LGA 1700 Alder Lake "Servers"

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
May 20, 2020
43
27
18
The spec sheet only advertises 4400Mhz, so I highly doubt it, but next time I can get into the BIOS I'll check (running an 8 hour long test now).
Confirming: BIOS allows frequency to be set, but 4400MHz is the highest. I left mine at Auto. No stability issues so far.

As hoped, this CPU is about 20% faster than the Ryzen 5900 for our CFD application, even though the Ryzen was tested on an enthusiast platform (using Ryzen Master to tweak CPU and RAM), vs. this conservative Supermicro platform (with bonus IPMI).

Side note: It's WAY faster than the older i9-7960x HEDT it replaces, solving the same benchmark in 2.79 hours on all 8 cores vs 4.60 on the 16 core 7960x. My engineers are going to like this. Now I need to build 2 more - though part of me really wants to test the Ryzen 5800x3d to see if that extra L3 helps - when I tested 2x Zen 2 chips, the 3900x and a 7302P, the extra cache on the EPYC did not seem to help that much. I'd rather not have another platform to support and last I checked ASrock was about the only show in town for Ryzen server boards so I think I'll leave than as an unknown.
 

mirrormax

Active Member
Apr 10, 2020
226
86
28
good to hear it worked out for you, Iam not a fan of ryzen consumer parts for workstations, had several issues with 5000series and ram stability etc, theres seems to be a lot of people having issues aswell. threadripper/epyc have been rock solid but unfortunately there isnt really a high clock SKU of those, ryzen 7000 series could be interesting ,might compete with the 12900k in low thread count.

looking at the supermicro board its has the worst VRM design ive seen of any lga17000 board even all the other w680 have at least 9 phases for cpu, supermicro went with 7. kinda regret ordering it now but theres nothing else to be found yet so guess it will do.
 
May 20, 2020
43
27
18
good to hear it worked out for you, Iam not a fan of ryzen consumer parts for workstations, had several issues with 5000series and ram stability etc, theres seems to be a lot of people having issues aswell. threadripper/epyc have been rock solid but unfortunately there isnt really a high clock SKU of those, ryzen 7000 series could be interesting ,might compete with the 12900k in low thread count.

looking at the supermicro board its has the worst VRM design ive seen of any lga17000 board even all the other w680 have at least 9 phases for cpu, supermicro went with 7. kinda regret ordering it now but theres nothing else to be found yet so guess it will do.
Yeah if TR 5000 (non-pro) was available that would have been on my short list too. A high clocked 12-16 core system with quad channel memory would be optimal for this workload.

FWIW per HDinfo and the IPMI VRM temps were a non issue during a sustained~ 5 hour P95 blend run, 300W system power consumption. Idle is ~31W with fans in optimal mode, 40W on full.
 

tjk

Active Member
Mar 3, 2013
485
201
43
I picked up one of the X13SAE-F boards to build a CFD solver for work. Only hiccup so far is the 1Gbit LAN disappeared (from BIOS and device manager), but a cold reboot fixed it. Running a 12900k and watercooling since these will see sustained high CPU usage. Board is revsion 1.02. Just using regular Crucial DDR5 (2x32GB), no extra ECC.

I've confirmed the out of box BIOS settings allow 24/7 241W CPU usage. I don't think I got a very good chip, it rarely hits any cores at 52x multiplier, but it'll do 49x all P core happily.

Our niche CFD software is only Win 10 approved and it scales poorly past 8 cores, so this seemed a good choice. I have E cores disabled while I run benchmarks on the CFD software.

Anyway, it'll be in testing for the next few days. If you have any questions or things for me to try, please reply.
Awesome! Can you detail your build in terms of chassis (desktop or rack mount), cooling, why no ECC, etc.
 

mirrormax

Active Member
Apr 10, 2020
226
86
28
Ddr5 ecc udimm chips are not readily available yet it seems.
Yeah if TR 5000 (non-pro) was available that would have been on my short list too. A high clocked 12-16 core system with quad channel memory would be optimal for this workload.

FWIW per HDinfo and the IPMI VRM temps were a non issue during a sustained~ 5 hour P95 blend run, 300W system power consumption. Idle is ~31W with fans in optimal mode, 40W on full.
yeah non pro tr5000 would be nice, good to hear the vrm temps where fine, thats actually surprising they stay that cool with the tiny heatsink. what case/rack do you have it in? mine should be here any day now hopefully the KS works even if its out of spec
 
May 20, 2020
43
27
18
...Can you detail your build in terms of chassis (desktop or rack mount), cooling, why no ECC, etc.
Ddr5 ecc udimm chips are not readily available yet it seems.

yeah non pro tr5000 would be nice, good to hear the vrm temps where fine, thats actually surprising they stay that cool with the tiny heatsink. what case/rack do you have it in? mine should be here any day now hopefully the KS works even if its out of spec
Agreed - no ECC because it's not available. My (limited) understanding is that there's already some form of ECC built into DDR5 as well which should cover most bases other than reporting of errors.

The existing solvers, now 4 years old, do not have ECC and it has not caused any known problems.

Here's the BOM:

BOM itemdescription / P/NApprox $
MBSupermicro X13SAE-F
500​
CPUi9-12900K
600​
RAM (2x32)Crucial CT2K32G48C40U5
420​
CaseFractal Meshify 2 XL
200​
CPU cooler, thermal compoundArctic Liquid Freezer II 420, Noctua NT-H2
150​
TPMAOM-TPM-9670V-S
50​
HDKXG60ZNV512G (picked because it's on the SM approved list)
140​
Case Fans (6)NF-A14 PWM
150​
PSUSeasonic Focus SSR-750PX
160​
Total:
2370​

This build was really all about the CPU. As mentioned upthread, the workload here is a niche CFD software; this software uses networked solvers, so the engineers configure the simulation using a frontend that can run on a laptop, then push it to a solver to do the number crunching.

It's all CPU, no GPU. It needs 2 channels of DDR4 (or 5) to scale well per each ~6 cores, so 8 cores with 2 channel RAM or 12-16 cores with 4 channel RAM are the sweet spots. More cores than that and CPU frequency drops off faster than the extra RAM channels help. Scaling is non-linear, so more than 16 cores does not really help regardless.

Minimum hardware list indicates a GPU is necessary, and it runs fine with the iGPU so I'm sticking with that. Part of the PSU sizing was so I could add a discrete GPU if needed.

The solutions generally take 1.5 - 4 hours to compute, though if an optimization problem is run (i.e. a DOE), back to back solutions are iterated that can take days of solid running. So the thermal solution needs to be designed to handle that.

I've done in house benchmarking on this software off and on over the last 4 years (when we first got this software) and based on that I suspected Alder Lake's P cores would be perfect (i.e. better than anything else on the market) if I could get 8 fast ones. Over and over I was reading that thermals were the primary constraint and that water cooling was needed to prevent thermal throttling under several types of heavy load.

Case was picked to for the ability to fit the 420mm AIO CPU cooler without compromise and to house enough case fans to allow for slight positive pressure in the case and with built in dust filters. I manually masked off selected openings in the case to control the airflow, preventing short circuiting, and directing the exhaust air where I want it (e.g. I blocked off most of the PCI slots, openings in the sheet metal, etc.).

If this AIO was not able to keep up my fallback was to get a 2U case and do custom watercooling and hook them into our plant's chilled water (60 deg F). Though a fun project, it has proven not to be necessary to run this CPU on its 8 P cores flat out and it would have added complexity (though the BOM cost would be about a wash, in house labor would be a lot higher). I'll be building 3 of these total - we have licensing for 3 simultaneous solvers, so another variable was "one big solver" vs 3 smaller ones.

Here's HWinfo screen cap after running P95 blend for about 55 minutes. I reset the counters about 4 minutes into the run. Temps are lower on the actual CFD software. In previous testing I have found P95 blend to be the best proxy for CPU frequency the actual CFD application will be able to achieve.

As you can see, VRM temps are healthy with this setup and the CPU is not thermally throttling, it is limited by the 241W power budget. Fan config is 5 NF-A14 as intake, 1 as CPU exhaust and the AIO as exhaust out the top of the case (stock fans are on the AIO, though I had higher pressure Noctuas on hand in case a swap was deemed appropriate after testing).

Resulting noise on full is just above background in our office environment. It was more noticeable at home (I did most of the build and initial testing there), but with the large diameter fans the frequency was not objectionable.

IPMI is currently set to fans at "Full". Since these will be sitting in a server room I may just leave them that way; currently playing with IPMI fan minimum thresholds to see if I can do otherwise. Putting them on optimal (variable speed) would help with long term dust collection as these are not active 24/7. If I do that, I'll move them all to the CPU Fan headers so the PWM on the case fans scales with CPU temps (as that's the only load in the system).

P95 blend temp testing, fans on full pic2.PNG
2022 Solver build 1 case pic - shrunk and redacted.jpg
 

MrBios

New Member
Feb 21, 2022
17
3
3
any vendors selling the supermicro or any other brands motherboards with IPMI option for 12th gen ?
 

nightcralwer

New Member
Jun 22, 2022
2
0
1
Bought a x13sae-f and put in i7-12700k for home server. I wonder if this board is faulty. I can get into IPMI, BIOS, and do memtest86 test without issues. But it just refused to boot into any os transferred from old server. Windows 11, Ubuntu Linux, and Proxmox. All the same. Trying to freshly install os also hangs the machine, it just stops right at the moment the installation process is about to begin.

Thougnts?
 
May 20, 2020
43
27
18
Bought a x13sae-f and put in i7-12700k for home server. I wonder if this board is faulty. I can get into IPMI, BIOS, and do memtest86 test without issues. But it just refused to boot into any os transferred from old server. Windows 11, Ubuntu Linux, and Proxmox. All the same. Trying to freshly install os also hangs the machine, it just stops right at the moment the installation process is about to begin.

Thougnts?
The recent supermicro boards I’ve had (not many, N=3) are really cumbersome and unintuitive when it comes to setting up secure boot. Is it disabled or did you go through that process?

honestly can’t remember if that’s something you set up before you load the OS (with the usb drive plugged in) or after.

I loaded Windows 10 Pro from the latest Microsoft.ISO and it went just fine via a USB drive.
 

MrBios

New Member
Feb 21, 2022
17
3
3
Bought a x13sae-f and put in i7-12700k for home server. I wonder if this board is faulty. I can get into IPMI, BIOS, and do memtest86 test without issues. But it just refused to boot into any os transferred from old server. Windows 11, Ubuntu Linux, and Proxmox. All the same. Trying to freshly install os also hangs the machine, it just stops right at the moment the installation process is about to begin.

Thougnts?
could be ram issue ?
 

vincococka

Member
Sep 29, 2019
50
26
18
Slovakia
How about small home/LAB server build based on ASUS Q670M-C board with some vPRO Alder Lake CPU ?
Maybe some users will find this combination as sweet spot between price and functionality.
 

zer0sum

Well-Known Member
Mar 8, 2013
876
490
63
How about small home/LAB server build based on ASUS Q670M-C board with some vPRO Alder Lake CPU ?
Maybe some users will find this combination as sweet spot between price and functionality.
vPRO is trash compared to a real IPMI chipset though! :(

At least it's relatively cheap, but that's an oddball motherboard.
  • DDR5, but no PCIe 5?
  • Only 1G lan port?
  • 2 x PCIe x1 slots?
 

vincococka

Member
Sep 29, 2019
50
26
18
Slovakia
vPRO is trash compared to a real IPMI chipset though! :(

At least it's relatively cheap, but that's an oddball motherboard.
  • DDR5, but no PCIe 5?
  • Only 1G lan port?
  • 2 x PCIe x1 slots?
While I agree that real IPMI is usually lot better - for somebody vPRO can be sufficient and also cheaper.
At least with vPRO there is KVM + remote on/off which for home lab is sufficient.

I believe also that for most users PCIe 4.0 is also more than acceptable - it is not easy to saturate PCIe 5.0 link (now).
1GbE LAN is suboptimal, but one can buy 2.5GbE NIC if required.

It was meant for some small server / remote workstation which can be sufficient and not so expensive as "bigger brothers based on W680 + IPMI".
 

shadowx360

New Member
Jun 20, 2018
2
0
1
I picked up one of the X13SAE-F boards to build a CFD solver for work. Only hiccup so far is the 1Gbit LAN disappeared (from BIOS and device manager), but a cold reboot fixed it. Running a 12900k and watercooling since these will see sustained high CPU usage. Board is revsion 1.02. Just using regular Crucial DDR5 (2x32GB), no extra ECC.

I've confirmed the out of box BIOS settings allow 24/7 241W CPU usage. I don't think I got a very good chip, it rarely hits any cores at 52x multiplier, but it'll do 49x all P core happily.

Our niche CFD software is only Win 10 approved and it scales poorly past 8 cores, so this seemed a good choice. I have E cores disabled while I run benchmarks on the CFD software.

Anyway, it'll be in testing for the next few days. If you have any questions or things for me to try, please reply.
What’s idle wattage like at the wall? My home workstation also hosts my Plex VM so I run it 24/7.