|
Interesting. How did you detect it? Swap out modules or run something like memtest86?
This last crash at 7:13 this morning mentioned something about a chrome.sys driver. This crash happened when I launched chrome.
Quite oddly, google returns no info about this file.
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
Swap modules.
RAM testing tools really don't find intermittent faults. They're good for finding solid faults though, which are surprisingly rare in my experience.
|
|
|
|
|
just swapped the modules. I have 64GB of ram in this unit. The crash seems to happen when I have a number of VMs running. This would certainly push the ram usage up.
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
If it happens relatively consistently when pushing memory to its limit, take them all out. Try to reproduce it again with a single module still in it. If it doesn't happen, take it out and replace it with one of the other ones, and try again. With luck, you might be able to narrow it down to the one faulty module.
Even if you find one, keep going until you've gone over all of them.
Once you're pretty confident you've found which is faulty, add the others by themselves. Then re-add the bad one. The idea is that you keep trying until you're very confident you've isolated it.
It's a time-consuming process, but might be worth it in the end.
|
|
|
|
|
(obligatory it works on my computer).
Never had issues with solid state hardware (motherboard, cpu , memory, ssd, gpu... )
Most/All of the time it was with hard-drives starting to have bad sectors.
It's your work computer, I'd start looking to change it now and hope you can find something that is not too backorder.
5 years is not that long, but you should be able to find something shiny to replace it.
good luck and do backups.
CI/CD = Continuous Impediment/Continuous Despair
|
|
|
|
|
This can easily happen if RAM got fragmented from too many allocations and deallocations. Usually best solution in this case is to increase size of your page swap file. Should be 4 times of physical memory (make sure the value is in the power of 2).
|
|
|
|
|
Valid point, but I'm not really pushing the machine that hard. Sure, I've got a few browsers up and VS Studio assorted editions. The biggest thing for me is that I made sure to install enough RAM to handle multiple VMs. I see very little paging.
Plus - I've seen the crash at very odd times with little or nothing running (but with the VMs operational).
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
To add to wahat Dave said, there are a huge range of faults, which are spectacularly hard to pick up with a memory test: stuck or floating address lines are always a good one (and can appear only after a period of use). If a AL gets stuck, the wrong memory is affected but it reads back as good - the only way to spot it is to write different values to each memory location and then check they are what and where you expected, and that they don't "echo" in a different location as well. That's just not practical in a reasonable timeframe for a memory tester because the sun will change to a red dwarf first ... 
If it's floating (i.e. a track has broken) then it fails sometimes but seems to work much of it.
Swap the module out, see if the problem goes away.
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
"Common sense is so rare these days, it should be classified as a super power" - Random T-shirt
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
you allude to my question - in all my years futzing around with PCs, I have yet to see a ram test fail. I get the address bar issue and understand the concept. It really boils down to how much time do you have?
It's my understanding that when these memory boards are made they can be tested resting on a bed of pins. The test system can easily check each circuit. Once it gets to the PC, well...
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
charlieg wrote: 1) Using RAM test tools, have you ever found a bad RAM? Replaced it and all was well?
Yes.
Your issues may be related to the memory controller though, so you can try and try and never catch the issue until it manifests. Issues on ground (half signals, floating signals) on some line may cause that, or heat altering some line / memory cell characteristics enough to kill it.
GCS d--(d-) s-/++ a C++++ U+++ P- L+@ E-- W++ N+ o+ K- w+++ O? M-- V? PS+ PE- Y+ PGP t+ 5? X R+++ tv-- b+(+++) DI+++ D++ G e++ h--- r+++ y+++* Weapons extension: ma- k++ F+2 X
|
|
|
|
|
Yes. I have a nice pile of defected memory...
"The only place where Success comes before Work is in the dictionary." Vidal Sassoon, 1928 - 2012
|
|
|
|
|
Kornfeld Eliyahu Peter wrote: defected
So it is working... but for the enemy.
GCS d--(d-) s-/++ a C++++ U+++ P- L+@ E-- W++ N+ o+ K- w+++ O? M-- V? PS+ PE- Y+ PGP t+ 5? X R+++ tv-- b+(+++) DI+++ D++ G e++ h--- r+++ y+++* Weapons extension: ma- k++ F+2 X
|
|
|
|
|
Marketing? 
Software Zen: delete this;
|
|
|
|
|
I may have in my desktop. I'd been getting crashes that my crash analysis tool suggested were probably memory related over the last year and which got markedly worse recently, memtest86 ran clear overnight. Friday morning I down-clocked my ram from DDR3-2400 to DDR3-1600 and have been crash free over the weekend; long enough to suggest the issue is solved.
That said I'm sure sure it's an inherent memory problem; my CPU is watercooled so there's not a ton of airflow around the ram; and since early this year I've been using an aircooled GPU again (and more recently a hotter/faster one). My ram is hot to the touch, so it's possible I'm just seeing an overheating issue that memtest86 by leaving the rest of the system at idle didn't trigger. I'm probably going to order a ram cooler later this week and see if that lets me restore to full speed operation.
Did you ever see history portrayed as an old man with a wise brow and pulseless heart, weighing all things in the balance of reason?
Is not rather the genius of history like an eternal, imploring maiden, full of fire, with a burning heart and flaming soul, humanly warm and humanly beautiful?
--Zachris Topelius
|
|
|
|
|
Well after suffering a weekend and a few days of BSODs, I have new ram installed. If it hasn't crashed in 8 hours, Crucial is getting their ram back.
Charlie Gilley
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
Has never been more appropriate.
|
|
|
|
|
How many modules?
If more than one (this is a laptop afterall ...) be sure to "jiggle" them. Power down/power up see. If that doesn't help, try swapping them. Power down/power up see.
How to "jiggle" them, you ask? Obviously you've got to take the shell off and be able to see the edge-connectors. Etc ...
On a big graphics-card-heated desktop where the first memory slot is located exactly due north of that hotspot, that DIMM is always coughing and wheezing. Swap always seems to do the trick.
And something I've never tried but might work ... duct tape a tab of paracetemol to each module. Power-down/power-up. See.
|
|
|
|
|
The only bad memory I have is in my skull. 
Entropy will win in the end.
|
|
|
|
|
well the jiggle thing is a valid suggestion. I removed both SO-DIMMs, cleaned their contacts it alcohol, cleaned the sockets with the same alcohol, let it dry.
crash after many hours.
Swapped the SO-DIMMs per one suggestion - crash after some hours.
I'm going to pull the trigger for new memory. Interestingly, my SO-DIMMS are labeled 16GB. They show up in the BIOS and Windows as 32GB. 4+ years ago, I think I looked at the plastic case containing the SO-DIMMS. So I tried chatting with Crucial about an exact substitute. My wife wonders why I drink. It was hopeless. I mean utterly braindead hopeless.
As another poster said, I'll download the analyzer...
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
The saga continues. Two new updates.
First, yesterday while the laptop was idling, I received a BSOD I've never seen before: driver_overran_stack_buffer. Hmm. According to Microsoft:
"A driver overran a stack-based buffer (or local variable) in a way that would have overwritten the function's return address and jumped back to an arbitrary address when the function returned.
This is the classic "buffer overrun" hacking attack. The system has been brought down to prevent a malicious user from gaining complete control of it."
Double hmm. Full virus scan found nothing. Of course, all of the other references indicate bad drivers, which I have had a terrible time with Microsoft replacing drivers. Over the weekend, I'll run a deep scan booting off a USB drive. We'll see if anything else pops up.
----------------
Second weird issue - RAM. Since I've had this laptop, I have had 64GB in it. I purchased two Crucial SO-DIMMs. When I pull up the BIOS, it sees 64GB. When I check task manager, it says 64GB. But the label on each module says 16GB. I sent the picture to Crucial. They confirmed 16GB. Suggested maybe a labeling issue with the vendor. Dude, it's YOUR label on the SO-DIMM  . Support did not understand. I'm in touch with Micron to determine the true density of the chips which I'm pretty sure are sized for 32GB.
The entire point of the RAM exercise is if I replace the RAM, I need to make sure I order the correct part. I got these from Amazon which is totally useless from a customer support perspective. I'll go to NewEgg next time.
Charlie Gilley
|
|
|
|
|
As an aside, does anyone have a recommendation for a good virus/malware (I guess they are the same these days).
I'm specifically looking for something I can boot off a USB that is an excellent scanner.
Charlie Gilley
<italic>Stuck in a dysfunctional matrix from which I must escape...
"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759
|
|
|
|
|
fliers put soarers in disarray (10)
|
|
|
|
|
Is that a word?
"The only place where Success comes before Work is in the dictionary." Vidal Sassoon, 1928 - 2012
|
|
|
|
|
I think so, not knowing which word you have in mind.
|
|
|
|
|
The right one of that flying dinosaur
"The only place where Success comes before Work is in the dictionary." Vidal Sassoon, 1928 - 2012
|
|
|
|
|
Pterosaurs - that's all I can make of it
"I didn't mention the bats - he'd see them soon enough" - Hunter S Thompson - RIP
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
