old i7 920 issues

timbo2410

timbo2410

timbo2410

Associate
Joined
27 Dec 2009
Posts
764
Ok, I've had the overclock for the old girl for a while now all stable at 4.2ghz. All stress testing done and sorted and running fine. I upgraded to win10 x64 and also installed 12gb of tripple channel 1600mhz ram. Everything is fine and cool. Now i've had a couple of reboots that I haven't been able to track down. It happened overnight while i was downloading and also running Nero 2018 (converting an mp4). Today I had an error that i managed to catch during a crash/reboot (while converting an mp4 in Nero). It said there was a memory error. I have run all stress tests I can think of and memory is fine. I then went to event viewer and found that Win10 is reporting that "ALL" of my Samsung hard drives are failing? Now I have a New SSD for my OS and I have 3 1tb Samsung HDD's that I use for storage. I cannot believe that all 3 are failing at the same time. Can there be a windows reporting issue? What do I do from here?
 
Being "Samsung" drives, I imagine they are now around... 6-7 years old at least, so failures seem likely at this age.

Firstly your data is most important so I'd start by downloading CrystalDiskInfo to indeed verify if they are failing (identified by amount of errors). https://osdn.net/projects/crystaldiskinfo/downloads/68590/CrystalDiskInfo7_5_0.exe/

You can also use SeaTools for Windows to perform a long disk check whilst still being able to use the PC.
https://www.seagate.com/gb/en/support/downloads/seatools/seatools-win-master/
 
thanks for the reply :-) I'm just running sfc/scannow .. is it worth doing or shall I just go straight to CrystalDiskinfo? Surely I wouldn't have all 3 drives go at the same time?
 
just ran crystaldiskinfo .. 1 of the 3 drives is saying caution .. the other 2 seem ok .. although they are all at around 18000 hours of use (wow) .. while everything is still working should i just replace the 3 of them? Didn't really want to fork out for more for 3tb worth of SSD .. mechanical seems very cheap .. ideas welcome ..
 
Memory overclocked or just CPU?

I'd leave it running memtest for a few hours to rule out a RAM error. Transcoding can use seldom used instructions in the CPU, could also be related to that. I'd dial the OC back a notch, see if it helps.
 
the memory is using xmp to run at it's rated 1600mhz. I've had memtest running overnight after I installed it and it ran just fine. I only caught the memory error as the pc went to blue screen for a few seconds prior to rebooting. The HDD errors were picked up when i went to look at Windows events .. the events do NOT show any memory errors at all .. just the imminent HDD failures ..
 
So, if I keep having a "memory" error then best to reduce the overclock back to 4ghz. I have run aida for the last 2 hours at 96% stress test on ram .. still running just fine ... so ... if it's not the ram .. could the failing hdd cause a weird memory fault on crash? .. if the memory fault is not showing in the events log?
I have 2 new HDD's arriving Tuesday to replace the 3 that are suspect .. if the issue keeps happening after that I shall drop the overclock back to 4ghz and see how it goes then.
 
Try replacing the sata cable and try a different sata port for the failing hdd

Try lowering the of down a bit to maybe 4ghz instead of 4.2ghz as this may help
 
You must log in or register to reply here.
Back
Top Bottom