PC crash and BSOD when play FF14 only after latest update and window update.

[sigina]

Hi. THanks and I tried disable SLI and use single GPU. It works for a few min before going BSOD.

[Kirika-]

Hi. THanks and I tried disable SLI and use single GPU. It works for a few min before going BSOD.

I see from your original post edit, you fixed the issue

A dead fan on a GPU will certainly cause issues.
Although, generally, a BSOD isn't triggered by a GPU overheating.
Normally, the GPU BIOS will just shut off the GPU when it greatly exceeds thermals, while throttling it to a much lower performance when approaching thermal limits.
Basically, the game would start slowing down on FPS very quickly and noticeably, and then the display would shut off, sometimes with the computer also turning off/rebooting.

You mentioned in your edit that it was a GPU purchased through Aliexpress (unless I misread that??)... If so....
It is very common for sellers on that platform to provide counterfeit or "hacked BIOS" GPUs, based on much older/lower capacity GPU chips and DRAM, that are designed to trick driver packages into thinking it's a newer / higher performance GPU.
As in the card loads the drivers for and reports itself as a 1070ti, but it's hardware is actually a much older GT series or such.
A very very very common side effect of the hardware vs driver miss-match is BSODs, as that driver set will try to make calls and functionality to hardware that isn't actually present on the fake card triggering a crash and BSOD.

Because driver packages and most hardware testing program only rely on the GPU's BIOS to let them know what it is, and the fake cards have that BIOS hacked to report false information, it can be very challenging to ferret out a fake card, with the symptoms created by the hardware/driver miss-match being the only clue as to this issue.
Easiest way to determine if it is a fake 1070ti, is to remove the heat sink, wipe the thermal compound off the GPU chip, and see what model GPU die it says.
For an nVidia GTX 1070 ti, it should read:

• GP104 <-- GTX 1070
• GP104-300-A1 <-- GTX 1070 ti version

https://www.techpowerup.com/gpu-spec...-1070-ti.c3010

If your GPU die is the correct model number for the card, then all is good; Apply fresh thermal compound to the die and reassemble, thus knowing your have a real card and that you just extended it's life more and lowered it's operating temperature a bit with fresh thermal compound (that stuff does dry-out/loose capacity over time, so fresh paste every few years is a good thing)
If, however, it is a fake, then you will know and can make informed claims about the fake and/or to get it replaced.

[sigina]

I see from your original post edit, you fixed the issue

A dead fan on a GPU will certainly cause issues.
Although, generally, a BSOD isn't triggered by a GPU overheating.
Normally, the GPU BIOS will just shut off the GPU when it greatly exceeds thermals, while throttling it to a much lower performance when approaching thermal limits.
Basically, the game would start slowing down on FPS very quickly and noticeably, and then the display would shut off, sometimes with the computer also turning off/rebooting.

You mentioned in your edit that it was a GPU purchased through Aliexpress (unless I misread that??)... If so....
It is very common for sellers on that platform to provide counterfeit or "hacked BIOS" GPUs, based on much older/lower capacity GPU chips and DRAM, that are designed to trick driver packages into thinking it's a newer / higher performance GPU.
As in the card loads the drivers for and reports itself as a 1070ti, but it's hardware is actually a much older GT series or such.
A very very very common side effect of the hardware vs driver miss-match is BSODs, as that driver set will try to make calls and functionality to hardware that isn't actually present on the fake card triggering a crash and BSOD.

Because driver packages and most hardware testing program only rely on the GPU's BIOS to let them know what it is, and the fake cards have that BIOS hacked to report false information, it can be very challenging to ferret out a fake card, with the symptoms created by the hardware/driver miss-match being the only clue as to this issue.
Easiest way to determine if it is a fake 1070ti, is to remove the heat sink, wipe the thermal compound off the GPU chip, and see what model GPU die it says.
For an nVidia GTX 1070 ti, it should read:

• GP104 <-- GTX 1070
• GP104-300-A1 <-- GTX 1070 ti version

https://www.techpowerup.com/gpu-spec...-1070-ti.c3010

If your GPU die is the correct model number for the card, then all is good; Apply fresh thermal compound to the die and reassemble, thus knowing your have a real card and that you just extended it's life more and lowered it's operating temperature a bit with fresh thermal compound (that stuff does dry-out/loose capacity over time, so fresh paste every few years is a good thing)
If, however, it is a fake, then you will know and can make informed claims about the fake and/or to get it replaced.

hi Thanks for replying. I followed your advice and apply fresh thermal paste on both GPU. As I clean it, both GPUs show GP104-300-A1, Nvidia, A taiwan 1744A1 PG9s96.MOP. I assume it the real one.

The one that i bought from Aliexpress if the replacement fan. Do you think it will cause BSOD?

[Kirika-]

hi Thanks for replying. I followed your advice and apply fresh thermal paste on both GPU. As I clean it, both GPUs show GP104-300-A1, Nvidia, A taiwan 1744A1 PG9s96.MOP. I assume it the real one.

You are golden there

The one that i bought from Aliexpress if the replacement fan. Do you think it will cause BSOD?

Ah, I did miss-read, sorry.

I do not see how a fan on a video card can directly cause a BSOD, however it is possible to a fan to be the wrong spec for the video card, as in maybe too much power draw or such, which then could cause the video card to fault.
It could also be spinning the wrong direction, thus effecting the heatsink's cooling efficiency... Or it's Tach signal is the wrong value for the card, so the card's BIOS thinks its spinning slower or faster than it actually is, which would also effect cooling.
As such, it's always best to replace the fan which one from the video card manufacturer or one that is confirmed to be the same spec.
Again, generally a thermal issue with a video card doesn't trigger a BSOD, it normally just triggers a card or system shutdown.

That said, a GTX 1070 ti isn't a video card that can be happy without active cooling when gaming.

My suggestions for fan replacement on your cards:

• See if your card's manufacturer offers OEM fans.
• Try and find a used or "for parts" card of the same make/model and salvage the fans / heatsink from it.
• Research what other brand manufacturers made for their version, that happens to use the same PCB / card base as yours, then look for parts from those cards and "convert" yours.
• Water cooling!!!!! Wraaaaa!!! Go JayzTwoCents on da cardz!!! ... However, that can get expensive... https://www.amazon.com/gtx-1070-wate...070+waterblock .......sooo, next....
• nVidia just this week did a massive price reduction on RTX 30 Series card, ahead of 40 Series roll out next quarter, so new RTX 30 Series cards are now much more reasonably priced than the past 2 years, so consider an upgrade to a single GPU over the SLI you currently have.

Sad to say, but SLI (and AMD Crossfire) has become obsolete for modern gaming, with even mid-range 20 Series cards besting 10 Series SLI rigs.