Thank you, dafarquaad, for updating us on the situation.
I can only imagine how discouraging it must be that clearing the shader cache was only a temporary fix and the crashes have returned. It’s such a shame—it’s Season 3, after all, and you can’t play with peace of mind when the entire system is crashing.
After reviewing the second DxDiag and the latest details, I’ve noticed a few key points. If you’re still interested in troubleshooting the issue yourself, please refer to the information below.
1. New error code (0xD1) detected
In the +++ WER0 +++ section of your second DxDiag, a new BlueScreen P1: d1 has been recorded. This is a DRIVER_IRQL_NOT_LESS_OR_EQUAL error, indicating that a specific driver accessed an invalid memory area.
When considered alongside the first 0x9F (DRIVER_POWER_STATE_FAILURE) and LiveKernelEvent 117/193, it is highly likely that a serious conflict or communication failure is occurring within the Windows kernel or device driver layer.
2. Is the true cause a Windows Update bug, not DICE (the game)?
In fact, since the May 12 Windows 11 security update (KB5089549), numerous serious issues have been reported in the market, including a known bug causing incomplete installations on devices with small EFI System Partitions (ESP), as well as reduced network speeds and instability in NVIDIA GPUs. Following the April update, Microsoft has now experienced incidents that shake the system’s foundation for two consecutive months.
The Battlefield 6 relys heavily on inter-process communication (IPC) with the Windows kernel, are inevitably hit hard by these kinds of OS bugs. It is highly likely that the root cause of the issue was not a DICE patch, but rather a bug in the Windows Security Update released.
3. Suggestions for potential self-troubleshooting
If you’re thinking, “I want to try a few more things while waiting for the fix,” the following steps may help you narrow down the cause.
・ Check the CBS log (to narrow down errors)
Path: C:\Windows\Logs\CBS\CBS.log
Open this log file and use Ctrl + F to search for the strings “SpaceCheck: Insufficient free space” or “fail.” If you find any matches, this indicates that the OS update did not complete successfully and that system files are corrupted (potentially affecting graphics or power management drivers).
・ Roll Back Windows Update
If possible, it’s worth trying to uninstall the Windows 11 security update (KB5089549) applied on May 12, reverting to the state prior to April, to see if BF6 runs more stably.
I previously encountered a similar OS-related update issue on Windows 10 and successfully identified and manually worked around the cause. I have a feeling that we can find a workaround for your environment (Windows 11) as well, before Microsoft releases a revised version to fix the bug.
However, since forced PC reboots place a strain on both your mental state and the hardware, I think it’s a good idea to “step away from the game until an official fix is released.”
[Breakdown Suspects Based on Error Codes]
The bug check codes (P1 values) appearing in the old and new logs have the following meanings, respectively.
Suspect 1: NVIDIA Display Driver (TDR Delay)
Relevant Error: LiveKernelEvent 117 (VIDEO_TDR_TIMEOUT_DETECTED)
Meaning: This indicates that the GPU stopped responding for a certain period (typically 2 seconds), causing Windows to determine that the system had “frozen” and attempt to restart the driver.
Breakdown: This matches perfectly with what you said: “Clearing the shader cache temporarily fixed the issue.” The GPU becomes unresponsive (hangs) either during cache loading or at the rendering instruction stage.
Suspect 2: Windows Power Management (ACPI / PCIe Bus)
Relevant Error: BlueScreen 9F (DRIVER_POWER_STATE_FAILURE)
Meaning: As “P2: 3 (Device object blocked IRP for an extended period),” this is a BSoD caused by a specific driver refusing to respond when the system transitions to power-saving mode (or returns to normal operation).
Breakdown: When the GPU or chipset experiences sudden load fluctuations (changes in power state) during gameplay, the drivers are failing to hand off control properly.
Suspect 3: Hardware Device (or PLX chip, NVMe, etc.)
Relevant Error: LiveKernelEvent 193 (WHEA hazard related / STORPORT malfunction, etc.)
Relevant Error: LiveKernelEvent 1B8
Meaning: This is a LiveKernel event (a log generated when Windows detects a hardware anomaly—particularly with the PCIe bus, storage, or power rails—without triggering a BSoD) that is logged in the background.
Suspect 4: Most Likely Memory conflict at the kernel level (failure of interrupt handling)
Relevant Error: BlueScreen D1 (DRIVER_IRQL_NOT_LESS_OR_EQUAL) - New occurrence on the second instance.
Relevant Error: BlueScreen 50 (PAGE_FAULT_IN_NONPAGED_AREA)
Meaning: Both errors indicate that a driver attempted to access an “invalid memory address” that it should not have accessed or violated the kernel's highest-priority processing (IRQL) rules, causing the system to shut down forcibly (BSoD) as a protective measure.
Breakdown: In addition to the first occurrence (50), the decisive factor is that (D1) surfaced as the top-level error (WER0) during the second occurrence. This is evidence that the game and the driver are conflicting in memory as a result of changes in kernel (memory management) behavior caused by an OS update.
● Why the “second DxDiag” showed differences
The most notable features of the second log are that “a new 0xD1 error appeared at the top (most recent)" and that “hardware-related errors such as 1B8 were present, with kernel/driver errors occurring in rapid succession.”
It appears that when you tweaked the shader cache, the timing of the GPU’s data processing (specifically, the order of memory accesses) shifted slightly. As a result, what previously resulted in a “No Response” error (codes 9F or 117) has now escalated to a “fatal collision that directly corrupts the kernel’s memory space” (code D1) due to this timing discrepancy.
● Final “Narrowing Down the Suspects” and Conclusion
Ryzen 7600X + RTX 5070 + ASUS B650E is a state-of-the-art monster machine. The performance of the components themselves is more than sufficient.
Nevertheless, the cause of such a wide range of simultaneous kernel errors (50, 9F, D1, 117, 193) is undoubtedly that “the May 12 Windows Update is causing bugs by interfering with power management for the latest GPUs (RTX 50 series) and the memory allocation for inter-process communication (IPC) required by DICE’s game engine.” It is also possible that issues from the April Windows Update are still lingering.
"Comparing old and new logs reveals that after a cache clear, the error has changed to a fatal memory conflict error, 0xD1 (DRIVER_IRQL_NOT_LESS_OR_EQUAL).
The primary culprit is not the game (DICE), but rather a mismatch between kernel memory management after Windows Update and the NVIDIA driver’s power and rendering processes that are being thrown off by it. The OS infrastructure is holding back the potential of the latest RTX 5070, causing it to go into a panic state.
Sorry for the long post.
