Fighting swarms of demons is awesome until you hit those random micro-stutters. The Fanxiang S790 4TB has an overly aggressive I/O scheduling policy for fragmented assets, with response times jumping between 1ms and 25ms, causing frame times to spike from 12ms to 35ms. I tried killing all background apps in Windows, but while CPU usage dropped, the I/O jitter stayed—just a waste of time. I went into the BIOS, disabled PCIe Link State Power Management, and forced the drive into High Performance mode. In RTSS, the frame time variance dropped from 10-30ms down to a tight 13-16ms. The game feels way smoother now. I did notice the idle temp went up by 4℃ after disabling power management, but I fixed that by tweaking my case airflow. Drive temps are now steady at 52-58℃ and the motherboard is at 60-66℃. Monitoring confirms the scheduling mode is active and holding steady. Last updated onMarch 19, 2026 6:09 PM.
The asset loading in this game is a joke; scenes just go black and then pop in out of nowhere. My drive was tanking to 50MB/s at the worst moments. Because the Intel 660P uses QLC NAND, once the SLC cache runs dry, the performance falls off a cliff, pushing my frame times to 40ms. I tried dropping the resolution to 1080p, but it looked like a blurry mess, which was just masochism. I eventually manually locked my virtual memory to 32GB in system settings and dialed the texture quality down from 'Ultra' to 'High'. In Resource Monitor, the disk active time finally settled between 15-25%, and the popping disappeared. My boot time took a 3-second hit after the VM change, but disabling Fast Startup fixed that. Drive temps are between 40-48℃ with fans at 1200 RPM. I exported the read curves via analysis tools, and the fan speed is now stable at 1400-1600 RPM, though QLC drives will always have these limits. Last updated onMarch 16, 2026 9:26 PM.
The game would just vanish and dump me back to the desktop after about an hour, which is an absolute nightmare when you're mid-mission. Monitoring showed the FireCuda 540 controller hitting a blistering 82-88℃ during 2GB large-file writes, triggering the hardware thermal shutdown. I tried slapping two more fans on the top of my case, but that only dropped the temp by 3℃—the core was still hovering above 80℃, so the crashes kept happening. I finally went into Device Manager and changed the write caching policy from 'Maximum' to 'Balanced' and set the PCIe power management to 'Maximum Performance'. In OCCT stress tests, the controller temp dropped to a safe 65-72℃, and the crashes stopped entirely. I noticed loading times increased by about a second after the cache tweak, but enabling Re-Size BAR brought that performance back. Temps are now steady at 58-64℃. After 5 hours of gaming, no crashes, and the controls feel incredibly responsive. Last updated onMarch 13, 2026 2:08 PM.
Driving through Night City felt like a slideshow with massive screen tearing, and my frame times were jumping all over the place from 18-55ms. It totally killed the immersion. I dug into the logs and found the Kioxia Exceria Pro was glitching between PCIe 3.0 and 4.0 under heavy load, causing throughput to fluctuate between 3.5GB/s and 7.2GB/s. I tried enabling low-latency mode in the drivers, but while the mouse felt snappier, the frame drops stayed—it was just a band-aid on a bullet wound. I eventually went into the BIOS, forced the M.2 slot to Gen4 instead of Auto, and flashed the latest firmware. RTSS confirmed the frame intervals tightened to 14-18ms, and the smoothness is night and day. Interestingly, my boot time slowed by 2 seconds after forcing Gen4, which I only fixed by disabling Windows Fast Startup. Drive temps are stable at 55-62℃ with fans at 1500 RPM. After three hours of testing, the I/O bottleneck is dead, and memory temps are holding at 58-63℃. Last updated onMarch 12, 2026 4:57 PM.
Whenever I hit the dense buildings in Saint Denis, the drive struggles with the massive stream of high-res textures. I saw random read speeds swinging wildly between 45MB/s and 120MB/s, which tanked my frame rate from 75 FPS down to a choppy 30 FPS. I initially tried pinning the page file to 16GB, but that software tweak did absolutely nothing for the hardware I/O bottleneck—it actually messed up some texture loads, which was incredibly frustrating. I finally shifted gears and installed the latest official Western Digital NVMe controller drivers and set the HDD turn-off time to 0 minutes in Windows Power Options. Checking AIDA64, the random read latency tightened up from a messy 15-45ms range to a rock-steady 8-12ms. I did hit a snag where the drive wasn't recognized immediately after the update, but a quick M.2 reseat and cleaning the gold pins fixed it. Temps sat between 42-51℃ with the heatsink feeling warm. After a three-hour stress test, the read curve is back to baseline and frame times are locked at 5.1-6.4ms. Last updated onFebruary 17, 2026 12:20 PM.
