Setting a sustained thermal cap for an overclocked Ryzen 9 5950X
Setting a sustained thermal cap for an overclocked Ryzen 9 5950X
I've recently upgraded my Ryzen Zen 2 3700x for a Ryzen 9 (Zen3) 5950X and as I have a Corsair H150 AIO and an Asus tuf gaming x570-plus MB I wanted to take advantage of some overclocking.
First let me say I'm using a 32GB 3200MHZ cl14 RAM, I'd love to upgrade to 64GB with 2x32GB, but any sticks or RAM I can find at 32GB per stick and 3600MHz are cl18+
😵
I briefly considered buying another pair of exact same modules I have now t up my RAM to 64GB, but they (G.skill cl14 3200MHZ) cost more than what I paid when I bought them 3+ years ago...
So I'm staying with 3200MHz RAM. I upped my PBO2 to 260/tdc 160/edc 180, I set my thermal limit at 85C and I created a per core under volt that makes my cores boost to 5050MHz when one is used and 4600 when all are used. The CPU draws 230W of power and stays at around 85 package temp (~82C core temps).
And I wonder. These CPUs have factory set 95C as throttle limit. Would I be loosing lots of potential longevity by setting it at 90C or even the original 95C? That should give ma a good 100MHz of boost at least hopefully. Going from 4300 to 4600 upped my cinebench R23 score from ~27k to ~29.5k. I'd love to get to 30k (my previous CPU was a bit under 12k).
So is anyone using heavy multicore loads ona ryzen 5000 series with high EDC and power almost double the spec for a long time? There is a lot of conflicting info online, but if AMD set the default thermal throttle limit at 95 it should be safe for the CPU?
What do you think?
If your cooling is functioning correctly, you shouldn't exceed 90°C. The CPU should reach 4.7 to 4.8 GHz under all core loads and 4.9 to 5GHz on a single core. Avoid manually forcing 4.6 or 4.6 as it reduces performance. Just manage the voltage via CO.
Because a semiconductor's lifespan is largely influenced by its operating temperature, it makes sense to reduce useful life when working at higher temperatures, provided other factors remain constant. Still, keep in mind that AMD calculated that the processor would retain a reasonable service life even under its rated maximum temperature of 90°C. However, consider that this assumes typical user behavior—running standard desktop tasks with sufficient cooling—and only reaching 90°C during occasional peak loads, not continuously. If you maintain constant high-intensity processing, such as running complex all-core workloads or rendering large video batches, the 85°C limit could extend longevity. Alternatively, a lower thermal threshold might support better core performance at higher clock speeds, though this likely benefits more tasks like gaming than general workloads.
You're asking about the meaning behind the statement "if my cooling works I shouldn't see 90C". You're referring to the situation where the cooling system is functioning well and you're concerned about reaching temperatures around 90 degrees Celsius. You also mention having successfully achieved temperatures as low as 80 degrees in single-core tasks with high power consumption, and you're curious about the limits of your setup. Additionally, you're discussing performance benchmarks like Cinebench r23 and the importance of temperature management for longevity versus speed.
Considering practical performance, such as adhering to AMD's 90C threshold or lower, it seems unlikely we'll encounter many genuine "real world" data points. These chips, built with dimensions under 10nm, are still relatively new. Nonetheless, AMD has accumulated substantial test data through the FIT (Failures In Time) parameters they created for the boost algorithm, which relies heavily on core temperature measurements. While their data remains confidential, it appears confident enough to release hundreds of millions into the global market with these specifications.
I also remain skeptical about internet overclocking claims. Many users prioritize immediate functionality—booting Windows and games for a few hours—over long-term reliability. Modern processors aren't as stressed by intensive tasks like heavy video encoding (h.264 with AVX instructions) as commonly assumed, especially since such operations typically use only one or two threads. Running a full h.264 encoding on all cores simultaneously could cause crashes or produce corrupted output. Some argue that a 16-core processor isn't fully optimized for h.264. Perhaps launching two instances would render videos faster than doing them one after another, which is what I usually do.
I question whether the motherboard's default 95C thermal limit stems from only certain models—like the 5800X/X3d, 5900X, and 5950X on AM4—while others like the Ryzen 5000 and 3000 series processors exceed it. Regardless, it's a clear oversight on the manufacturer's side.
Certainly, they also offer a three-year warranty. It would really tempt me to grind that IHS (IHS flatter) (I have the capability to achieve around half a micron across the size of a Ryzen IHS, and to measure flatness down to 100nm - thanks to a different hobby) or replace it, but I’m not very sure about its reliability over these three years. I also doubt I’d be able to find another one in a year or two going forward.
I don’t think grinding or delidding would make the chances of failure worse. As you mentioned before, AMD doesn’t seem to push users to run CPUs at full capacity 24/7 at maximum temperatures, and if they do, they might just accept a higher failure rate for those users.
I wish there was long-term failure data with thermal information, similar to the kind Google provides for spinning drives they released some years ago. If I had access to typical failure rates at stock settings during heavy continuous use, near the maximum temperature (tjmax), I’d feel confident enough to grind or delide.
Someone might wonder why? Isn’t this CPU considered "not fast enough"? Well, yes it is "fast enough" without overclocking, but I still enjoy getting as much performance out of the hardware I own within my personal limits.
Exactly. Although I’ve heard on YouTube that some modern games are starting to use AVX2 and multiple cores when available, I’m not sure how widespread that is.
I'm uncertain about the benefits of refining the IHS, but if your temperature reaches the 80s, any adjustments that improve cooling—while keeping everything else constant—should boost performance due to the way the boost algorithm limits maximum voltage increases. Although it could potentially work better, I wouldn't necessarily aim to lower the temperature excessively (again because of the algorithm's behavior), but it might.
Investigate undervolting using PBO2 and Curve Optimizer. CO supports the boost algorithm, allowing improved performance from a cooler processor while also extending its lifespan through reduced voltage. Avoid merely reducing Vcore, as this disrupts the boost process and can hurt performance.
Most current games are significantly more demanding in terms of threading than before. However, any simulation remains constrained by one primary thread; all others wait to contribute. This doesn't mean they're ineffective—offloading tasks from the main thread can still enhance its efficiency. Some turn-based strategy titles also rely heavily on multiple threads, likely for simultaneously calculating outcomes after moves.
While games may utilize AVX and AVX2 instructions, this is inconsistent and not as intensive as applications like rendering software (Cinebench, Handbrake/h.264) that fully engage all available threads with dense AVX code. The most noticeable thread bottleneck in the games I play occurs during decompression of assets before they're loaded into GPU memory. This decompression process relies on AVX instructions.