BSOD Memory_management

[Regeneration]

1. Take off the CMOS battery.
2. Re-seat the CPU and check the pins aren't bent.
3. Take off the memory and clean it and the socket.
4. Put the CMOS battery back.
5. Load default BIOS configuration.
6. Use AIDA64 to read the RAM's SPD information, and enter timing manually in the BIOS.
7. Make sure DDR, IMC, cache voltages are set correctly.
8. Some motherboards change the memory RTL value on reboot. Try to set RTL manually.
9. If there is a RAM turbo feature in BIOS, try to disable it.
10. Try to increase DDR voltage from 1.5v to 1.52v.

Use MemTest64 and Prime95 blend to test for stability. But long runs, if a single memory chip is failing, it will take time to catch it.

 

[Shambles1980]

try and loosen the memory timings and run at lower ram speeds.

 

[IceScreamer]

1. Take off the CMOS battery.
2. Re-seat the CPU and check the pins aren't bent.
3. Take off the memory and clean it and the socket.
4. Put the CMOS battery back.
5. Load default BIOS configuration.
6. Use AIDA64 to read the RAM's SPD information, and enter timing manually in the BIOS.
7. Make sure DDR, IMC, cache voltages are set correctly.
8. Some motherboards change the memory RTL value on reboot. Try to set RTL manually.
9. If there is a RAM turbo feature in BIOS, try to disable it.

Use MemTest64 and Prime95 blend to test for stability.

I actually did exactly all that, apart form the CPU re-seat, will have to try that as well. All the voltages and timings are correct, will try to loosen them up a bit or adding a bit of voltage. The BIOS does have the memory turbo/normal option and that was the first thing I changed and it gave a BSOD on both options.

 

[Bill_Bright]

I replaced the PSU for now, as soon as I get my hands on some memory stick I'll check on that.

Hold on! One step at a time. You replaced the PSU. Wait and see if that fixed your problem before testing, removing, replacing, reseating, increasing, loosening or changing anything else.

 

[IceScreamer]

Hold on! One step at a time. You replaced the PSU. Wait and see if that fixed your problem before testing, removing, replacing, reseating, increasing, loosening or changing anything else.

Of course, I really wanna get to the bottom of this, I'll wait and see if this fixes it, and if it doesn't, move forward. Problem is, I can't seem to force the BSOD, everything worked fine on W10 for a couple of days and this morning it just crashed randomly while browsing the internet.

 

[Bill_Bright]

Problem is, I can't seem to force the BSOD

Duplicating a problem "at will" is always a desired goal when troubleshooting any problem. If you can duplicate it "at will", then you typically can easily (or more easily) isolate the fault, and repair or replace the offender. That's why troubleshooting intermittent problems are the most difficult.

But that is also why you must take one step at a time. If you do several things at once, and it fixes the problem, you will not know what step fixed the problem - in other words, you don't learn how to prevent recurrence in the future, or how to fix it should you encounter it again (on this machine or another).

 

[IceScreamer]

Duplicating a problem "at will" is always a desired goal when troubleshooting any problem. If you can duplicate it "at will", then you typically can easily (or more easily) isolate the fault, and repair or replace the offender. That's why troubleshooting intermittent problems are the most difficult.

But that is also why you must take one step at a time. If you do several things at once, and it fixes the problem, you will not know what step fixed the problem - in other words, you don't learn how to prevent recurrence in the future, or how to fix it should you encounter it again (on this machine or another).

I know that, but with the things they are now, this is a waiting game, my BSODs can sometimes be weeks apart from what I've gathered.

 

[Bill_Bright]

but with the things they are now, this is a waiting game

Exactly! So "Patience, Young Grasshopper!" One step at a time.

I must also point that BSODs weeks apart may be totally unrelated to each other too! And of course, in that time frame, it is likely Windows and perhaps drivers and other programs will be updated too - that ends up resolving the problem without you ever know what was wrong.

 

[Regeneration]

I actually did exactly all that, apart form the CPU re-seat, will have to try that as well. All the voltages and timings are correct, will try to loosen them up a bit or adding a bit of voltage. The BIOS does have the memory turbo/normal option and that was the first thing I changed and it gave a BSOD on both options.

Gigabyte X58 motherboards used to set extreme RTL values time to time causing memory corruption. The trick was to add +5 to RTL for each channel.

 

[ArbitraryAffection]

This might be completely unrelated but I'mma put it here anyway just in case. A while ago I was getting almost constant BSODs playing Fallout 4 with MEMORY_MANAGEMENT on my PC. I had 2600X and Vega 56 at the time. It was all at stock so I assumed it was windows or the game being derpy. I thought nothing of it but now its stopped.

 

[IceScreamer]

Gigabyte X58 motherboards used to set extreme RTL values time to time causing memory corruption. The trick was to add +5 to RTL for each channel.

I can't seem to find the RTL option to modify.

This might be completely unrelated but I'mma put it here anyway just in case. A while ago I was getting almost constant BSODs playing Fallout 4 with MEMORY_MANAGEMENT on my PC. I had 2600X and Vega 56 at the time. It was all at stock so I assumed it was windows or the game being derpy. I thought nothing of it but now its stopped.

Maybe that happens to me too, but I'd rather find out what's wrong tbh.

 

[Regeneration]

I can't seem to find the RTL option to modify.

Maybe that happens to me too, but I'd rather find out what's wrong tbh.

It's in RAM timing. RTL or round trip latency. Gigabyte has a secret hotkey (CTRL+F1) for advanced options.

 

[IceScreamer]

It's in RAM timing. RTL or round trip latency. Gigabyte has a secret hotkey (CTRL+F1) for advanced options.

Noted, will check after testing with the current change.

 

[Regeneration]

Noted, will check after testing with the current change.

Sorry for the mistake, but it was B2B Cas Delay causing memory corruption on old Gigabyte boards.

Errors from B2B appear fast and random per boot session. You'll need to boot to MemTest86, give it a few seconds, and reboot to MemTest86 again.

 

[IceScreamer]

So the PSU change didn't help, still got a BSOD.
It kept bugging me why did the BSODs start on the day they did, December 10th, so I dug around the logs and whatnot and surprsingly found nothing, no software installed or updated around that time. But, I was scrolling through my local ebay-like site messages and noticed that right around that time (a day before) I sold the Cooler Master AIO and replaced it with my current air cooler.
That cooler has a mounting system that could possibly be overightnened, and I figured it could cause these issues. So I just took everything apart, re-seated the CPU and the cooler and made sure to tighten everything just enough, so now I wait again. Also checked for bent pins and found nothing out of order.

 

[er557]

It could be a start of a faulty mobo, might wanna dig a bit in that direction.

 

[IceScreamer]

It could be a start of a faulty mobo, might wanna dig a bit in that direction.

I hope not, if it gives a BSOD after this the only thing left is new memory stick which I finally got. I'll wait and see.

 

[eidairaman1]

Typically mem management bsods relate to the ram, after that cpu and mobo

 

[IceScreamer]

Typically mem management bsods relate to the ram, after that cpu and mobo

I know, I mean it's in the name and all. I've done day long tests in memtest86, various versions, with both sticks and different combinations and no errors whatsoever.

 

[eidairaman1]

I know, I mean it's in the name and all. I've done day long tests in memtest86, various versions, with both sticks and different combinations and no errors whatsoever.

I ran into that on DDR1 on my 462 platform, I just happened to kill a module or 2...

 

[IceScreamer]

Just to update on the situation, been more than a week and no crashes whatsoever. Seems the cooler refitting did the job. I'm guessing over tightening the mounting system messed something up the first time, looks ok now.

 

[eidairaman1]

Just to update on the situation, been more than a week and no crashes whatsoever. Seems the cooler refitting did the job. I'm guessing over tightening the mounting system messed something up the first time, looks ok now.

Yeah board warpage can cause a partial contact connection-possible arcing too.

 

[trparky]

over tightening the mounting system

On the processor and heatsink/fan assembly?

 

[Bill_Bright]

On the processor and heatsink/fan assembly?

I have seen this but it was many many years ago. These days, modern mounting mechanisms, like the quarter-turn locking screws, prevent over-tightening so not sure why it still gets mentioned.

 

[IceScreamer]

On the processor and heatsink/fan assembly?

Yea, the cooler has to be tightened to the plastic ring that is mounted on the board itself, by phillips screws, and you gotta do it by feel so I probably overdid it the first time.