Single Socket AMD EPYC 7000 FAQ Answers to Common Questions

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

Pri

Active Member
Jul 30, 2014
124
52
28
Thanks Patrick this was a great post that answered questions I had. I really wish we had dates for when we can buy motherboard and processors to build ourselves. I only really need a board and CPU to upgrade my server, I don't need a whole system.
 
E

Edu

Member
Aug 8, 2017
55
8
8
33
Very interesting article.
I remember a while back you said that Epyc servers had some type of hot plug problem. I'm wondering if that has been resolved yet?
 
B

bitrot

Member
Aug 7, 2017
95
25
8
Great article as always, thanks for that.

If availability was better and if I actually needed all that horsepower for my home server, I would have definitely considered a single socket EPYC setup, most likely an EPYC 7351P.

The major deal breaker for me was power consumption, though. Not just because of costs (energy prices are ridiculous over here), but also for 'thermal' reasons: cooling 300 and more Watts in my 2U Supermicro 826 chassis is definitely not possible with an acceptable acoustic footprint, which is a very important criteria to me. Hence, I went the Xeon Scalable route (Xeon 4114 + Supermicro X11SPH-NCTF, which is one hell of a board btw., made my SAS controllers obsolete and adding NVMEs a breeze).
 
L

loggedinasroot

Member
Jul 30, 2017
53
9
8
54
"A: This one surprised us a bit. We saw about a 3W difference in the P part versus the non-P part in terms of power consumption. It was small, but it was fairly consistent."

Does this mean that the P part uses 3W more or less?
If the P part uses less it would probably have something to do with the lack of cross CPU infinity-fabric right?
But any idea what may cause this if it uses more?
 
  • Like
Reactions: Edu
Patrick

Patrick

Administrator
Staff member
Dec 21, 2010
12,513
5,804
113
loggedinasroot said:
"A: This one surprised us a bit. We saw about a 3W difference in the P part versus the non-P part in terms of power consumption. It was small, but it was fairly consistent."

Does this mean that the P part uses 3W more or less?
If the P part uses less it would probably have something to do with the lack of cross CPU infinity-fabric right?
But any idea what may cause this if it uses more?
Click to expand...
Less. IF takes some power. Small difference but it was surprising since in theory that power budget gets re-allocated to the rest of the chip (hence ever so slightly higher performance.)
 
You must log in or register to reply here.
Top Bottom