CPU hitting 65C, GPU staying cold under mining load

[1stMidnight]

I mine on my GPU and CPU sometimes, and my cpu temps are stupid for a custom loop. I’m running 2 360mm radiators, and my gpu stays cool just fine.

 

For reference, in the pictures, gpu is overclocked to -150 core clock, +1000 memory clock, and cpu is stock mining on only 8 threads.

 

All fans and pump are on 100%. 
 

[Chris Pratt]

No info on CPU, GPU, or even what your cooling solution is.

[1stMidnight]

6 minutes ago, Chris Pratt said:

No info on CPU, GPU, or even what your cooling solution is.

View the pictures? And I said 2 360mm radiators. The side is an EK SE coolstream, so pretty bad and the top is some bitspower touchaqua that came with the case. Fans are HD120s and Bitspower notos. I know I could get something better, but it should be a problem for a 5600x and 1080.

[Chris Pratt]

Zen 3 runs hot, and 65C under load, even water cooled, is pretty normal. You can try undervolting it.

[1stMidnight]

1 minute ago, Chris Pratt said:

Zen 3 runs hot, and 65C under load, even water cooled, is pretty normal. You can try undervolting it.

Oh, really? What software would you recommend, and how. I still would think 65c is high, as my gpu is running overclocked 25c less under load. Could I have put too much thermal paste?

[Chris Pratt]

2 minutes ago, 1stMidnight said:

Oh, really? What software would you recommend, and how. I still would think 65c is high, as my gpu is running overclocked 25c less under load. Could I have put too much thermal paste?

No, too much paste wouldn't cause that, and that's literally normal, as strange as it may seem. Everyone water-cooling Zen 3 is pretty much seeing the same highs.

 

As far as undervolting goes, AMD added the curve optimizer for Zen 3. It's going to be slightly different per board manufacturer, but you just go into your BIOS and whatever advanced CPU overclocking section it has. From there, there should be a listing for Precision Boost Overdrive. Set that to advanced/manual. That should give some additional options. Set power limits to disabled, offset direction to negative (undervolt) and enter a number of "steps" up to 30. Each step represents 3-5mV for a maximum of 150mV. Most Zen 3 chips can do the full 30. Save and restart. Run some stress tests on the CPU just to make sure it's stable. If you do have any instability, back off on the number of steps, but like I said: that shouldn't be an issue.

[1stMidnight]

5 minutes ago, Chris Pratt said:

No, too much paste wouldn't cause that, and that's literally normal, as strange as it may seem. Everyone water-cooling Zen 3 is pretty much seeing the same highs.

 

As far as undervolting goes, AMD added the curve optimizer for Zen 3. It's going to be slightly different per board manufacturer, but you just go into your BIOS and whatever advanced CPU overclocking section it has. From there, there should be a listing for Precision Boost Overdrive. Set that to advanced/manual. That should give some additional options. Set power limits to disabled, offset direction to negative (undervolt) and enter a number of "steps" up to 30. Each step represents 3-5mV for a maximum of 150mV. Most Zen 3 chips can do the full 30. Save and restart. Run some stress tests on the CPU just to make sure it's stable. If you do have any instability, back off on the number of steps, but like I said: that shouldn't be an issue.

Alright, thanks for all the help! I really appreciate it.

[TVwazhere]

19 minutes ago, 1stMidnight said:

Could I have put too much thermal paste?

Besides conductive thermal paste or messy applications, there's no such thing as too much thermal paste. Not enough could be an issue though, or potentially improper screw tension (meaning it's either not tightened down enough or evenly). 

 

65ºC does seem a little warm, but it's definitely nothing that says "there's a problem" to me. What CPU block are you using?

[1stMidnight]

1 minute ago, TVwazhere said:

Besides conductive thermal paste or messy applications, there's no such thing as too much thermal paste. Not enough could be an issue though, or potentially improper screw tension (meaning it's either not tightened down enough or evenly). 

 

65ºC does seem a little warm, but it's definitely nothing that says "there's a problem" to me. What CPU block are you using?

A bitspower summit ms that came with my case. If I could find another block I could swap it to without having to redo the tubing, I would. I really want cpu temps as low as possible as I’m gonna be mining about 18 hours a day or so

[1stMidnight]

I’m also a first time builder so there is a great chance it’s uneven, under tightened (I was scared to tighten it too much) or not enough paste as I was trying not to emulate the verge.

[TVwazhere]

4 minutes ago, 1stMidnight said:

A bitspower summit ms that came with my case. If I could find another block I could swap it to without having to redo the tubing, I would. I really want cpu temps as low as possible as I’m gonna be mining about 18 hours a day or so

2 minutes ago, 1stMidnight said:

I’m also a first time builder so there is a great chance it’s uneven, under tightened (I was scared to tighten it too much) or not enough paste as I was trying not to emulate the verge.

I will say not tightening them enough can cause the block to not make full contact with the Chip Die (CPU or GPU) which can cause higher than expected temps. 

 

Looking at reviews there dont seem to be too many complaints about that block underperforming, so if you did have to remove that block, I'd double check the thermal paste has enough (usually a pea sized amount in the center is enough) and re tighten the mounts, hand tight at first and then tighten down every half turn or so in a cross pattern until it gets pretty difficult. If you remove the block and there's not a lot of paste on the die and block as if most of it got squished out, then that means you had good contact. If you see a lot of gray paste thats still left on the block and die, likely you didnt have enough contact. 

Photo source and article

[1stMidnight]

3 minutes ago, TVwazhere said:

I will say not tightening them enough can cause the block to not make full contact with the Chip Die (CPU or GPU) which can cause higher than expected temps. 

 

Looking at reviews there dont seem to be too many complaints about that block underperforming, so if you did have to remove that block, I'd double check the thermal paste has enough (usually a pea sized amount in the center is enough) and re tighten the mounts, hand tight at first and then tighten down every half turn or so in a cross pattern until it gets pretty difficult. If you remove the block and there's not a lot of paste on the die and block as if most of it got squished out, then that means you had good contact. If you see a lot of gray paste thats still left on the block and die, likely you didnt have enough contact. 

Photo source and article

I let it mine for like 20-30 mins and it reached 67c and climbing. I’ll try that later. The paste I used is a small thing of ek-ectotherm that the person I bought my mobo from gave to me, so I’m going to buy some Arctic mx-5 and retry. I can clean off the old thermal paste with isopropyl alcohol on a cotton ball, right?

[1stMidnight]

2 minutes ago, 1stMidnight said:

I let it mine for like 20-30 mins and it reached 67c and climbing. I’ll try that later. The paste I used is a small thing of ek-ectotherm that the person I bought my mobo from gave to me, so I’m going to buy some Ic diamond 7 Karot and retry. I can clean off the old thermal paste with isopropyl alcohol on a cotton ball, right?