High temperatures on the Ryzen 7 5800X3D are puzzling.
High temperatures on the Ryzen 7 5800X3D are puzzling.
he isn't going to... the chip throttles hard at around 81c ... but yeah, there's something off and that something is probably the whole setup instead of a too weak cooler, because the chip isnt that hard to cool at all ... doesn't really matter, ryzen is all about temps... this is my little guy, with a noctua u12s as said, and a "bad airflow case" (inwin 301) i did push it to ~125w at some point, doesn't matter performance doesn't improve as it'll just run hotter lol...
Most individuals can manage with a fundamental grasp of thermodynamics, often overlooking its complexity. This view is overly simplistic because many use appropriately sized heatsinks for their processors, allowing them to rely on a more basic interpretation: high temperatures are undesirable—either increase fan speed or airflow, or switch to a larger cooler. Conversely, low temperatures are favorable, enabling reduced fan operation or the use of smaller coolers. The challenge arises when talking about 3D v-cache CPUs, which are relatively low-power components. Their design introduces additional thermal management layers, each needing precise temperature control. This often leads to higher CCD temperatures, explaining why a delidded 5800x3D performs similarly to a standard 5800x when both have comparable heat transfer structures. The differences become clearer in this scenario, as the extra copper bonding layers in 3D v-cache affect thermal distribution. Many still mistakenly believe that higher temperatures equate to more heat, but in reality, it's the nuanced temperature gradients that matter most. My deeper explanations are limited here, since many share this common misunderstanding about how thermodynamics applies. I suspect any flaws in the die are likely due to mounting issues rather than a fundamental design problem, especially considering the performance of both 5800x and 5800x3D under similar conditions.
You're right, I didn't mention the candle implication earlier. It seems better to ask directly if the flame made the space warmer than expected. A candle burns at extremely high temperatures, producing yellow-orange light, but its heat output is minimal compared to modern lighting. For context, old daylight bulbs were very hot and intense, yet they emitted only a small amount of thermal energy. This could help clarify things without confusing them.
This shows a solid grasp of the core concepts behind nuclear reactors and their electronic applications. Using everyday comparisons can make these ideas clearer. The initial discussion seemed to address most setup challenges, but it’s possible the issue lies elsewhere—like a faulty IHS, solder joint, or component itself. Unless there are repeated problems during remounting, it probably points to a hardware defect.
I acknowledge my grasp of thermodynamics is quite limited when it comes to silicone, yet I have a solid understanding of aerodynamics—it's clear that removing heat quickly improves performance. Many people overlook this point, assuming flashy RGB fans mean effective cooling, but that’s not the case. Temperature control remains crucial because of the underlying algorithms. Interestingly, I discovered Ryzen temperatures are better managed with negative voltage offset rather than just airflow alone. Both factors must be sufficient, particularly with advanced 3D chips. I agree with you, but I think simpler fixes exist—probably just a few high-quality exhaust fans, maybe around -20 to -30°C across all cores, which should bring things down by 10 to 20 degrees.
I didn't like the 5800x3D at all; even a -10mV drop wasn't reliable. It seems to depend on the specific chip. Undervolting assumes the chip is fairly typical, because it's likely that a given 5800x3D only runs at normal voltages. Based on their production numbers, many chips might only handle -10mV or less, but this isn't guaranteed given how inconsistent results can be. If someone can't run stock boost clocks and high power loads on an all-core setup, that's a red flag for this model. A 10k on R23 is a strong indication of a problem with the 5800x3D. There are some discussions online about aggressive -20mV or more undervolting attempts and the resulting drop in frame rates. People using Cinebench R23 often see around 12k or lower, which raises concerns.
my results improved slightly with -30... but silicone lottery is real. i still believe there’s an issue with OPs cooling—it’s a fairly typical situation, and the fan configuration might be insufficient. We also saw many discussions about people earning 90c+; when they mentioned PBO2, they often got around 75c, which is enough to run the CPU without throttling. It’s not a certainty, but I’d definitely give it a shot before calling the CPU faulty.
Consider that PBO2 isn't available on the 5800x3D. Certain motherboards could have added it unofficially, but it was intentionally turned off on this model to avoid significant problems. The benefit of using the 7000x3D came from supporting PBO2.
ah, a misunderstanding i guess... i know its not officially supported - i didn't know pbo2 is an official name (not sure?) https://github.com/PrimeO7/How-to-underv.../README.md even if it's supported on the mobo or otherwise it still boils down to silicone lottery i guess. me i just set it to -30 all cores, tested it not too extensively*, and even after a year not a single crash, seems stable! : D *i always do cb, firestrike, and superposition + p95** at the same time which i think tests mostly heat and power but in my experience is also pretty sensitive to unstable "overclocks". **and i always chicken out after like 5 minutes, its just so hot and loud haha (still enough to crash if its not stable typically)