Optimal method for lowering the CPU "Socket" temperature on ASUS M5A99X EVO R2.0
Optimal method for lowering the CPU "Socket" temperature on ASUS M5A99X EVO R2.0
Hi,
I’m looking for advice on how to cool the socket or if there’s a better way to locate the sensor in the MB. I’m currently using air cooling and planning to make a water-based build, which should improve performance significantly. However, I’m not too excited about the socket itself, even with a high-quality water build. I’ve already installed a 7000 RPM fan on the CPU’s back, but it seems ineffective—temperatures stay the same whether or not the fan is running. If you could point me to the exact location of the sensor, that would really help. Alternatively, do you have suggestions for a more effective cooling method? Also, I’m considering an extreme OC for my CPU once I finish the water build, but if the socket won’t cool properly—even with a water build—I might run into issues. I’ve successfully boosted my FX 8350 to 5.0 GHz on air cooling (Evo 212), and it runs well under load, though stress tests cause overheating and the system enters safe mode. Please advise.
Thank you!
All I need is some guidance on socket cooling.
Don’t worry about air cooling—I’m focusing on the water build now.
The fan in the rear of the socket won't aid the CPU, but it might help cool the mainboard and the back of the VRM, which could become warmer. The VRM chips are placed directly on the mainboard, allowing them to heat up more than the CPU, which is protected by the socket.
you're sure about that ?
well i can't test it right away because i'm stil waiting for my water tanks & pomps to arive but i hope you are right ...
That could fix the problem but do i need to place an extra sensor on the back of the MB at the socket ?
Thanks.
Yes, I'm certain about FX processors and Athlon II and Phenom II processors. Previously, Athlon featured a flexible probe in the middle of the socket and connected to MB, so they likely kept using the "Socket" temperature name.
Moreover, AMD chips lack separate temperature sensors for each core, meaning what is shown for one core is identical across all cores. The "socket" temperature reported by some programs actually reflects package-wide readings.
The fan in the rear of the socket won't aid the CPU, but it might help cool the mainboard and the back of the VRM, which could become warmer. The VRM chips are placed directly on the mainboard, allowing them to heat up more than the CPU, which is protected by the socket.
I'm thinking about installing copper heat pipes on the VRM to improve cooling performance.