This post is more than 5 years old

1 Rookie

 • 

77 Posts

3621

April 7th, 2018 10:00

Random freezes and BSODs on T7910

I'm getting random random system freezes about once per week, since about a month ago. At some point, under high load, the system just freezes: even mouse is not moving, no BSOD is generated. 

The system is T7910.

The issue might be related to running specific CUDA 9.0 code, but I doubt that.

  1. Well, usually even worst CUDA errors just crash a the videodriver, and Windows is able to cope with that.
  2. Almost the same code was working OK on CUDA 8.0, and it works OK with CUDA 9.0 for about a week now. 

First, I've tried to "Save Controller Events" via pre-boot LSI SAS config tool (without any luck).

Next, I've found the following in the Windows Event Log:

Event 11, LSI_SAS3i, The driver detected a controller error on \Device\RaidPort0.

Sometimes it's RaidPort0, sometimes it's RaidPort1. See more details on intel forums (cause here my post is getting automatically banned when I add it). First such event is from 2018-01-04. Not sure why it started, maybe some Windows update? Their number was greatly increased since end of March or so. Here's a histogram:

all_hist.png.

Trying to fix this, I've updated

  • BIOS to A25 (changelog says "Updated Chipset support")
  • "LSI 3008 Firmware, Ph15" to "15.01.00.00 ,A04"
  • maybe something else. not sure.

However, I was still getting a fresh portion of those events after every system boot.

Next, I've noticed "LSI SAS3008 Windows Drivers for Win10-64" to "2.51.21.0 ,A04" update. I was slightly more difficult to install, but it actually fixed the issue. I'm no longer getting those events now. This is the first thing I'd like to report. Cause googling those message was giving almost no relevant information. Hope this would help someone..

1 Rookie

 • 

77 Posts

May 8th, 2018 04:00

First, it looks like images I've uploaded to my posts are broken. So - here's an alternative version of the same picture:

 

Next, as discussed on intel forums, for me the solution was to replace the Intel driver with Microsoft "Standard SATA ACHI Controller". 

No Events found!

Top