Power Options
- In Device Manager (USB hubs), disable “Allow the computer to turn off this device to save power” where applicable
Windows gaming and graphics settings
These can be workload-dependent; if you change them, re-test and keep what improves frame-time stability.
- Game Mode: usually best left On
- Hardware-accelerated GPU scheduling (HAGS): try On and Off and keep the better one for your system
- Fullscreen optimizations: generally fine, but if you see odd frame pacing in borderless, test disabling it for the sim
- Windows Variable refresh rate setting: keep Off when using RaceLab overlays
Windows 10 vs Windows 11
Windows 11 enables VBS (Virtualization-Based Security) by default on clean installs, which can reduce gaming performance by 5–10%. If you notice lower performance after a clean Windows 11 install, check Settings → Privacy & Security → Device Security → Core Isolation → Memory Integrity — consider disabling it on a dedicated racing PC. Windows 11's DWM compositor behavior is also slightly different from Windows 10; if you see unexplained frame pacing changes after upgrading, the baseline recommendations in this guide still apply.
Ensure both apps use the discrete GPU
On systems with an iGPU + dGPU, make sure both the sim and RaceLab are set to use the High performance GPU in Windows Graphics settings. If you're on a laptop, see Laptop and iGPU+dGPU Systems for additional steps.
Mixed Refresh Rate Monitors
If you run triple screens (or any multi-monitor setup) with different refresh rates, DWM composites the desktop at the lowest common refresh rate. This means your higher-refresh display(s) will effectively run at the slowest monitor's rate, causing stutter and uneven frame pacing on those screens.
Recommendation
- Match all monitors to the same refresh rate if possible — this is the simplest fix.
- If your monitors have different maximum refresh rates, set them all to the same rate in Windows Display Settings (even if one monitor supports a higher rate).
- This is separate from VRR (G-SYNC/FreeSync) — this is about static refresh rate mismatch between monitors.
This is a common cause of "stutter on triples" that no amount of in-game or driver tuning will fix.
GPU Drivers
Keep the driver simple
- Prefer stable/released drivers over beta
- Avoid enabling extra driver “enhancements” for the sim profile (sync, caps, overlays)
Do a clean reset if you’re chasing unexplained stutter
If your system used to be smooth and suddenly isn’t:
- NVIDIA: use the installer’s clean install option when updating
- AMD: use factory reset (or the equivalent clean-install option)
Then re-apply the minimal settings from the sync tables above.
Laptop and iGPU+dGPU Systems
If you're racing on a laptop (or a desktop with both integrated and discrete graphics), there are a few extra things to check beyond the general settings above.
Force the discrete GPU for both apps
In Windows Settings → System → Display → Graphics, add both your sim and RaceLab and set them to High performance (your discrete GPU). If either app falls back to the integrated GPU, you'll see major performance drops and potential stutter.
Disable dynamic GPU switching
If your laptop supports it, use a MUX switch or set dGPU-only mode in BIOS to bypass Optimus / AMD Switchable Graphics entirely. Dynamic GPU switching adds latency and can cause frame pacing issues when the system decides to route through the iGPU.
Keep the power adapter plugged in
Battery mode throttles both CPU and GPU clocks to save power. Always race on AC power.
Watch for thermal throttling
Laptops have limited cooling. Keep the laptop well-ventilated, consider a cooling pad, and monitor temperatures — if the CPU or GPU hits its thermal limit, clocks drop and frame times spike.
Disable manufacturer power-management software
Software like Asus Armoury Crate, Lenovo Vantage, or MSI Dragon Center often includes "Silent" or "Battery Saver" modes that override your Windows power plan and throttle performance. Make sure your system is set to its highest-performance profile in these tools, or disable them entirely while racing.
Background Overlays and Capture Features
With RaceLab overlays, stacking multiple overlays/capture layers increases the chance that the composited frame gets delayed.
While troubleshooting, disable:
- Xbox Game Bar / background recording
- Discord overlay
- Steam overlay
- GeForce Experience overlay / AMD Adrenalin overlay
After you have a stable baseline, re-enable one-by-one if you really need them.
Network Stability
Not all "stutter" is a rendering problem. Network issues can cause other cars to teleport or rubber-band, which also affects RaceLab position and gap data.
- Use wired Ethernet whenever possible — WiFi packet loss and latency spikes are a common source of jitter that looks like stutter but can't be fixed with graphics settings.
- Disable bandwidth-heavy background tasks while racing (cloud sync, streaming, large downloads, OS updates).
- If you must use WiFi, connect on the 5 GHz band and stay close to the router to minimize interference and packet loss.
This won't fix rendering or frame pacing issues, but it eliminates a common source of confusion when diagnosing "stutter" that's actually netcode jitter.
RaceLab Overlay Tips
When troubleshooting, start with the minimum number of RaceLab overlays you need, then add more once frame pacing is stable.
If RaceLab provides options for overlay refresh/update rate or visual effects, prefer the most conservative settings while testing.
Avoid “Magic Tweaks”
These frequently get recommended online, but are not good baseline advice for sim racing stability:
- Disabling random services without understanding impact
- Timer/HPET tweaks and registry hacks
- Core parking tools and forced affinity/priority scripts
- Third-party “optimizer” apps and driver-tweak packs
If a change can’t be validated with frame-time testing, it doesn’t belong in your baseline.
Discord: Asking for Help in #i-need-help
If you’re stuck and want help from the community (and our team when available), please use our Discord #i-need-help channel.
To make it possible for others to help you, please include the info below in your first message.
Please include a CapFrameX A/B capture
The most helpful thing you can provide is a CapFrameX capture that compares the same replay segment with and without RaceLab:
- Save a replay and pick a specific lap to use as your test segment.
- Play the replay and capture 20–30 seconds of that lap without RaceLab running.
- Restart the same replay and capture the same lap and duration with RaceLab running.
- Share the result/screenshot(s) in #i-need-help (label which is "with" vs "without").
Using a replay keeps the workload identical between captures, so any difference in the frame-time graph is caused by RaceLab — not random traffic or network variation.
System info to include
Copy/paste this and fill it out:
- Sim:
- CPU:
- GPU:
- GPU driver version:
- RAM:
- Monitor setup (single/triples, refresh rate(s), mixed refresh yes/no):
- Display mode (Borderless/Windowed):
- VSync/VRR status (in-game / driver / Windows):
- RaceLab version:
- Which RaceLab overlays are enabled:
- Other overlays/capture apps running (Discord/Steam/Game Bar/GeForce/AMD overlay/etc.):
If you include both the capture and the system info, it’s much easier for others to suggest the right next step.
Summary
For the best sim racing experience:
- Disable sync and driver/external FPS limiters
- Run windowed or borderless with RaceLab
- Keep all CPU cores and SMT enabled
- Avoid BIOS “gaming” modes
- Use high-performance Windows power
And validate changes with a repeatable test and frame-time metrics.
This produces the smoothest physics, the most consistent frame pacing and the best driving feel.