Hi

I run a 13900k with an Asus z690 Prime-A, DDR 5 ram 2 x 32gb (no XMP). 360mm AIO. RTX 3080Ti.

I built this system in December last year. Everything has been running fine until start of this month (November).

It started with the software I use for work crashing every 5 minutes. Error code being 0xc0000005. I contacted developer and they created 3 debug builds to pin poin the error. The latest crash dump confirmed our suspicion, “It might be a CPU bug”.

Today, I saw another post saying ASUS board are causing high core density CPU’s to crash?

Any testing software that can pin point which part is causing this? All tests are coming back negetive for errors.

From the developer

Now it is obvious it is not a problem in the Editor.
Here is the problematic instruction sequence:
00007FFFCE5489DD  movups      xmmword ptr [rsi],xmm0
00007FFFCE5489E0  movups      xmm0,xmmword ptr [rax-60h]  
00007FFFCE5489E4  movups      xmmword ptr [rsi+10h],xmm1  
00007FFFCE5489E8  movups      xmm1,xmmword ptr [rax-50h]  
00007FFFCE5489EC  movups      xmmword ptr [rsi+20h],xmm0  
00007FFFCE5489F0  movups      xmm0,xmmword ptr [rax-40h]  
00007FFFCE5489F4  movups      xmmword ptr [rsi+30h],xmm1  
00007FFFCE5489F8  movups      xmm1,xmmword ptr [rax-30h]  
00007FFFCE5489FC  movups      xmmword ptr [rsi+40h],xmm0  
00007FFFCE548A00  movups      xmm0,xmmword ptr [rax-20h]  
00007FFFCE548A04  movups      xmmword ptr [rsi+50h],xmm1  
00007FFFCE548A08  movups      xmm1,xmmword ptr [rax]  
00007FFFCE548A0B  movups      xmmword ptr [rsi+60h],xmm0  
00007FFFCE548A0F  movups      xmm0,xmmword ptr [rax-10h]  
00007FFFCE548A13  movups      xmmword ptr [r9-10h],xmm0  
00007FFFCE548A18  movups      xmm0,xmmword ptr [rax+10h]  
00007FFFCE548A1C  movups      xmmword ptr [r9],xmm1  
00007FFFCE548A20  movups      xmm1,xmmword ptr [rax+20h]  
00007FFFCE548A24  movups      xmmword ptr [r9+10h],xmm0  
00007FFFCE548A29  movups      xmm0,xmmword ptr [rax+30h]  
00007FFFCE548A2D  movups      xmmword ptr [r9+20h],xmm1  
00007FFFCE548A32  movups      xmm1,xmmword ptr [rax+40h]  
00007FFFCE548A36  movups      xmmword ptr [r9+30h],xmm0  
00007FFFCE548A3B  movups      xmm0,xmmword ptr [rax+50h]  
00007FFFCE548A3F  movups      xmmword ptr [r9+40h],xmm1  
00007FFFCE548A44  movups      xmm1,xmmword ptr [rax+60h]  
00007FFFCE548A48  mov         rax,qword ptr [rax+70h]  
00007FFFCE548A4C  movups      xmmword ptr [r9+50h],xmm0  
00007FFFCE548A51  movups      xmmword ptr [r9+60h],xmm1  
00007FFFCE548A56  mov         qword ptr [r9+70h],rax  
00007FFFCE548A5A  mov         byte ptr [rsi+44h],bl  
00007FFFCE548A5D  mov         byte ptr [rsi+40h],0  
00007FFFCE548A61  mov         byte ptr [rsi+42h],dil  
00007FFFCE548A65  mov         byte ptr [rsi+41h],dil  
00007FFFCE548A69  mov         byte ptr [rsi+43h],0  
00007FFFCE548A6D  mov         qword ptr [rsi],6Fh
The rsi regirter changes its value somewhere in the middle, because the first and the last instructions write to different memory addresses (the last one was captured by the hardware data breakpoint and generated this crash dump, while the first one was executed normally. And there no instructions in between that can change the rsi register.