Processing load is extremely high.
Processing load is extremely high.
Hey everyone, I just upgraded to a 5900x and did a ton of testing. I'm getting really high temps compared to my 3900XT. Here's some info: Ambient temps: 28-32 C (in a heat wave unfortunately) Idle: 50-60 C 100% Load: ~75-80 C stock Cinebench R20, ~90 C w/ PBO enabled Gaming (w/ PBO & Curve offset -20 & all power/current limits at stock, makes barely any difference to temps vs regular PBO or even stock): AC Valhalla (usually ~20% CPU usage) = 70-75 C Forza Horizon 4 (usually ~20-30% CPU usage) = 73-80 C NFS Heat (usually ~40-50% CPU usage, Frostbite games always seem very intensive) = 85-88 C For reference, I have a 360mm AIO, used Noctua NT-H2 thermal paste (applied with one blob in the middle, 4 blobs around), and I have an RTX 3080. I used my 3900XT yesterday (around the same ambient temps) and my temps were ~10+ C cooler. I don't think it's my thermal paste/cooler application, since my coolant heats up a lot more and my fans spin much faster. Any help would be appreciated!
The video mentions temperatures were recorded in a room at around 21°C. That means the readings were 7–11°C lower than your own setup. Adding that to Steve’s 71°C measurement gives an average of roughly 78–82°C. At 4:50, Cinebench temperatures appear near 65°C. Considering the temperature difference, your results match what you observed. Regarding the 3900XT, it likely benefited from improved silicon quality and lower voltage operation due to AMD’s 7nm optimization, resulting in better thermal performance. Variations exist naturally, so your unit might have had a better silicon chance, while the 5900X performed solidly with a 55°C drop under similar conditions. Your AIO also outperformed by about 8°C, which aligns with expectations for its size.
It makes sense now, I realized the numbers were affected by the environment. Your 3900XT ran cooler because of that. Appreciate the clarification!