Benchmarking
Last night I spent my time benchmarking my original CPU cooler, and then installing and benchmarking this ridiculous behemoth. The dimensions are 165x161x150mm, or 6.5x6.34x5.91 inches. It's really quite absurd. I need a banana for scale, really this picture doesn't do it justice. The results weren't good so I'm going to reseat the cooler after work today. On average it runs 8 degrees Celsius higher than my previous AIO, triggers thermal throttling sooner and reduces my core clock speed more, and was recorded topping out at 100c in the same test where previous cooler topped out at 82c. Because of the size and weight of the cooler, I also benchmarked it with the case on its side, so that the weight of it would only increase the contact with the heatsink and CPU. It's all because I saw some dumb video on Youtbube or something where they were showing top of the line air coolers were outperforming low end AIOs, which is what I had/have. I bought this basic ass Thermaltake Water 3.0 Performer C 120mm AIO Liquid Cooling System https://www.amazon.com/gp/product/B00MPIDYTO/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1 three years ago in 2018 for a pretty cheap $131.21 which includes tax & shipping. This Noctua NH-D15 was $110.19 by comparison. So yeah, I'm going to reseat it with less thermal paste, run a few more benchmarks, reseat the AIO and also use a lot less thermal paste, because it's possible I can get even better numbers from my AIO by minimizing the amount of thermal paste I'm using. The science behind it is that copper for example has a great thermal conductivity, so let's say that the heatsink surface area that is against the CPU is copper, and that rating is something like 390W/m-K, which is a great and high number, and you're adding thermal paste because when the two surfaces come together there will be gaps filled by pockets of air, with air being composed mostly of oxygen and nitrogen both of which have a rating of 0.02 W/m-K, using thermal paste even though it only has a rating of something like 15 W/m-k, is an extreme improvement. But if you're using so much thermal paste that there is no metal to metal contact you're really missing out on that iron 205 W/m-K to copper 390 W/m-K heat transfer.