Need guidance on 7800x3d? Here’s some advice.
Need guidance on 7800x3d? Here’s some advice.
You're navigating some tricky thermal dynamics here. Upgrading to a 7800x3d with a Noctua DH15 cooling solution sounds solid. The negative offset of -20 on all cores seems like a reasonable approach, similar to your previous setup. However, the shader compilation reaching 85°C is concerning—this suggests your system is pushing close to its limits during rendering tasks. The Prime 95 stress test showing 100% usage without exceeding 70°C might be misleading; it could indicate unstable thermal behavior or insufficient cooling under load. The latest Asus BIOS updates can indeed complicate things, making settings feel inconsistent. Consider checking the current BIOS version, ensuring your fans are running at optimal speeds, and verifying that your CPU cooler is properly seated. If temperatures remain high during intensive tasks, it might be worth consulting the manufacturer’s support or exploring more robust cooling options.
That's actually quite typical. The AMD boost algorithm behaves oddly on X3D chips because of strict constraints on voltage, power, and clock speeds. Prime95 tends to perform better in this scenario compared to Cinebench. Running at 85C isn't risky, and it's one of the few tasks that pushes the chip to its limits, which is expected.
The chip is behaving as anticipated, but it dropped quickly after the shader compilation finished. Check the BIOS settings to see if any adjustments are needed for this model.
Welcome to contemporary processors where 90°C+ is quite typical. 85°C is now seen as quite high for maximum temperature. Beyond the standard XMP/EXPO settings, it's not much more interesting. PBO can be enabled, though it appears you likely already have it active. Generally, there aren't many things requiring direct adjustment unless you have a particular reason to modify them.
To illustrate this, consider using prime95. It forces all CPU cores to operate identically, pushing them toward maximum efficiency. In reality, you'll hit your power goal quickly, and cores adjust voltage and current to maintain optimal performance. If you design for less parallel tasks, one core handles everything, never reaching full capacity but using more power per unit. This raises heat density, making cooling more challenging. For instance, a 5900X running single-core loads at about 30W, while multi-core loads average around 104W. If every core performed the same as a single core, total draw would skyrocket to 360W—far exceeding the actual 104W. Similar issues arise with workloads targeting specific core counts. Overall, power density often matters more than raw total consumption.