F5F Stay Refreshed Hardware Desktop XMP configuration leading to GPU driver failure Issue arises when XMP settings trigger instability in GPU drivers.

XMP configuration leading to GPU driver failure Issue arises when XMP settings trigger instability in GPU drivers.

XMP configuration leading to GPU driver failure Issue arises when XMP settings trigger instability in GPU drivers.

D
Draker59
Member
126
01-27-2020, 11:39 AM
#1
I recently faced strange errors after upgrading my GPU, where the drivers would frequently crash while playing games. I consulted Microsoft, EA, and Nvidia support, but they couldn’t resolve the problem. I then reduced the RAM speed from 3600 to 3000 and stopped getting crashes. It’s puzzling given my specs: 3060ti, 16GB 3600 Kingston Fury, i5 12400f, and MSI B660A Pro. How could this be linked?
D
Draker59
01-27-2020, 11:39 AM #1

I recently faced strange errors after upgrading my GPU, where the drivers would frequently crash while playing games. I consulted Microsoft, EA, and Nvidia support, but they couldn’t resolve the problem. I then reduced the RAM speed from 3600 to 3000 and stopped getting crashes. It’s puzzling given my specs: 3060ti, 16GB 3600 Kingston Fury, i5 12400f, and MSI B660A Pro. How could this be linked?

I
Isoba5101
Junior Member
3
01-30-2020, 04:49 AM
#2
The memory controller on the 12400F performs quite poorly. I think if you pushed it to 3600MT/s with a stress test like Y cruncher VST, it would likely fail right away. Only certain high-quality chips can handle that speed consistently, and I suspect yours is just adequate enough for basic operation but not stable at that rate. A full 3466MT/s should work without issues, though.
I
Isoba5101
01-30-2020, 04:49 AM #2

The memory controller on the 12400F performs quite poorly. I think if you pushed it to 3600MT/s with a stress test like Y cruncher VST, it would likely fail right away. Only certain high-quality chips can handle that speed consistently, and I suspect yours is just adequate enough for basic operation but not stable at that rate. A full 3466MT/s should work without issues, though.

S
SpiRiiT
Junior Member
32
01-30-2020, 04:58 AM
#3
I stored it at 3600 for about four months after moving from 1050ti to 3060ti. It began to fail after that change. Not sure why the GPU is causing all these issues.
S
SpiRiiT
01-30-2020, 04:58 AM #3

I stored it at 3600 for about four months after moving from 1050ti to 3060ti. It began to fail after that change. Not sure why the GPU is causing all these issues.

P
PimOla_PvP
Member
166
02-02-2020, 08:24 AM
#4
PCI and RAM are managed by the CPU.
P
PimOla_PvP
02-02-2020, 08:24 AM #4

PCI and RAM are managed by the CPU.

G
GewoonMex
Member
88
02-02-2020, 09:10 AM
#5
ryzens cpus tend to be picky about RAM, which explains why compatibility issues arise. It’s usually not a full match. Consider upgrading to a more reliable option or stick with 3000 MHz. Isn’t RAM also important? On my 3600, the Hyundai chips didn’t work above 3200 MHz, while the Samsung ones did fine at 3200. My guess is either better RAM speed or higher voltage requirements. But I’m not sure what the maximum limits are for combining OPs RAM and CPU speeds.
G
GewoonMex
02-02-2020, 09:10 AM #5

ryzens cpus tend to be picky about RAM, which explains why compatibility issues arise. It’s usually not a full match. Consider upgrading to a more reliable option or stick with 3000 MHz. Isn’t RAM also important? On my 3600, the Hyundai chips didn’t work above 3200 MHz, while the Samsung ones did fine at 3200. My guess is either better RAM speed or higher voltage requirements. But I’m not sure what the maximum limits are for combining OPs RAM and CPU speeds.

T
thebomb1216
Member
76
02-12-2020, 08:09 AM
#6
RAM overclocking is what the OP meant, not what they were asking about. Stands aren't assured to exceed the XMP limits (like 3200), meaning results vary—some succeed, others don’t. The key voltages (VCCSA mainly) needed for 3600MT/s are capped below 1V on certain chips. On past Intel processors and unlocked 12th-gen boards, you’d need 1.2–1.4V to reach higher speeds than 3200MT/s. Thus, achieving such speeds is unpredictable. I suspect the issue for the OP’s setup is a minor memory controller instability at 3600MT/s. With older 1050 Ti systems, little stress was applied, so problems didn’t appear. Newer 3060 Ti chips hit the memory more often, increasing crash chances. This can be confirmed with a stress test like Y Cruncher VST.
T
thebomb1216
02-12-2020, 08:09 AM #6

RAM overclocking is what the OP meant, not what they were asking about. Stands aren't assured to exceed the XMP limits (like 3200), meaning results vary—some succeed, others don’t. The key voltages (VCCSA mainly) needed for 3600MT/s are capped below 1V on certain chips. On past Intel processors and unlocked 12th-gen boards, you’d need 1.2–1.4V to reach higher speeds than 3200MT/s. Thus, achieving such speeds is unpredictable. I suspect the issue for the OP’s setup is a minor memory controller instability at 3600MT/s. With older 1050 Ti systems, little stress was applied, so problems didn’t appear. Newer 3060 Ti chips hit the memory more often, increasing crash chances. This can be confirmed with a stress test like Y Cruncher VST.

Z
Zologa19
Member
65
02-12-2020, 09:06 AM
#7
No, the BIOS isn't fully updated yet.
Z
Zologa19
02-12-2020, 09:06 AM #7

No, the BIOS isn't fully updated yet.

F
Fullalexis10
Member
167
02-14-2020, 04:41 AM
#8
Update planned, will retry in 3600 seconds
F
Fullalexis10
02-14-2020, 04:41 AM #8

Update planned, will retry in 3600 seconds