Does FSX use multiple CPU cores, and how do I optimise it?
Learn how FSX uses multiple CPU cores, when AffinityMask helps, whether to disable Hyper-Threading, and which settings improve CPU-limited FPS.
Yes, but only partly. Fully updated boxed FSX and FSX: Steam Edition automatically use additional CPU cores for terrain and texture work, while the main simulation and rendering thread remains the usual FPS limit. Leave multicore scheduling enabled and AffinityMask unset initially; reduce AI traffic, autogen and complex add-on load before changing CPU affinity.
How does FSX use multiple CPU cores?
FSX spreads selected scenery jobs across worker threads but keeps much of each frame dependent on one main thread. This is why a modern processor may show activity across several logical processors without delivering an equivalent increase in frame rate.
| Workload | Typical CPU behaviour | What the simmer sees |
|---|---|---|
| Main simulation and rendering | Heavily dependent on one main thread | One busy logical processor can limit FPS |
| Terrain and texture synthesis | Uses additional worker threads | Extra cores can reduce delayed or blurry terrain loading |
| AI traffic, autogen and draw calls | Much of the cost feeds back into the main thread | Large airports and cities produce sharp FPS drops |
| Gauges and aircraft systems | Depends on how the add-on was built | Complex cockpits may be much slower than default aircraft |
| External utilities and Windows | Can run on other available processors | Spare cores help the whole system without accelerating FSX's serial work |
The unpatched boxed release is the main exception. Bring boxed FSX to the SP2 or Acceleration code level so it has the later engine improvements. FSX: Steam Edition already incorporates that later code, and boxed service packs should not be installed over it.
FSX being a 32-bit application limits its address space, not its ability to schedule work on more than one core. Windows may also move the main thread between logical processors, so the bottleneck can appear to hop between Task Manager graphs rather than leaving one graph permanently at 100%.
Do Concorde and other complex aircraft distribute work across cores?
Workload distribution depends on the aircraft package, not on the fact that it models Concorde or another sophisticated aeroplane. Gauge and systems code running inside FSX can add pressure to its constrained threads, while a separate companion process may run on another core.
Compare the same airport and weather using a default aircraft. If performance recovers, the aircraft's gauges, displays or systems simulation are part of the load; changing global CPU affinity is unlikely to be the best fix.
Does multicore rendering improve FPS, and should it be on or off?
Multicore processing can improve scenery throughput and frame-time consistency in FSX, but it rarely makes FPS scale directly with core count. The main rendering thread still has to complete before the next frame can be presented.
FSX has no normal in-simulator switch called Multicore Rendering. Keep all processor cores available and let FSX and Windows schedule them automatically. If a driver or add-on presents a similarly named option, that setting belongs to that specific component; leave it at its default unless a repeatable FSX test proves otherwise.
Disabling cores can make terrain loading or background utilities slower. It may produce a higher momentary FPS in an unusual configuration, but minimum FPS, frame pacing and texture loading matter more than a peak number.
How do I improve CPU performance in FSX?
The largest CPU gains usually come from reducing the number of objects and simulations that FSX must process. Lowering display resolution mainly helps when the graphics card, rather than the main CPU thread, is the limiting component.
- Update boxed FSX. Install the official updates to reach the SP2 or Acceleration engine level before attempting configuration-file tweaks. Steam Edition does not need those boxed updates.
- Create a repeatable test. Use the same aircraft, gate, airport, weather, time and cockpit view. Allow the flight to settle before judging it, and change only one variable at a time.
- Reduce traffic first. Lower airline and general aviation AI, airport vehicles, road traffic, ships and leisure boats. Dense AI at a large hub is one of the clearest FSX main-thread stress tests.
- Lower Autogen Density and Scenery Complexity. These settings increase object processing and draw calls. Tune them at your most demanding airport rather than over an empty rural area.
- Test aircraft and add-ons separately. Compare a default aircraft, then restore complex aircraft, weather software, traffic packages and detailed scenery one category at a time.
- Choose a sustainable frame-rate target. A limit that the system can maintain often feels steadier than an unlimited rate that repeatedly overloads the main thread. Test limited and unlimited modes because the result varies with scenery and aircraft.
- Prevent CPU throttling. Use a Windows power mode that permits full processor performance, close unnecessary workloads and check that heat is not reducing sustained clock speed.
If resolution and antialiasing changes have little effect, use our FSX CPU-versus-GPU bottleneck checks. Once the limiting component is clear, our complete FSX tuning sequence for traffic, scenery, weather and add-ons covers the remaining settings without relying on speculative processor tweaks.
Should I set FSX AffinityMask in FSX.cfg?
Leave AffinityMask absent unless controlled testing demonstrates a repeatable improvement on your exact processor and add-on combination. It restricts where FSX may run; it does not turn serial code into multicore code.
The entry belongs under [JOBSCHEDULER] in the active FSX.cfg. Its decimal value represents a bitmask in which bits correspond to Windows logical-processor numbers. Those numbers include Hyper-Threading or SMT siblings and, on hybrid processors, potentially different core types.
A mask copied from another computer can therefore exclude useful physical cores, place workers on logical siblings of the same core, or leave too little capacity for terrain work. Values recommended for older four-core processors are not portable to a Ryzen 5 7600X, a modern hybrid Intel processor or a machine with a different firmware configuration.
- Back up the active configuration file. Boxed FSX, Steam Edition and side-by-side installations may use different user configuration files. Editing the wrong copy is a common reason a change appears to do nothing.
- Establish the default result. Remove the
AffinityMaskline, restart FSX and test the same demanding flight several times. - Change affinity alone. Add one calculated mask under the existing
[JOBSCHEDULER]section. Do not create duplicate sections or change process priority at the same time. - Judge more than average FPS. Compare minimum frame rate, visible stutter, texture sharpness and scenery loading. A small peak-FPS gain is not useful if terrain arrives later.
- Revert unless the gain repeats. Remove the entry if results are inconsistent or no better than automatic scheduling.
Should I disable Hyper-Threading or SMT for FSX?
No; leave Hyper-Threading or AMD SMT enabled by default. FSX may gain little from two logical threads sharing one physical core, but Windows, terrain workers and external add-ons can still use the available scheduling capacity.
Disable it only for a controlled before-and-after test that shows better minimum FPS without worse loading or background performance. A global firmware change affects every application, making it a much broader intervention than an FSX-specific affinity test.
What does “main thread priority 2” mean?
Main thread priority 2 is not a standard FSX tuning value and has no universal meaning. If a third-party utility displays that label or number, its definition belongs to that utility; adding an invented MainThreadPriority=2 line to FSX.cfg will not distribute rendering across more cores.
Keep the FSX process at the normal Windows priority unless testing proves a benefit. High priority can deprive weather tools, audio, drivers and scenery utilities of CPU time, while Realtime priority can make the computer unresponsive. Our assessment of CPU-affinity and priority booster applications explains why these tools seldom fix an FSX main-thread bottleneck.
How can I tell whether one CPU core is limiting FSX?
FSX is probably main-thread limited when detailed airports, AI traffic or complex aircraft lower FPS but reducing resolution makes little difference. Several signs together are more reliable than the overall CPU percentage:
- Performance falls sharply near dense autogen, terminal buildings or large AI schedules.
- One logical-processor graph approaches full use, or the heavy load moves between graphs.
- Total CPU usage remains modest because Windows averages activity across every logical processor.
- Lowering traffic or autogen helps more than lowering resolution or texture filtering.
- A default aircraft runs significantly faster than a complex glass cockpit or systems-heavy Concorde package.
Low total CPU use does not mean FSX has unused performance available. On a processor with many logical processors, one saturated main thread can delay every frame while most of the chip remains lightly loaded.
Will a Ryzen 5 7600X or Intel Pentium 2020M run FSX well?
A Ryzen 5 7600X has ample core capacity and strong per-core performance for FSX, while an Intel Pentium 2020M is much more likely to require substantial compromises. Neither processor changes the simulator's basic main-thread design.
| Processor | Expected FSX behaviour | Recommended approach |
|---|---|---|
| AMD Ryzen 5 7600X | Strong main-thread performance with spare capacity for terrain workers and external tools | Leave SMT enabled and AffinityMask absent initially; tune heavy airports and add-ons rather than chasing uniform CPU usage |
| Intel Pentium 2020M | Older mobile dual-core performance can become the limit at modest scenery and traffic levels | Use restrained AI, autogen and scenery complexity, avoid an unlimited frame-rate target and verify that laptop cooling is not throttling the processor |
| Modern high-core-count CPU | Extra cores help Windows and external applications, but FSX still waits on its main thread | Prioritise sustained single-thread performance over core count alone |
Clock speeds from different processor generations cannot be compared directly. Architecture, sustained boost behaviour, memory performance and cooling all affect the work completed by each core; our guidance on balancing CPU and graphics-card upgrades for FSX covers the hardware decision in more detail.
Does the same advice apply to MSFS 2024 or frame synthesis?
No. Microsoft Flight Simulator 2024, including its Sim Update 5 code, uses a different engine and threading model from FSX. It can distribute more work but may still become main-thread limited in demanding situations; FSX.cfg, [JOBSCHEDULER] and FSX affinity-mask recommendations do not apply to it.
Frame synthesis, usually called frame generation, is also separate from multicore CPU rendering. On supported hardware it creates intermediate displayed frames, which can make motion look smoother and raise the reported output rate, but it does not increase the simulator's underlying simulation rate or remove a CPU main-thread bottleneck.
Enable frame generation when the base frame rate is stable and the smoother presentation outweighs added latency or visual artefacts. Leave it off when the base rate is erratic, cockpit displays show distracting artefacts, or immediate control response matters more than the displayed FPS figure.