P3D v5 BSOD problems

[captain420 3,099]

Im experiencing a lot of BSOD lately from P3D v5 HF1. I have made some notes to see if I can figure out the cause and hope that you guys can help me pinpoint the problem. I am going to do some tests and attach some pictures with stats of my systems while in different stages of the sim.

Here's me idling at Orbx KHAF. In this area I have Flightbeam's KSFO HD, Orbx Northern California, Global Base, Buildings HD, Trees HD, TerraFlora v2, OpenLC NA, Active Sky P3D, Truesky Enabled, RAAS Pro, GSX, Global Ship Traffic v2, AIG AI and running at 4K.

This is me flying past KSFO right before where my machine BSOD.

My CPU Utilization while idling on the runway at KHAF

My GPU Utilization while idling on the runway at KHAF

Edited May 13, 2020 by captain420

[Skywolf 992]

It looks like your overclock is failing.  Lower some clocks.  See on Utilizations - CPU 2 and 4 is at 21%ish rest are high.  Have you done a stress test on your cpu with your overclock.  I will into KHAF for you and post a screenshot.  If it is a scenery problem then Orbx forums will be your best bet

Edited May 13, 2020 by Skywolf

[captain420 3,099]

Hmmm I noticed the max temp my CPU got was 85C right before BSOD. Could it be it BSOD to protect the CPU from overheating? What is the max temperature a CPU can hit before it shuts off the computer and goes into protection? If that's the case, them it could be the problem. Insufficient cooling of my CPU.

Update: This time I'm flying south instead, away from KSFO and noticed that my CPU at it's max temp is 87.

Edited May 13, 2020 by captain420

[Sethos 4,344]

6 minutes ago, captain420 said:

Hmmm I noticed the max temp my CPU got was 85C right before BSOD. Could it be it BSOD to protect the CPU from overheating? What is the max temperature a CPU can hit before it shuts off the computer and goes into protection? If that's the case, them it could be the problem. Insufficient cooling of my CPU.

TJ Max for the 3770k is 105C, then it would start to throttle and if it got any higher, it would do a shutdown. 

Sounds more like a bad component, unstable overclock, faulty RAM stick, wonky driver or the likes. 

 

 

Edited May 13, 2020 by Sethos

[captain420 3,099]

1 minute ago, Sethos said:

Sounds more like a bad component, unstable overclock, faulty RAM stick or the likes. 

RAM is not faulty, already performed a memory test. Don't see how my OC is unstable. All these years I've been running without any problems and I play many other games on my machine as well. I'm thinking it's a heating issue.

Edited May 13, 2020 by captain420

[Sethos 4,344]

Just now, captain420 said:

RAM is not faulty, already performed a memory test. Don't see how my OC is unstable. All these years I've been running without any problems and I play many other games on my machine as well.

Overclocks can become unstable as silicone degrade due to usage and heat over the years, so that can easily happen. You need to analyze your BSOD error message, find out how to get the most accurate dump and data reading, then proceed from there. Because a BSOD isn't just a general thing, it holds a ton of information pertaining to the issue. Also, how did you perform said memory tests?

[captain420 3,099]

[captain420 3,099]

Not sure how to get a detailed log and analyze it but this is what Bluescreenview captures:

[Sethos 4,344]

22 minutes ago, captain420 said:

Yeah, that's pretty much useless. If you want an accurate test, you need a third party tool that does a lot of passes and completely fills the RAM to actually find out whether there's any read errors or not. The current 'gold standard' is the RAM Test from Karhu Software. It has a small initial purchase fee but it's the fastest tool out there. Some of the other brands can spend 12-24 hours to do enough coverage. You can Google it, check various Reddit threads, it comes highly recommended. Then you let that run to perhaps 1200% coverage, if you manage that with zero errors, you could comfortable rule out your RAM I'd say. 

Your BSODs don't really tell much. At that point, you probably need to get into the more advance stuff with Microsoft's own debugging tools, which you definitely need to Google to use and understand. That will give you a much more detailed view and understanding of the BSOD dumps. 

In any case, you've hit the fun jackpot of obscure troubleshooting. 

Edited May 13, 2020 by Sethos

[cyyzrwy24 127]

10 minutes ago, Sethos said:

Overclocks can become unstable as silicone degrade due to usage and heat over the years, so that can easily happen. You need to analyze your BSOD error message, find out how to get the most accurate dump and data reading, then proceed from there. Because a BSOD isn't just a general thing, it holds a ton of information pertaining to the issue. Also, how did you perform said memory tests?

While I have you all here....similar question...I am reading GPU utilization on my GPU (same monitor) 100% is that normal..? and constantly runs at 100%https://imgur.com/a/6JanctH

 

Edited May 13, 2020 by cyyzrwy24

[Sethos 4,344]

Just now, cyyzrwy24 said:

While I have all here....similar question...I am reading GPU utilization on my GPU (same monitor) 100% is that normal..? and constantly runs at 100%https://imgur.com/a/6JanctH

 

You seem to be looking at the Max value, that will always be 100% if you've hit that once. The three brackets represents Min, Current and Max I believe. 

[cyyzrwy24 127]

5 minutes ago, Sethos said:

You seem to be looking at the Max value, that will always be 100% if you've hit that once. The three brackets represents Min, Current and Max I believe. 

Nope...I am looking at it now and it is 100% as Value and 100% as a max but min is 0%...just not sure if that's correct to be always at 100%....Thanks

Edited May 13, 2020 by cyyzrwy24

[spokes2112 179]

Having owned a shedding dog along with many of her friends stopping over, a BSOD with ntoskrnl.dll was the alarm clock -- "clean me up you i_d_i_o_t!" (~ once a month) Fine after that.
Max rated temps of the cpu may not mean anything..
Directly underneath your cpu, on the motherboard, there are usually some simple cpu support circuits. They get heated from convection and are not cooled except for air flow. If those start acting up due to heat then your CPU will wonk out -or- send a signal to OS  that something is wrong.
Just cool it down, you should be good.  

[vgbaron 3,627]

7 hours ago, Sethos said:

Yeah, that's pretty much useless. If you want an accurate test, you need a third party tool that does a lot of passes and completely fills the RAM to actually find out whether there's any read errors or not. The current 'gold standard' is the RAM Test from Karhu Software. It has a small initial purchase fee but it's the fastest tool out there. Some of the other brands can spend 12-24 hours to do enough coverage. You can Google it, check various Reddit threads, it comes highly recommended. Then you let that run to perhaps 1200% coverage, if you manage that with zero errors, you could comfortable rule out your RAM I'd say. 

Your BSODs don't really tell much. At that point, you probably need to get into the more advance stuff with Microsoft's own debugging tools, which you definitely need to Google to use and understand. That will give you a much more detailed view and understanding of the BSOD dumps. 

In any case, you've hit the fun jackpot of obscure troubleshooting. 

I have not used the tool you recommend so it may compensate for it but the ideal way to test memory is stick by stick. When all slots are populated the built in error correcting of the mobo can give false reports. Now, if the memory test isolates the sticks, which most do not, then results would be more accurate.

Vic

[Bob Scott 7,721]

1 hour ago, vgbaron said:

When all slots are populated the built in error correcting of the mobo can give false reports.

That would only hold true if it were ECC (error checking and correcting) memory, which would be a very rare find on a consumer-grade mobo.  That's the kind of memory generally found on mission-critical server systems with Xeon CPUs.

I use MemTest 86+, and a thorough multi-pass test profile (10 or more full passes) needs to run for probably 8-10 hours minimum.  An iffy DIMM that only throws the occasional error when under heavy load is not likely to show itself right away.

Given that the minidump shows mostly 3B errors, I'd start by backing off any GPU/VRAM overclock to nVidia stock settings (including a factory overclock), and consider clean install of a different set of graphics drivers.  3B errors are often associated with driver problems.  Notably, however, there was one 101 error listed in there too, and those are usually a symptom of an unstable CPU overclock.

So if the memory test comes back clean after a thorough overnight run, I'd recommend setting chip manufacturer's stock speeds on both the CPU and GPU, and see if the error occurs under those conditions.  If not, it's pretty clearly an unstable OC.  If so, then more butt pain lies ahead.