2 min read

HTPC Part 2

B34rd

It has been a while since I did part 1. Once the system was up and running my kids didn’t give me any time to run benchmarks, change settings, or really have fun with the overclocking. I’ve finally finished everything and have numbers!

I went a bit wild to start. I set the clock speed on GPU (EVGA GeForce GTX 1080Ti Founders Edition) to 2 GHz and set the power consumption and memory as fast as it would go. This paired with the 4.3 GHz (Ryzen 7 2700X) on all cores of CPU things were rocking at 4k! My kids play mostly Lego games and some emulated games so things were running stable and at great temps! As I’m sure you all guessed, when I finally had a chance to hop on Far Cry 5 things tanked, and they tanked HARD.

After the crash, I sat down and really thought about what I wanted to achieve. This system sits in an entertainment center with minimal airflow. It only has one 240mm radiator for both the CPU and GPU. I really shouldn’t be pushing it to the limit. The previous iteration at least had an additional 120mm radiator to help out and I added the 1080 Ti to replace the fried 1080 which only increased thermals. After thinking long and hard about what my goals were, I decided. I’m pushing it as far as I can without noise or heat being obnoxious. Here’s how I did…

CPU — 4.3 GHz on all cores. I could honestly go further but I think that’s a good spot to be for now. Maybe in the future, I can see about a bit more (I can probably get to 4.5 but with the entertainment center I don’t know).

GPU — Base clock is at 1780 MHz and boost clock is at 2000 MHz. Memory is clocked at 5000 MHz.

RAM — I was able to crank quite a bit out of RAM, it went from 3200 up to 4400 without any trial and error. I just set the base multiplier and it took it right away.

After all of this, gameplay max thermals are the high 60s. I’m completely amazed by the ability of the aluminum water cooking kits from EK. Mini ITX is notoriously bad for case airflow, and the Fractal Design Node 500 is a great case but with everything I have packed inside airflow is certainly not the best.

I was going to post benchmarks and do a full video as well but unfortunately, I haven’t had the time recently. If there is interest, I can do a follow up with benchmarks and further info.