[SOLVED] Hard freezes, no idea about the guilty.

[Phazoner]

How can I find out the responsible component?

The first freeze came the last thursday playing Halo Reach on 4K/60FPS with a 50-60% of GPU usage after around 10 minutes. Cleaned drivers, had the GPU at 99% for some minutes with Unigine and I was able to play the whole first level without further issues.

The second freeze happened yesterday playing Gears 5, max settings, dynamic resolution with 60FPS target, again after 10 minutes or so. This time I stress tested the CPU with Intel XTU for some minutes and no issues there.

So I'm confused because usually hard freezes are PSU wise and once they start they just keep getting more usual and worse, it makes no sense to me to freeze with Reach with medium GPU usage, not freeze with Unigine with maximum GPU usage and temps (75°C in a 1070), freeze again with Gears 5 (which in this case is GPU intensive and therefore drains some wattage).

Can it be the MoBo or is the PSU messing with me before dying? Are this cheap PCI MoBo testers really useful? I'm all ears.

Thanks a lot for anyone who tries to help me. I'll appreciate it very much.

 

[Darkbreeze]

Full specs?

Motherboard BIOS version?

Exact PSU model?

 

[OllympianGamer]

How can I find out the responsible component?

The first freeze came the last thursday playing Halo Reach on 4K/60FPS with a 50-60% of GPU usage after around 10 minutes. Cleaned drivers, had the GPU at 99% for some minutes with Unigine and I was able to play the whole first level without further issues.

The second freeze happened yesterday playing Gears 5, max settings, dynamic resolution with 60FPS target, again after 10 minutes or so. This time I stress tested the CPU with Intel XTU for some minutes and no issues there.

So I'm confused because usually hard freezes are PSU wise and once they start they just keep getting more usual and worse, it makes no sense to me to freeze with Reach with medium GPU usage, not freeze with Unigine with maximum GPU usage and temps (75°C in a 1070), freeze again with Gears 5 (which in this case is GPU intensive and therefore drains some wattage).

Can it be the MoBo or is the PSU messing with me before dying? Are this cheap PCI MoBo testers really useful? I'm all ears.

Thanks a lot for anyone who tries to help me. I'll appreciate it very much.

However unlikely it could simply be a software issue, I keep getting crashes in the outer worlds on the xbox app and it is the games fault.
But you could always check event viewer if you have another crash, see if there are any issues with drivers around that time.
You could roll back try a slightly older gpu driver.
Run userbenchmark.
Run memtest 86.
Forgetting about gpu utilisation does you gpu actually boost to its correct max clock speed?

 

[Phazoner]

Sorry for the late response, I've been very busy.

Full specs?

Motherboard BIOS version?

Exact PSU model?

MSI MS-B1061 (NUVOTON NCT6685D), it's the motherboard of the MSI Nightblade X2B computer, slightly custom version of a standalone one I guess. According to msinfo the version is 1.0 (SMBIOS 3.0, controller 255.255). I tried to update it because of the support for seventh generation of Intel Core but I didn't accomplished it, trie to start from the USB where the new bios and chipset update was but it kept starting Windows.

I don't know either the PSU, not even the wattage (I'm guessing it is some 550W plus silver one according because the other option would be 350W looking at different builds specs and 350W doesn't seem enouth for my 1070 and a 6700k).

However unlikely it could simply be a software issue, I keep getting crashes in the outer worlds on the xbox app and it is the games fault.
But you could always check event viewer if you have another crash, see if there are any issues with drivers around that time.
You could roll back try a slightly older gpu driver.
Run userbenchmark.
Run memtest 86.
Forgetting about gpu utilisation does you gpu actually boost to its correct max clock speed?

Userbenchmark said everything is correct, it surprisingly said that my SSHD is benchmarking very slow but nothing I've noticed and I think Gears 5 is in the SSD so it shouldn't be the problem. The 6700k is slightly underperforming but is OK.

Memtest86 ran its test for around 3 hours and said everything was OK, 0 issues.

In the first freeze I checked for errors in Windos event viewer (I think that's the english name) and it just pointed to a pair of disks disconnection but them were like disk 6 and disk 8 which don't really exist (and I think there aren't even ports for that much drives.

I still need to check the GPU clock speed but it would be odd. Anyways About its drivers I've been playing with them the past weeks to Gears 5 without an issue, after some time issues started, cleaned drivers, installed the newest and the error persisted so I don't think drivers have nothing to do with this.

I still need to check maximum clock speeds.

 

[Phazoner]

I think I may have found the guilty. The clock speed was racing to 1949Mhz, when the official boost clock for the 1070 Armor OC is 1740Mhz. I've tuned it down with MSI Afterburner to maximum 1740Mhz with its default voltage at that speed.

I think 1949, being a high clock, isn't still unusually high as I've seen other people peaking this speed too, probably some GPU Boost 3.0 mess. As I'm not getting artifacts I just pray for it to be freezing due to natural PSU aging and not it being dying and neither be the graphics card.

Anyways I bought a bigger box for future upgrades and will probably get a Tier A PSU for myself and use mine for my father's PC which has a poor quality one and isn't even being stressed.

Thanks a lot, I will update if I keep getting freezes (I haven't really tested it yet).