10C contrast in cores!
10C contrast in cores!
Fascinating, I struggled to understand why my CPU throttles at 95°C instead of the 100°C Intel claims. No PROCHOT adjustment appears anywhere (TS 9.5 is what I use). 95°C feels reasonable—I don’t want to boost PROCHOT since 95°C seems adequate. All four cores running around 3.1GHz at 95°C (possibly 3.2GHz with the extra 5°C?) makes sense. I’m thinking about buying a MX4 soon or next week and swapping it out.
Have you checked where I mentioned? A gold-colored lock icon indicates the PROCHOT Offset register is locked and won’t change. I haven’t noticed any Intel Core i CPUs from the past 15 years functioning this way. Intel processors keep trying to boost as much as they can during a thermal throttle event. They only reduce speed enough to stay under the safe temperature limit. Most tracking software fails to measure CPU speed accurately when throttling occurs. ThrottleStop is one of the few programs that follows Intel’s suggested approach. It keeps showing correct MHz even during thermal or power limits. The first image in this discussion confirms this. Some laptop users have experienced bad long-term outcomes with MX-4. It can begin to overheat within a week or two, causing temperatures to rise. Do some online research. I’ve heard Noctua NT-H2 performs better under extended use. The issue with most thermal paste reviews is they test on desktop CPUs with built-in heat spreaders. Another issue is that they evaluate 10 pastes for short periods then switch to the next one. These tests don’t reflect real-world performance when applying paste directly to a mobile CPU or after weeks of replacement. This explains why Noctua developed a new formula to handle the high temperatures many laptops face today.
I'm sorry, I found it... I'm really confused... Apologies for my oversight. I attempted a benchmark and the extra 5c headroom seems to maintain performance at 3.1 GHz (instead of dropping to 2.9 then returning to 3.1). However, I've noticed another problem that hasn't appeared in about two years—power throttling. It appears TS 9.5 doesn't have a toggle to bypass power limits, unlike TS 8.7. I fixed it by deleting the current version and running the older one, which worked perfectly. I read positive feedback for the MX-4, so I decided to try it out. I'll do some proper research and keep the NT-H2 in mind.
The document mentions several points about checks and settings, but it seems some details are unclear. It references a ReadMe file in the ThrottleStop folder, which might be missing. The discussion focuses on avoiding outdated TS versions and addressing thermal management issues with PROCHOT Offset. It also touches on CPU shutdown temperatures and the need for consistent monitoring over time.