Question hard freeze when gpu temperature reaches ~60C

Apr 5, 2019
System gets into hard freeze when GPU temperature reaches nearly 60C, no networking, no sysrq, no signal to monitor and gpu fans are spinning at ~40% (~1200 rpm).
Note, temperature, not load/power comsumption - if I set fans to run at max it takes longer to reach hard freeze temp, so in particular, it's enough to fully pass glmark2 (while with driver managed cooling it freezes on 3-4 scene).
Tested with rx580 and rx470, one card per test. rx470 is proven to be good on another setup.
tested on different linux kernels (2016-2019) and window 10 with different radeon driver versions.

HW info:
AsRock x370 Fatality Gaming X, bios 4.50
Chieftec GPS-1250c (cwt, considered strong middle quality psu on market in my country)
Ryzen 7 1700
GOODRAM Irridium 12GB (4 + 8) 2400Mhz
Sapphire RX580 Nitro+
Sapphire RX470 Nitro+

Why I do not think that it's PSU:
apparently, card is taking even a bit more power with fans manually set to max rpm, though there is no freeze in this case for same "freeze" load with driver managed cooling.
I will try to get another psu and run test with it, but my intuition suggests me that the problem is with mobo (for now its also a problem to test on other board).

Maybe somebody faced similar/same issues.
Thanks in advance for your comments and help.
Last edited: