10700kf Help!

[Mr Mcgoo]

So i have my 10700kf at stock undervolted. However, it would seem i can get 1.15v but my temps are hitting between 60-65c during multicore cinebench r23. My water temp at the end of the test is hitting 26c. i have seen others hitting lower than this at higher vcore. I even reseated my cooler and changed thermal paste twice. At 5ghz (1.25v) all core i see between 72-77c. Why are my temps higher than other with a custom loop? Also I have the MSI z590 gaming edge wifi and can see socket sense and vcc sense for voltage reading. Which is more accurate?

 

[Mr Mcgoo]

bump

 

[Wing-Man]

why is a peace of string this long, and not that long...


room temp, flow rate, fan speed, rad size/quality.. and the number one reason people lie so there is better..
77c under stress testing is good

 

[decto]

Would agree with above, likely pushing 200W at stock and more with the 5Ghz all core.
AIO's are fairly decent these days so unless you heat soak an AIO, a custom loop doesn't have that much advantage on a CPU only stress test.

Main issue is getting the heat from the CPU to the block to the water, looks like you have a fairly generic block that's not optimised for the socket but plenty of rads to keep water temp low.
If others are reporting better temps, take a look at what block they are using.

You're temps are good though, considering the power throughput so personally I wouldn't change anything.

What are you peak CPU temps after a gaming session?

 

[Mr Mcgoo]

why is a peace of string this long, and not that long...


room temp, flow rate, fan speed, rad size/quality.. and the number one reason people lie so there is better..
77c under stress testing is good

My room temp is 21c, flow rate 200lph, fan speed 1000rpm under 30c water temp and 1250rpm anything above that, 2x240mm xr5 rads and 1x360 xr5 rad. Water temp after heavy gaming session hits 29.7c

Would agree with above, likely pushing 200W at stock and more with the 5Ghz all core.
AIO's are fairly decent these days so unless you heat soak an AIO, a custom loop doesn't have that much advantage on a CPU only stress test.

Main issue is getting the heat from the CPU to the block to the water, looks like you have a fairly generic block that's not optimised for the socket but plenty of rads to keep water temp low.
If others are reporting better temps, take a look at what block they are using.

You're temps are good though, considering the power throughput so personally I wouldn't change anything.

What are you peak CPU temps after a gaming session?

At stock 4.7ghz 1.15v i pull 135watts in cinebench multicore and at 5ghz 1.25v i pull 175watts in cinebench. After a gaming session they peaked at 58c at 75% load at 4.7ghz and peaked at 66c at 71% load at 5ghz. while i know the temps are well within tolerance. Just seeing a fair few people with 65-70c at 5ghz pulling 190-200watts at 1.3v.

 

[Wing-Man]

After a gaming session they peaked at 58c at 75% load at 4.7ghz and peaked at 66c at 71% load at 5ghz.

that genuinely is good.

peaked at 66c at 71% load at 5ghz

Just seeing a fair few people with 65-70c at 5ghz

so your bang on the money

as asked by @decto its probably down to the block your using, a lot of water-cooling manufactures are pumping out the same blocks they did 10 years ago.
if you cant get the heat from the CPU to the water nothing else matters.

But again that said, your number look good

 

[decto]

My room temp is 21c, flow rate 200lph, fan speed 1000rpm under 30c water temp and 1250rpm anything above that, 2x240mm xr5 rads and 1x360 xr5 rad. Water temp after heavy gaming session hits 29.7c

At stock 4.7ghz 1.15v i pull 135watts in cinebench multicore and at 5ghz 1.25v i pull 175watts in cinebench. After a gaming session they peaked at 58c at 75% load at 4.7ghz and peaked at 66c at 71% load at 5ghz. while i know the temps are well within tolerance. Just seeing a fair few people with 65-70c at 5ghz pulling 190-200watts at 1.3v.

With this type of block it is important to get the in/out correct and even the orientation can make a difference.

The water is pushed through the slot to impinge in the centre of the fins, it then flows to the sides and exits.
The slot in the block should be alighed with the CPU die, water is a poor conductor of heat but a good carrier so it's the impingement that maximises the heat transfer.
From what I can see a line drawn from the in to the out port should be parrellel to the socket lever... but that's just from looking at a few pictures and a delid so you may want to confirm.

Seems I can't add a video at specific point so image here:

Spoiler

https://i.imgur.com/avg6qvp.png

Video - image at 11:23

Spoiler

 

[Wing-Man]

one more thing to take into account is the motherboard and its sensors. you are assuming your board the reporting the right Watts, Voltage and temp.
Some boards can be way out and if you have an High End board doesn't mean its any better.

putting the same watts into 2 CPU's can give 2 different temps based on the CPU its self but also the way the VRM deliver the power

 

[Freddie1980]

My room temp is 21c, flow rate 200lph, fan speed 1000rpm under 30c water temp and 1250rpm anything above that, 2x240mm xr5 rads and 1x360 xr5 rad. Water temp after heavy gaming session hits 29.7c

At stock 4.7ghz 1.15v i pull 135watts in cinebench multicore and at 5ghz 1.25v i pull 175watts in cinebench. After a gaming session they peaked at 58c at 75% load at 4.7ghz and peaked at 66c at 71% load at 5ghz. while i know the temps are well within tolerance. Just seeing a fair few people with 65-70c at 5ghz pulling 190-200watts at 1.3v.

Your numbers look fine, those chips aren't the easiest to cool due to the density of the silicon but what I would say is your running a lot of radiators in your setup (2x 240 and 1x360). I'm wondering if your putting to much load on your pump to get the full benefit, for single CPU a single thick 240 rad should be more then sufficient even with that power draw.

 

[Deleted member 258511]

Some people are reporting much higher temps. You should be pleased.

 

[Mr Mcgoo]

With this type of block it is important to get the in/out correct and even the orientation can make a difference.

The water is pushed through the slot to impinge in the centre of the fins, it then flows to the sides and exits.
The slot in the block should be alighed with the CPU die, water is a poor conductor of heat but a good carrier so it's the impingement that maximises the heat transfer.
From what I can see a line drawn from the in to the out port should be parrellel to the socket lever... but that's just from looking at a few pictures and a delid so you may want to confirm.

Seems I can't add a video at specific point so image here:

Spoiler

https://i.imgur.com/avg6qvp.png

Video - image at 11:23

Spoiler

Orientation is correct including inlet and outlet. the inlet and outlet are marked inside the block thankfully.

one more thing to take into account is the motherboard and its sensors. you are assuming your board the reporting the right Watts, Voltage and temp.
Some boards can be way out and if you have an High End board doesn't mean its any better.

putting the same watts into 2 CPU's can give 2 different temps based on the CPU its self but also the way the VRM deliver the power

Yes my board is reporting correct power use for my cpu. I am using a wall plug wattage meter and can see the usage at the wall matches my efficiency of my psu under load.

Your numbers look fine, those chips aren't the easiest to cool due to the density of the silicon but what I would say is your running a lot of radiators in your setup (2x 240 and 1x360). I'm wondering if your putting to much load on your pump to get the full benefit, for single CPU a single thick 240 rad should be more then sufficient even with that power draw.

I know my rad setup is overkill for my current system. But i got it ready for an rtx 4070/80 once prices drop a little and i save a little more, which is hard due to uk inflation right now.