Running this on Win11 24H2, AIDA64 logged write bandwidth peaks at 3.6GB/s - 4.2GB/s with controller temps at 58℃ - 63℃. I initially had the polling interval set to 1 second, which actually ate too much CPU and caused the frame times to look like a jagged saw. I went into AIDA64 $
ightarrow$ Sensor Settings and bumped the polling rate to 2 seconds. This dropped resource overhead by 11% and the tearing felt way less aggressive. After three reboots, the FPS variance shrank to ±3fps. However, in crowded areas, the game's weird way of calling the fast external lanes still causes brief hitches. That's a software flaw the hardware just can't fix. Last updated onMarch 3, 2026 1:18 PM.
Looking at Disk Report 2026-015, I used FPS Monitor and noticed the frame time graph looked like a sawtooth. I dug into the AIDA64 sensor panel and found the controller temp staying between 57℃ - 62℃, with write bandwidth peaking at 3.5GB/s - 4.1GB/s. I tried setting the sampling interval to 1s, but the software itself started eating too much CPU, making the game even laggier. Once I bumped it to 2s, CPU overhead dropped by 12%. RTSS then confirmed the frame curve had finally flattened out, and the tearing was way less noticeable. However, because of the game's streaming engine, you'll still get those brief hitches when flying at max speed. It's a known flaw in the current build. Last updated onMarch 4, 2026 1:47 PM.
Looking at disk report 2026-015, I used FPS Monitor and saw the frame time graph looking like a jagged mountain range. I dove into the AIDA64 sensor panel and found the controller temp staying between 56℃ - 61℃, with write bandwidth peaking at 3.4GB/s - 4.0GB/s. I tried setting the sampling interval to 1s, but the software itself started eating too much CPU, making the game even laggier. Once I bumped it to 2s, CPU usage dropped by about 10%. I verified with RTSS that the frame curve finally flattened out and the tearing stopped. However, because of how the game handles streaming, you still get these brief hitches during high-speed flight. It's just a broken part of the current build. Last updated onMarch 5, 2026 2:26 PM.
This was a classic case of monitoring software conflict. In test DS2-2026-T1, I realized the default 1s polling rate in HWMonitor was causing tiny instruction blocks while the CPU hammered the sensors. Analyzing the waveforms, I saw that when temps hit 70℃ - 75℃, the aggressive polling created frame time spikes of 15ms - 20ms. I bumped the sampling interval to 2s and killed unnecessary voltage monitoring. Cross-referencing with AIDA64, the data fit the actual load curve at 98% accuracy, and resource overhead dropped by 10%. The panel is rock steady now. Just keep in mind this only fixes the reporting accuracy—it won't magically fix physical heat pipe latency in extreme heat. Last updated onMarch 8, 2026 10:19 PM.
The issue is a mismatch between the software polling interval and hardware response. I went into HWiNFO -> Settings and forced the sensor scan interval from 2000ms down to 500ms. In the [Env-S2-2026] setup, RAM temperature refresh latency shrank from 30ms - 60ms to a crisp 12ms - 18ms. It turns out the software was just merging samples to save CPU cycles. Now, temperatures hover realistically between 45C - 56C without those terrifying fake peaks over 80C. The trade-off? CPU background usage climbed by about 1% - 2%, which might cause tiny frame drops on bottom-tier CPUs—a typical 'accuracy over performance' compromise that I'm fine with. Last updated onDecember 3, 2025 1:42 PM.
