280W TDP CPU in Supermicro 2U?

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

neobenedict

Member
Oct 2, 2020
64
21
8
Has anyone had experience running high TDP Epyc CPUs (e.g. 7763) with 280W TDP in a Supermicro 2U (826)? The only 2U heatsink (SNK-P0063AP4) is rated just 180W TDP so I am worried the cooling will not be sufficient and 4U is needed.
 
Last edited:

i386

Well-Known Member
Mar 18, 2016
4,315
1,592
113
35
Germany
I have ordered a SNK-P0063AP4 to test it with a threadripper pro system...
I asked the support, SNK-P0066AP4 and SNK-P0064AP4 are the only verified heatsinks for these 280watt tdp cpus. The support mentioned a 1u liquid cooling solution (SNK-P3045A4) but I can't find any pictures or specifications for it.
 

i386

Well-Known Member
Mar 18, 2016
4,315
1,592
113
35
Germany
Threaripper Pro 3955WX on a M12SWA with 128GB ddr4 3200MHZ ram

Ambient temp: 28°C
Humidty: 40-44% (Analog hygrometer)
Fan settings: Optimum speed

Stress test with Prime95 v30.8 b15
SNK-P0064AP4: 90°C
SNK-P0063AP4 without airshroud: 89°C
SNK-P0063AP4 with airshroud: 90°C

Idle temps:
SNK-P0064AP4: 50°C
(Forgot to test without the airshroud)
SNK-P0063AP4 with airshroud: 52°C

With the fans setting full speed the idle temps drop with the 2u heatsink as low as 42°C while the 4u heatsink never got below 48°C

One thing I noticed on the SNK-P0064AP4 when I removed it was that the thermal grease trapped a lot of air bubbles...


Edit: after 20 minutes running prime95 the cpu temp stays most of the time at 89°C, occasionally 88°C (the thermal grease is probably still "melting")
Something different I noticed with the 2u heatsink: the midwall fans run pretty consistent at the same speed while with the 4u heatsink they would ramp up for a few seconds when you started a program or opened a new tab in a browser...
 
Last edited:

i386

Well-Known Member
Mar 18, 2016
4,315
1,592
113
35
Germany
(Forgot to test without the airshroud)
I rerun the tests to see what differences in temps I will get now that the thermal grease is set

Ambient temp: 27°C
Humidty: 50-52% (Analog hygrometer)
Fan settings: Optimum speed

Idle temps:
SNK-P0063AP4 without airshroud: 47°C
SNK-P0063AP4 with airshroud: 47°C

The SNK-P0063AP4 is definitely cooler than the SNK-P0064AP4 for idle temps. Under full load there is no measurable differences in temperatures.
 

RolloZ170

Well-Known Member
Apr 24, 2016
5,767
1,760
113
One thing I noticed on the SNK-P0064AP4 when I removed it was that the thermal grease trapped a lot of air bubbles...
how do you apply the grease ?
i make just a 1,5" line in the middle of the IHS, no bubbles possible.
note that thermal grease is not a good heat conductor(compared with metal), it just have to fill the air gaps.
 

i386

Well-Known Member
Mar 18, 2016
4,315
1,592
113
35
Germany
It was the preapplied thermal grease from the SNK-P0064AP4 heatsink.
Photo from another heatsink:
abc.jpg

I would use the "x method with dots" like amd shows in their videos:
(Funny thing: the guy uses the threadripper(pro) wrench and taped the logos in that video:D)
 

TXAG26

Active Member
Aug 2, 2016
402
123
43
I rerun the tests to see what differences in temps I will get now that the thermal grease is set

Ambient temp: 27°C
Humidty: 50-52% (Analog hygrometer)
Fan settings: Optimum speed

Idle temps:
SNK-P0063AP4 without airshroud: 47°C
SNK-P0063AP4 with airshroud: 47°C

The SNK-P0063AP4 is definitely cooler than the SNK-P0064AP4 for idle temps. Under full load there is no measurable differences in temperatures.
During your testing of the 63AP4 and 64AP4 heatsinks, was there any appreciable noise difference between the 2U and 4U heatsinks?