I'm checking whether my Ryzen 7 5700X3D is running hotter than it should. It uses a Thermalright Peerless Assassin 120 SE with both fans at full speed, MX-6 paste applied about 1.3 years ago, and a negative Curve Optimizer setting of -10 on one core and -30 on the others. During an OCCT Extreme all-core load, the CPU reaches about 119W package power, runs around 3.85GHz all-core, and peaks near 89°C. CPU voltage is roughly 1.20–1.24V.
My room is unusually warm, normally around 29–33°C and occasionally as high as 39°C. The case has three 120mm intake fans, one 140mm rear exhaust, and one 120mm top exhaust, although the fans are fairly inexpensive. I also have a 5070 Ti, but it is idle during this test. Four DIMMs are installed at 3000 CL16 and 1.35V.
Gaming temperatures are usually lower, though CPU-heavy games can reach around 80°C during summer. I also run AI workloads and other tasks that can keep the processor near full load. In cooler weather, I remember seeing more like 60–73°C, so I'm trying to determine whether the change is simply the hotter room or whether the cooler needs to be reseated or repasted.
Is roughly 89°C at 120W reasonable for this CPU and cooler with a 29–33°C ambient temperature? Would a 360mm AIO, such as an Arctic Liquid Freezer III Pro, make a meaningful difference, and would the Peerless Assassin be sufficient for a 7800X3D?
5 Answers
I would not buy a 360mm AIO just to lower this stress-test number. It could reduce peak temperatures and possibly prevent the CPU from dropping a little clock speed at the limit, but the improvement in gaming would likely be modest. A 280mm or 360mm radiator is stronger than the Peerless Assassin, but it also adds pump noise, cost, and eventual maintenance. Repasting or using a quality thermal pad could be worth trying first.
The Peerless Assassin should be adequate for a 7800X3D in normal gaming and most workloads, provided the case has decent airflow. The 7800X3D can still run near its temperature target under heavy synthetic loads, so a better cooler does not guarantee dramatically lower temperatures. Check the exact case clearance, mounting, and airflow before upgrading; an AIO is optional rather than necessary.
The 3.85GHz all-core clock looks normal for the settings you listed, and the CPU is staying just below its thermal limit rather than showing a serious fault. The difference from your old 60–73°C readings is consistent with moving from roughly 20°C ambient to a room in the low 30s, especially if the case is sitting near the floor or has restricted intake. Idle GPU temperature is not completely irrelevant if the card warms the case, but it is unlikely to be the main cause here.
That temperature is generally expected during an extreme, sustained all-core stress test when the room is already 33°C or hotter. The 5700X3D has a relatively small hotspot over the cache, and the CPU is designed to operate close to its thermal limit under heavy loads. Your gaming temperatures around 65–80°C are much more representative and are not concerning.
The result is plausible, but the temperature delta is a little higher than some people get with the same cooler. At 33°C ambient, 89°C means the CPU is about 56°C above room temperature. Before buying an AIO, reseat the cooler, verify that the mounting pressure is even, inspect the paste spread, and make sure the case panels and front intake are not restricting airflow. A custom fan curve may also help, although running the fans at 100% already suggests airflow or contact is worth checking.
I get under 80°C with a 5700X3D and the same general cooler at 26–28°C ambient, even without running the fans at full speed. The warmer room accounts for several degrees, but checking the mount and case airflow still seems sensible.
The stress-test result is mainly useful because I also run AI and other workloads that can keep all cores busy. I remember getting 60–73°C before, but the room was much cooler, so the summer ambient temperature may explain most of the difference.