F5F Stay Refreshed Hardware Desktop The 7980XE OC isn't performing as expected.

The 7980XE OC isn't performing as expected.

The 7980XE OC isn't performing as expected.

Pages (3): 1 2 3 Next
S
Skro74
Junior Member
37
07-16-2017, 09:21 PM
#1
Just installed the 7980XE and cleaned it up with liquid metal. It works fine at stock speeds, hitting 40°C per core in Cinebench R15. But pushing it to 4.7GHz causes temps near 90°C and sometimes higher. I reached over 90°C on six cores. Can you help me get it to 4.8GHz? I'm using three 480mm thickest rads, one 240mm thick rad, two Revo D5 fans at full speed, and 18 Noctua Industrial 3000 RPM fans running at 70%. Also, I have an X299 Dark in the room.
S
Skro74
07-16-2017, 09:21 PM #1

Just installed the 7980XE and cleaned it up with liquid metal. It works fine at stock speeds, hitting 40°C per core in Cinebench R15. But pushing it to 4.7GHz causes temps near 90°C and sometimes higher. I reached over 90°C on six cores. Can you help me get it to 4.8GHz? I'm using three 480mm thickest rads, one 240mm thick rad, two Revo D5 fans at full speed, and 18 Noctua Industrial 3000 RPM fans running at 70%. Also, I have an X299 Dark in the room.

M
M4blast
Member
67
07-17-2017, 05:34 AM
#2
The voltage applied was quite high for such a compact design.
M
M4blast
07-17-2017, 05:34 AM #2

The voltage applied was quite high for such a compact design.

K
kcaz56
Senior Member
664
07-17-2017, 01:44 PM
#3
Have you thought about smoothing the IHS? You might run a brief test and spin the CPU around the IHS to check if it rotates. If it’s convex, it should spin and would need polishing. Even if it’s concave and doesn’t spin, polishing is still necessary. How does the paste appear once you remove the water block?
K
kcaz56
07-17-2017, 01:44 PM #3

Have you thought about smoothing the IHS? You might run a brief test and spin the CPU around the IHS to check if it rotates. If it’s convex, it should spin and would need polishing. Even if it’s concave and doesn’t spin, polishing is still necessary. How does the paste appear once you remove the water block?

L
lok0000
Junior Member
18
07-17-2017, 03:23 PM
#4
Disable hyperthreading settings
L
lok0000
07-17-2017, 03:23 PM #4

Disable hyperthreading settings

R
randrake22
Junior Member
21
07-23-2017, 08:20 AM
#5
What CPU block is being considered? Could another block perform better? For my assessment, I’d evaluate removing the loop and adding a large reservoir of ice or cold water to absorb heat. Lowering the temperature further would require using only the most frigid water available.
R
randrake22
07-23-2017, 08:20 AM #5

What CPU block is being considered? Could another block perform better? For my assessment, I’d evaluate removing the loop and adding a large reservoir of ice or cold water to absorb heat. Lowering the temperature further would require using only the most frigid water available.

T
Tico_32
Senior Member
680
07-25-2017, 11:38 PM
#6
Also keen on voltage levels. These processors heat up quickly when voltage increases. On AVX-512 workloads (y-cruncher), I haven’t exceeded 1.25v, possibly a touch higher during lighter tasks like Cinebench R15.
T
Tico_32
07-25-2017, 11:38 PM #6

Also keen on voltage levels. These processors heat up quickly when voltage increases. On AVX-512 workloads (y-cruncher), I haven’t exceeded 1.25v, possibly a touch higher during lighter tasks like Cinebench R15.

K
kImbO95_
Junior Member
11
07-26-2017, 12:57 AM
#7
I operate at 1.3V now at 4.7Ghz and still experience comparable temperatures. My main issue is that the 7980XE only achieves 41 FPS in Physics score at 4.9Ghz, while others with similar settings reach around 51 FPS at 4.7Ghz. I have no idea what's causing this difference.
K
kImbO95_
07-26-2017, 12:57 AM #7

I operate at 1.3V now at 4.7Ghz and still experience comparable temperatures. My main issue is that the 7980XE only achieves 41 FPS in Physics score at 4.9Ghz, while others with similar settings reach around 51 FPS at 4.7Ghz. I have no idea what's causing this difference.

M
miknes123
Senior Member
646
07-26-2017, 06:41 AM
#8
Copper Velocity was handling it smoothly at 5.3Ghz on 9900K. The max voltage reached was 1.31V. The main issue is the poor TimeSpy physics score. CPU utilization sits around 50%, and Cinebench R15 confirms that priority setting to high gives solid results. 4650 is a solid number. Thermal Grizzly Kyronaut.
M
miknes123
07-26-2017, 06:41 AM #8

Copper Velocity was handling it smoothly at 5.3Ghz on 9900K. The max voltage reached was 1.31V. The main issue is the poor TimeSpy physics score. CPU utilization sits around 50%, and Cinebench R15 confirms that priority setting to high gives solid results. 4650 is a solid number. Thermal Grizzly Kyronaut.

C
Camsell
Member
207
07-26-2017, 08:42 AM
#9
I question if a standard benchmark tool can handle 18 cores, perhaps its stability isn't solid?
C
Camsell
07-26-2017, 08:42 AM #9

I question if a standard benchmark tool can handle 18 cores, perhaps its stability isn't solid?

D
Der_Zwieback
Member
148
07-27-2017, 07:42 AM
#10
Because the TimeSpy leaderboard contains many high scores, it can be unstable. If it weren't stable, it would crash.
D
Der_Zwieback
07-27-2017, 07:42 AM #10

Because the TimeSpy leaderboard contains many high scores, it can be unstable. If it weren't stable, it would crash.

Pages (3): 1 2 3 Next