Summary
The Forever beta is imposing a 60 FPS cap while its own graphics log reports a GPU temperature of just 32°C.
The log shows the client changing its thermal state from Normal to Serious and immediately imposing the cap. The associated NVIDIA status is:
NV_GPU_PERF_DECREASE_REASON_API_TRIGGERED
This happens despite the foreground FPS limiter and VSync being disabled. The game can initially run around 170 FPS, then become capped at 60 during play. Restarting cleared the cap on one occasion, but it subsequently returned. On the latest launch, the cap was imposed almost immediately.
I am also experiencing occasional brief freezes accompanied by a harsh buzzing sound through my headphones/speakers, including during startup. I do not know whether these freezes and the FPS cap share a cause.
System information
| Component | Details |
|---|---|
| Beta build | 1.60.1.70425 |
| GPU | NVIDIA GeForce RTX 5080, 16 GB |
| NVIDIA driver | 617.42, branch r617_07 |
| CPU | AMD Ryzen 9 9900X |
| RAM | 32 GB installed |
| OS | Windows 11, build 26300.9711 |
| Graphics API | DirectX 12 |
| Display | 2560×1440 at 180 Hz |
The client correctly selects the RTX 5080 and reports:
<IsGPUDriverOutOfDate> No
Actual behavior
- The game can run around 170 FPS before abruptly settling at 60 FPS.
- The cap occurs while actively playing with WoW in the foreground.
- Restarting can temporarily clear it, but it returns.
- The latest launch imposed the cap while the GPU was essentially idle-temperature.
- The client explicitly logs that it is imposing the cap, rather than this merely being an observed performance drop.
Relevant graphics log
From the October 15 session, immediately after startup:
10/15 21:58:45.089 [0] Dx12 Device Create Successful (486.9 ms)
10/15 21:58:45.089 [0] <IsGPUDriverOutOfDate> No
10/15 21:58:46.342 [0] App is background
10/15 21:58:46.342 [0] App is foreground
10/15 21:58:46.423 [0] Notify responders thermal state changed from Normal to Serious
10/15 21:58:46.423 [0] Imposing FPS cap of 60 FPS
10/15 21:58:46.423 [0] Periodic Gpu Status Report: Mem Budget: 2.7GB / 14.9GB (18.0%) Freq:2.61GHz Temp:32C GPU is throttled. Reason: NV_GPU_PERF_DECREASE_REASON_API_TRIGGERED
An earlier session showed the same sequence during play:
Notify responders thermal state changed from Normal to Serious
Imposing FPS cap of 60 FPS
Periodic Gpu Status Report: Mem Budget: 4.7GB / 14.9GB (31.5%) Freq:2.78GHz Temp:67C GPU is throttled. Reason: NV_GPU_PERF_DECREASE_REASON_API_TRIGGERED
Checks performed
Live in-game CVar output showed:
useMaxFPS = 0
maxFPS = 120
vsync = 0
The foreground limiter was therefore disabled; 120 was its stored limit value.
Other observations:
- The client detects the monitor at 180 Hz.
- A separate NVIDIA temperature reading shortly before the latest launch was 33°C, consistent with WoW’s subsequent 32°C reading.
- NVIDIA diagnostic samples taken after the earlier incident showed no active hardware or software thermal slowdown, with both thermal slowdown counters at 0 microseconds. These were subsequent samples, not simultaneous measurements of the triggering instant.
Expected behavior
The client should maintain uncapped rendering when the relevant FPS limits are disabled, unless a valid protective condition requires otherwise.
If a driver status triggers a protective cap, the log should make the reason clear. In this case, the reported reason is API_TRIGGERED, while the client categorizes the condition as a Serious thermal state, even at 32°C.
Reproduction pattern
I do not have a guaranteed manual trigger, but it has occurred repeatedly:
- Launch the Forever beta using DirectX 12.
- Enter the game with foreground FPS limiting and VSync disabled.
- Play normally.
- The game either starts capped or drops from substantially higher FPS to a sustained 60 FPS.
- Check
Logs\gx.logfor the thermal-state transition and explicit cap message above.
Request for investigation
Please investigate how the beta handles NV_GPU_PERF_DECREASE_REASON_API_TRIGGERED, particularly:
- Why that status results in a Serious thermal state and a 60 FPS cap.
- Whether the cap is correctly removed when the triggering condition clears.
- Whether the startup freezes and audio buzzing are related.
The evidence establishes that the client imposes the cap while reporting a low GPU temperature. It does not yet establish whether the underlying problem is NVIDIA’s reported status, the beta’s interpretation of that status, or an interaction between the two.