Need assistance with a Ryzen 5000 overheating issue? Let me know what you're experiencing.
Need assistance with a Ryzen 5000 overheating issue? Let me know what you're experiencing.
I assembled projects for friends unfamiliar with the process. They purchase components and I handle the assembly. Yesterday night I finished two builds and started testing their overclocks. After about ten minutes, both systems would suddenly spike in temperature and shut down immediately, without any throttling. This is unusual since both reached similar temperatures during Prime95 stress tests. How can two machines experience the same problem? System 1: AMD R5 5600X @ 4.5GHz @ 1.3v Corsair 240mm AIO (pump/fans Extreme) Asrock Phantom Gaming B550 LLC2 System 2: AMD R9 5950X @ 4.5GHz @ 1.2v Corsair 360mm AIO (pump/fans Extreme) ASUS Strix X470-F LLC4 The overclocks performed well in Cinebench, reaching 80*C. Then I ran P95 and the temperature jumped from idle (45*C) to 60*C—lower than what I see in Cinebench. After roughly ten minutes, it climbed by about one degree per second until shutdown. This happened with both systems. I haven’t seen such a rapid rise before. Typically, the temperatures in Cinebench should match those in P95. On the machines I’ve already overclocked, this is rare. What could be causing the sudden heat spike? Is there something missing with these new chips?
Increased audience engagement for both push and pull actions, maintaining a 100% target consistently. Issue remains unresolved.
The thermal paste inside the AIOs is likely stock. You might want to consider Thermal Grizzly Kryonaut for an upgrade.
Uncertain about the thermal paste; temperatures fall quickly after stopping tests, indicating heat moves to the pump.
If you've tried this paste in other configurations and it worked well during P95, I'll rely on your assessment. I'm willing to test my 5900X (on a Corsair 280 AIO) at stock settings to confirm the outcome. I had a poor experience with overclocking and even undervolting, so I prefer to stick to baseline performance.
The unusual aspect is that in P95 it remains stable at 60*C for about ten minutes, but then things go haywire afterward. I’d appreciate it if it were a gradual decline, but my spike is something completely new. Based on cooling solutions, if an XYZ cooler removes 10 units of heat per minute and the processor generates 11 units, you end up with a net gain of +1 unit per minute, leading to eventual throttling or shutdown. The speed at which it rises is consistent here at +1/min. In my situation, I have two systems that stay steady for ten minutes before suddenly jumping +X, causing them to overheat.
60C plateau on the 5950X? What's your ambient temperature? I'm still working here and can't test right now. In the meantime, would you mind running P95 with the CPUs at their default settings?
At the 10-minute mark of my stress test, all cores reached 3.7-3.8Ghz. Maintained a consistent 66°C temperature.
[Mar 12 17:43] Initiated a steady self-check on your machine. [Mar 12 17:43] Review stress.txt carefully. Select Test/Stop to conclude the evaluation. [Mar 12 17:43] Executed Test 1 with 7,000,000 iterations of Lucas-Lehmer for M83839 using FMA3 FFT of 4K length. [Mar 12 17:45] Proceeded to Test 2 with identical parameters for M82031, also in 4K FFT. [Mar 12 17:47] Continued with Test 3 applying the same method on M79745, maintaining a 4K FFT size. [Mar 12 17:49] Successful self-test completed in 4K! [Mar 12 17:49] Run Test 1 again with 6,000,000 iterations of M104799 using 5K FFT. [Mar 12 17:51] Followed up with Test 2 on M102991, utilizing a 5K FFT. [Mar 12 17:54] Performed Test 3 for M100705 with a 5K FFT. [Mar 12 17:56] Achieved successful self-test in 5K with favorable pass values. [Mar 12 17:56] Test 1 completed successfully with 5,000,000 iterations of M125759 using 6K FFT, Passes: 128, 48, clm=2. [Mar 12 17:58] Conducted Test 2 on M125281 with identical settings, also achieving a pass. [Mar 12 18:00] Executed Test 3 for M123713 using a 6K FFT, with Pass1=128, Pass2=48, clm=2. [Mar 12 18:03] Final self-test cleared in 6K with all criteria met. [Mar 12 18:03] Successfully finished Test 1 with 3,200,000 iterations of M172031 using 8K FFT, Pass1=128, Pass2=64, clm=2. [Mar 12 18:05] Completed the Torture Test across 9 trials in 21 minutes—no errors, no warnings. [Mar 12 18:05] Worker terminated. Performance reached a maximum of 68C. Strong indication it’s overclocking. Possibly preventing cores from downclocking, causing excessive heat buildup that the cooling system can’t manage efficiently, leading to thermal throttling. Hypothesis still under consideration.]