Question Temperature to expect with aio and 11700k

Phaaze88

Titan
Ambassador
Mission impossible - that's for you to find out and tell us.
-room ambient
-chassis ambient
-chassis airflow
-fan setup, model(s), and curve
-radiator position
-pump speed: should be maxed anyway
-bios settings: Intel defined, vendor custom, or user custom
-Windows power plan
-Gpu power consumption and cooler
-AVX offset, if in use
^All that, and probably others, is going to influence thermals.

Few would be able to answer this without the exact same setup, Windows/bios settings, and home cooling as you.

IF you installed the Neptune in the roof, that likely wasn't the best place for it.

How's the 980 Pro? I've seen a few threads where the drive is thermal throttling. The thermals of Gen 4 NVMe appears to be the unspoken elephant in the room.
 
Dec 4, 2020
45
0
40
1
Mission impossible - that's for you to find out and tell us.
-room ambient
-chassis ambient
-chassis airflow
-fan setup, model(s), and curve
-radiator position
-pump speed: should be maxed anyway
-bios settings: Intel defined, vendor custom, or user custom
-Windows power plan
-Gpu power consumption and cooler
-AVX offset, if in use
^All that, and probably others, is going to influence thermals.

Few would be able to answer this without the exact same setup, Windows/bios settings, and home cooling as you.

IF you installed the Neptune in the roof, that likely wasn't the best place for it.

How's the 980 Pro? I've seen a few threads where the drive is thermal throttling. The thermals of Gen 4 NVMe appears to be the unspoken elephant in the room.
My build is as follows:
  • motherboard - ASUS ROG STRIX Z590-E GAMING WIFI
  • cpu - Intel tray cpu core i7-11700K 3.6GHz 16MB
  • hard drive - os - SAMSUNG SSD 2TB 980 PRO M.2 NVMe MZ- V8P2T0BW
  • hard drive - data - SAMSUNG 2.5" SSD 4TB EVO870 MZ-77E4T0B
  • memory - G.Skill Trident Z RGB 2x16GB DDR4 4000Mhz CL17 Kit
  • graphics card - MSI GeForce RTX™ 3080 GAMING Z TRIO 10G
  • cpu cooler - antec neptune 240 argb liquid cooler
  • case - PHANTEKS ECLIPSE P600S PH-EC600PSTG_AG01
  • PS - Corsair 750W 80+ Gold RM750x
Room temperature: 24 - 27 C

Fan Setup:
  • Two 14” intake fans at the front (built in by phanteks)
  • One 14” exhaust fan at the back (built in by phanteks)
  • AIO radiator with two 12” fans mounted as exhaust at the top (AIO kit by antek)
  • Pump speed: full (no way to control it)
  • GPU cooler: MSI stock
bios settings: intel defined, no OC
Windows power plan: balanced

Please note that the case has two detachable noise reduction heavy plates - one on the top and on at front. Removing them improves thermals by up to 3-5 C at idle.

Regarding the 980 - it is very fast, delivering and even exceeding Samsung’s promises, but it does run (relatively) hot. I have a thread dedicated to that here:

I’m adding a snapshot of my build to give a better sense of the setup:
(M2 installed under the yellow rectangle)
 

Phaaze88

Titan
Ambassador
Regarding the 980 - it is very fast, delivering and even exceeding Samsung’s promises, but it does run (relatively) hot. I have a thread dedicated to that here:
Ok... so that was just idle?

I’m adding a snapshot of my build to give a better sense of the setup:
Now, I don't know how you run your PC, or how you will run it later, so I must look at this from a worst case scenario:
It is entirely possible for you to overwhelm the Neptune 240 in its current position.

When users do top mounted hybrids, it doesn't just cope with the cpu it's mounted to, but also the gpu, usually the biggest source of heat.
So the ideal thing to do with top mounts is to go with a bigger model to deal with both... you didn't really do that here.

The 11700K is 'another 9900K'.
When the 9900K came out, it was shown to kick the butts of 240mm hybrids in cpu only testing, be it Prime95, Blender, or some AVX application rears its ugly head.
Over 80C in those tests easily. And that was without overclocking.

The power efficiency has gone down with these RTX 30 cards compared to the last 2 gens. That's a ~400w card in there. Some of that heat will be passing through the top mounted Neptune.

Front mounting the Neptune should make things easier on it.


Also, the 3080 can heat up the 980 Pro positioned above it.
 
Dec 4, 2020
45
0
40
1
so i took some measurements today.

ambient: 25.5C

with the top and front panels off:
  • idle: ~31C (delta 5.5C)
  • stress multi core: 63C (delta 37.5C)
  • stress single core: 55C (delta 29.5C)
with the top and front panels on:
  • idle: ~34C (delta 8.5C)
  • stress multi core: 65C (delta 39.5C)
  • stress single core: 58C (delta 32.5C)
is that decent?

@Phaaze88 i use idle temp as reverence as that is closer to standardized
 

Phaaze88

Titan
Ambassador
Why did you take the panels off? Test it as is - how you will use it.

There's no mention what you tested with...
Prime 95: Small FFT, AVX-512 off, AVX2 off, AVX off, run for 15 mins.
OCCT: Cpu test, Small data set, Normal mode, Steady load type, SSE Instruction set, run for 15 mins.
^Pick one.

Furmark or Msi Kombustor: Select the monitor's native resolution and run stress test for 15 mins.
Unigine Superposition: Game mode - not the Benchmark - Custom Preset, Direct X API, Resolution(monitor native), Shader and Texture quality on high. Run for 15 mins.
^Pick one.

Run the selected applications at the same time. That'll be your worst case scenario(possible).
If cpu and gpu core temperatures exceed 85C, consider it a confirmation that the cooling isn't sufficient.

EDIT: Also sustained Vram thermals over 95C isn't great either. Hwinfo can monitor that one.
 
Last edited:
so i took some measurements today.

ambient: 25.5C

with the top and front panels off:
  • idle: ~31C (delta 5.5C)
  • stress multi core: 63C (delta 37.5C)
  • stress single core: 55C (delta 29.5C)
with the top and front panels on:
  • idle: ~34C (delta 8.5C)
  • stress multi core: 65C (delta 39.5C)
  • stress single core: 58C (delta 32.5C)
is that decent?

@Phaaze88 i use idle temp as reverence as that is closer to standardized
Your temps are decent, and actually very good.
I usually think of 10-15c. over ambient at idle.

Your case has good intake, and that is necessary for good cooling.

Temperatures on 11th gen do not greatly increase unless you are overclocking which requires extra voltage to be applied.
It is exponential so trying for the max overclock is what gets you in trouble.
 
Reactions: roieco

Phaaze88

Titan
Ambassador
@roieco
Cpu-Z stress test doesn't apply that hard of a load; we are in a time where there are cpu heavy titles harder than that, as well as ones that use AVX. Both can catch you off-guard.
CoD: Cold War, Horizon Zero Dawn, Cyberpunk 2077, and Battlefield V are a few examples.

OH, I completely misunderstood what you meant by panels off. You were referring to the mesh panel covers, not the entire panels. Sorry!
 

Windows22

Great
Dec 30, 2020
118
12
85
0
Mission impossible - that's for you to find out and tell us.
-room ambient
-chassis ambient
-chassis airflow
-fan setup, model(s), and curve
-radiator position
-pump speed: should be maxed anyway
-bios settings: Intel defined, vendor custom, or user custom
-Windows power plan
-Gpu power consumption and cooler
-AVX offset, if in use
^All that, and probably others, is going to influence thermals.

Few would be able to answer this without the exact same setup, Windows/bios settings, and home cooling as you.

IF you installed the Neptune in the roof, that likely wasn't the best place for it.

How's the 980 Pro? I've seen a few threads where the drive is thermal throttling. The thermals of Gen 4 NVMe appears to be the unspoken elephant in the room.
Phaaze88's right, these factors have to be taken into account in order to have a general idea.
 
Dec 4, 2020
45
0
40
1
following advice here, I've run Unigine_Heaven-4.0
profile:
  • api : dx11
  • quality: ultra
  • tessellation: extreme
  • sterep 3d: disabled
  • multi monitor: disabled
  • anti aliasing: X8
  • full screen
  • resolution: 4k
ambient: 25.5C

for CPU, most of the time the temperature was about 42C (delta 16.5C), with maximal temperature reaching 50C (delta 24.5C)
for GPU, it reached and hovered just under 80C (delta 54.5C), from idle temperature of ~47C (delta 21.5C)

i recount here the idle and cpuz stress temperatures i tested before (ambient was the same for both tests):

with the top and front panels off:
  • idle: ~31C (delta 5.5C)
  • cpuz stress multi core: 63C (delta 37.5C)
  • cpuz stress single core: 55C (delta 29.5C)
with the top and front panels on:

  • idle: ~34C (delta 8.5C)
  • cpuz stress multi core: 65C (delta 39.5C)
  • cpuz stress single core: 58C (delta 32.5C)
 
Last edited:
You are worrying for nothing.
Your temperatures are fine.
A processor will monitor it's own temperature and will slow down or shut off if it detects a dangerous temperature.
That is around 100c.

Samsung operating temperatures are 0 to 70c.
You are likely ok there.

If you run something like HWmonitor you will get current, minimum, and maximum temperatures for cpu and ssd.
If the cpu max ever gets to 100c. you have throttled.
If the ssd heats up under continuous reading(say 30 seconds) it will slow down for a bit.
Heavy random i/o(which is most of what we do) is not likely to heat up the ssd.

It is not bad for a processor to run hot.
It is excessive voltage from overclocking that can damage a processor.
Under full load, 75c. is OK
Under a stress test, 85c. might be the most you would want.
 
Reactions: roieco
Dec 4, 2020
45
0
40
1
Thank you @geofelt . Hwinfo is what I've been using for these tests.
Happy to hear that the tpepmperature are as would be expected. I'll redo these tests from time to time.
 

Phaaze88

Titan
Ambassador
@roieco
Credit goes to @CompuTronix for the guide.


Note 1: As per Intel’s Datasheets, TDP and Thermal Specifications are validated “without AVX. Prime95 with AVX test selections enabled will impose an unrealistic 130% workload which can increase Core temperatures by up to 20°C. To correctly set Prime95 to run your CPU at 100% workload, simply click on the AVX test selections that are not greyed out so that all three AVX boxes are checked, as shown above. Further explanation concerning AVX Instruction Sets are detailed later in this Section.

Utilities that don't overload or underload your processor will give you a valid thermal baseline. Here’s a comparison of utilities grouped as thermal and stability tests according to % of TDP, averaged across six processor Generations at stock Intel settings rounded to the nearest 5%:




Figure 11-2

Although these tests range from 70% to 130% TDP workload, Windows Task Manager interprets every test as 100% CPU Utilization, which is processor resource activity, not %TDP workload. Core temperatures respond directly to Power consumption (Watts), which is driven by workload. Prime95 Small FFTs (AVX disabled) provides the correct workload for testing thermal performance. If Core temperatures don't exceed 80°C, your CPU should run the most demanding real-world workloads without overheating.

Cpu-Z is light, and is no good for getting a thermal baseline.
Again, I don't know everything you plan to use/operate your PC for, but some titles already out there would already push the cpu harder than Cpu-Z does.

Heaven Benchmark doesn't even support DX12.
 
Reactions: roieco
Dec 4, 2020
45
0
40
1
@roieco
Credit goes to @CompuTronix for the guide.


Note 1: As per Intel’s Datasheets, TDP and Thermal Specifications are validated “without AVX. Prime95 with AVX test selections enabled will impose an unrealistic 130% workload which can increase Core temperatures by up to 20°C. To correctly set Prime95 to run your CPU at 100% workload, simply click on the AVX test selections that are not greyed out so that all three AVX boxes are checked, as shown above. Further explanation concerning AVX Instruction Sets are detailed later in this Section.

Utilities that don't overload or underload your processor will give you a valid thermal baseline. Here’s a comparison of utilities grouped as thermal and stability tests according to % of TDP, averaged across six processor Generations at stock Intel settings rounded to the nearest 5%:




Figure 11-2

Although these tests range from 70% to 130% TDP workload, Windows Task Manager interprets every test as 100% CPU Utilization, which is processor resource activity, not %TDP workload. Core temperatures respond directly to Power consumption (Watts), which is driven by workload. Prime95 Small FFTs (AVX disabled) provides the correct workload for testing thermal performance. If Core temperatures don't exceed 80°C, your CPU should run the most demanding real-world workloads without overheating.

Cpu-Z is light, and is no good for getting a thermal baseline.
Again, I don't know everything you plan to use/operate your PC for, but some titles already out there would already push the cpu harder than Cpu-Z does.

Heaven Benchmark doesn't even support DX12.
thank you @Phaaze88.
i ran the tests you suggested
  • Prime95 Small FFTs (AVX enabled) - sat on ~80C for a while, then dropped to 60C and stayed there
  • Prime95 Small FFTs (AVX enabled) - sat on ~70C for a while, then dropped to 50C and stayed there
 
Dec 4, 2020
45
0
40
1
I hope that's a typo. It should be disabled.
Yes, the correct results are

  • Prime95 Small FFTs (AVX enabled) - sat on ~80C for a while, then dropped to 60C and stayed there
  • Prime95 Small FFTs (AVX disabled) - sat on ~70C for a while, then dropped to 50C and stayed there
 

ASK THE COMMUNITY

TRENDING THREADS