The temperatures are normal.
The temperatures are normal.
I was bored today and started experimenting with my 5950X. After a successful run, I felt confident with the temps in check. I ran Cinebench now more for monitoring than for results. I’m still curious about some things I noticed while watching it. While Cinebench is running, I keep checking hwmonitor to see temperatures, power usage, and clock speeds. I observed that my 5950X stays cooler without overclocking during a Cinebench test compared to when it’s idle. When I ran Cinebench R23 without overclock, the temperature dropped to around 50°C. During Cinebench R32 with OC enabled, temps settled at about 78°C for a 256w package with no overclock, and 52°C for a 55-60w package without OC, which is idle. I’m new to this CPU and was surprised by the differences between loads. I’m also wondering if there’s a BIOS setting that helps the CPU stay cooler when idle. It seems like it prefers cooler operation. I’ve read about PBO and its effect on overclocking and Cinebench performance, but I’m still trying to understand everything.
Is your pump consistently operating at full capacity? The name hints at a minor automatic boost, which helps maintain cooler temperatures by applying a small negative voltage offset. It works well for a few extra benchmark points but comes with significantly higher heat levels.
Pump is a D5 model with consistent speed, which shouldn't cause any issues. I'm running it without overclocking, as performance isn't essential for me, even if I consider undervolting. Still, I'm interested in comparing temperatures during full load versus idle, since they remain similar. The loop was definitely warm—first I ran the benchmark, then a 4K test, and finally returned to stock clocks after restarting.
It seems a bit unusual, but are you certain it was truly idle? Could there be background updates or something similar happening? I think idle should be around 0-1%, especially since Ryzen tends to use more power under light loads. By the way, a voltage offset in PBO really helped—my 3600 ran cooler and performed better thanks to lower temperatures, just like my GPU does.
I'm sure it was just idle, though the noise was a bit loud. Core temperatures hovered between 1-3%. I've seen it tend to spike quickly when opening files. The 5950x and waterblock seem to handle it well, but it's odd that after overclocking to 256W TDP, cooling feels much better. I'll keep monitoring it. It's a bit puzzling. My Xeon behaves differently—reaching 48°C under heavy load, yet still runs smoothly. I've upgraded before, but this one keeps pushing limits. It's a tough machine.
I notice it's a bit odd, but that's typical of Ryzen... your idle temperatures should be near normal, which suggests you're getting solid cooling. And those "strange little Ryzen CPUs" are common too—generally they don't overheat much if the cooling works well. P.S.: I've tried three different Ryzen models—2200G, 3600, and 5800X3D—and all stayed around 85°C under airflow. The 2200G might have reached 90°C with a stock cooler and overclocked CPU/IGPU... so ironically, the slowest one was also the hottest.
Your 5950X running with a Noctua D15S hits about 33°C. In cooler conditions, idle falls into the 20–25°C range, while higher temps push it up to around 44°C at full load. If your max temps are within normal limits, you might skip worrying and enjoy a stable 16C/32T setup. :3
I’ve explored further. When the system isn’t busy, it sends 1.45x/1.57x voltage to the CPU. During Cinebench tests, the voltage falls to 1.0, which makes sense but doesn’t clarify why such a high voltage is required when idle. There might be a BIOS setting I should check for that.
I adjusted the voltage to 1.25v and it helped lower the idle temps to around mid-36°C. This seems to be the main problem with the 5950x, or at least what I’ve noticed after some research. The biggest concern for me is the significant 5000-point drop in Cinebench following that small voltage change.
AMD boosting method simply increases voltage for the CPU causing elevated temperatures, particularly during idle. Static voltage tends to provide superior performance compared to stock boosts while maintaining lower heat output. My 3600 operates at over 40°C at idle but stays around 25°C at 4225MHz with all cores at 1.288V (stock only reaches 4150/4100MHz, giving improved speed and cooler operation). I suggest beginning with a vcore below 1.3v, manually adjusting the SOC voltage (stock may raise it beyond safe limits—uncertain if this affects memory only), then tweak the CPU frequency. Save your settings and perform stress tests; p95 often exceeds 80°C, yet I continue using it. Unlike AMD stock boosting, manual OC doesn’t cause clock drops. Blender works better for normal stability checks, OCCT, Realbench, Linpack. Avoid Cinebench—it’s too light to detect issues, and I can safely add another 100MHz while leaving other parameters unchanged. Cinebench will run smoothly while browsers lag and real stress tests fail instantly.