- Joined
- Jun 24, 2015
- Messages
- 7,566 (2.34/day)
- Location
- Western Canada
System Name | ab┃ob |
---|---|
Processor | 7800X3D┃5800X3D |
Motherboard | B650E PG-ITX┃X570 Impact |
Cooling | U12A + T30┃AXP120x67 |
Memory | 64GB 6000CL30┃32GB 3600CL14 |
Video Card(s) | RTX 4070 Ti Eagle┃RTX A2000 |
Storage | 8TB of SSDs┃1TB SN550 |
Display(s) | 43" QN90B / 32" M32Q / 27" S2721DGF |
Case | Caselabs S3┃Lazer3D HT5 |
Power Supply | Corsair HX1000┃HDPlex |
Just started using the PC since yesterday.... Was verifying war-zone and updating CS:GO and bam another error!
Edit: I've tried 1.45v to no improvement. Could this be motherboard related?
I've got a Noctua NH-D14 on this CPU so i dont think I have an airflow issue and I can likely push 1.4v-1.5v without any heat issues. To be honest I've only got a GTX1060 so heat wont be too big an issue
I agree this is a ram issue, I am just struggling to make it fail any test which is what the company will do when RMAing it.
Will be trying HCI - do i need to run multiples of this overnight? I Saw another tool but paid to test all the ram in one go
I will be doing something about this very soon. I have another kit of the exact same ram from a friend to try tomorrow.
Passive airflow from a neighboring CPU cooler is good for ballpark 1.45V, but 1.5V is where you start to want a dedicated small fan for just your RAM, especially given B-die temp characteristics, because 1.5V is getting serious and into unsafe voltage territory for a lot of ICs that are not B-die. Long story short, if 1.45V doesn't help, this B-die isn't anywhere near acceptable performance for B-die.
Anyways, let's not get ahead of ourselves. HCI's free version is limited to about 3072MB per instance, so just run simultaneous instances in which each of them is testing 3072MB (just type it in the box), then once you get close to about 2GB of remaining free memory (use HWInfo's reading), run an extra instance that tests less, maybe 1-2GB, to top it off. Then just leave them to run until either 400% completion (which is pretty certain for most intents and purposes) on all of them, or overnight. It will let you know if errors occur. HCI errors are not acceptable for stability.
IMO HCI gets the memory hotter than MT86 does, make sure with HWInfo in the background that the sticks are staying below 50C.
I wouldn't say it's an impossibility that the board is playing a part in this, but knowing the usual quality of LPX B-die it's much less likely. Hopefully after you test this kit, you can pop in the other one and see if it makes a difference.
Do remember to run both chkdsk and sfc /scannow in Powershell to make sure that Windows is eliminating existing disk/file integrity issues. Can't remember the exact commands for chkdsk but you want the full one that's scheduled to run at the next system reboot (Windows logo, spinning circle and chkdsk status messages running below it).