Minisforum MS-01 PCIe Card and RAM Compatibility Thread

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

nicoska

Member
Feb 8, 2023
55
10
8
So, after I little bit more digging, I tried to disable Efficiency core (via BIOS).
So now my Proxmox host run on only "12" core, but at least seems [finger crossed!] stable.
1708469465878.png

I'm able to run the Windows 11 VM which caused the crash before.

I'll keep you posted.

Nico
 
  • Like
Reactions: GreenAvacado
H

homelabenthusiast

Member
Feb 19, 2024
30
32
18
FingerBlaster said:
my 13900h version is running the 96gb kit of Crucial ram that Patrick used and i'm running at 5200 MT/S with no problems yet. I do plan on doing an extended memtest before going into homelab "production"

Code: 
root@pve1:~# dmidecode --type 17
# dmidecode 3.4
Getting SMBIOS data from sysfs.
SMBIOS 3.5.0 present.

Handle 0x0028, DMI type 17, 92 bytes
Memory Device
        Array Handle: 0x0027
        Error Information Handle: Not Provided
        Total Width: 64 bits
        Data Width: 64 bits
        Size: 48 GB
        Form Factor: SODIMM
        Set: None
        Locator: Controller0-ChannelA-DIMM0
        Bank Locator: BANK 0
        Type: DDR5
        Type Detail: Synchronous
        Speed: 5600 MT/s
        Manufacturer: Crucial Technology
        Serial Number: E7AE8D4B
        Asset Tag: 9876543210
        Part Number: CT48G56C46S5.M16B1
        Rank: 2
        Configured Memory Speed: 5200 MT/s
        Minimum Voltage: 1.1 V
        Maximum Voltage: 1.1 V
        Configured Voltage: 1.1 V
        Memory Technology: DRAM
        Memory Operating Mode Capability: Volatile memory
        Firmware Version: Not Specified
        Module Manufacturer ID: Bank 6, Hex 0x9B
        Module Product ID: Unknown
        Memory Subsystem Controller Manufacturer ID: Unknown
        Memory Subsystem Controller Product ID: Unknown
        Non-Volatile Size: None
        Volatile Size: 48 GB
        Cache Size: None
        Logical Size: None

Handle 0x0029, DMI type 17, 92 bytes
Memory Device
        Array Handle: 0x0027
        Error Information Handle: Not Provided
        Total Width: 64 bits
        Data Width: 64 bits
        Size: 48 GB
        Form Factor: SODIMM
        Set: None
        Locator: Controller1-ChannelA-DIMM0
        Bank Locator: BANK 0
        Type: DDR5
        Type Detail: Synchronous
        Speed: 5600 MT/s
        Manufacturer: Crucial Technology
        Serial Number: E7AE8DC6
        Asset Tag: 9876543210
        Part Number: CT48G56C46S5.M16B1
        Rank: 2
        Configured Memory Speed: 5200 MT/s
        Minimum Voltage: 1.1 V
        Maximum Voltage: 1.1 V
        Configured Voltage: 1.1 V
        Memory Technology: DRAM
        Memory Operating Mode Capability: Volatile memory
        Firmware Version: Not Specified
        Module Manufacturer ID: Bank 6, Hex 0x9B
        Module Product ID: Unknown
        Memory Subsystem Controller Manufacturer ID: Unknown
        Memory Subsystem Controller Product ID: Unknown
        Non-Volatile Size: None
        Volatile Size: 48 GB
        Cache Size: None
        Logical Size: None
Click to expand...
So you don’t see any mce hardware errors at all in the logs?? Which proxmox version are you running? And kernel?
 
H

homelabenthusiast

Member
Feb 19, 2024
30
32
18
DataX said:
I've been running the 13900H with the 2x48G Crucial kit at 5200MT/s and so far so good. I've had a few load related crashes which may have been due to the E-cores, but nothing that would indicate memory instability.

I would agree with the other comment, try re-seating or swapping the DIMMs around, and test each one individually. You could also unplug AC and fully drain power from the system between tests, might help ensure memory retrains.

Code: 
# dmidecode 3.3
Getting SMBIOS data from sysfs.
SMBIOS 3.5.0 present.

Handle 0x0027, DMI type 16, 23 bytes
Physical Memory Array
        Location: System Board Or Motherboard
        Use: System Memory
        Error Correction Type: None
        Maximum Capacity: 64 GB
        Error Information Handle: Not Provided
        Number Of Devices: 2

Handle 0x0028, DMI type 17, 92 bytes
Memory Device
        Array Handle: 0x0027
        Error Information Handle: Not Provided
        Total Width: 64 bits
        Data Width: 64 bits
        Size: 48 GB
        Form Factor: SODIMM
        Set: None
        Locator: Controller0-ChannelA-DIMM0
        Bank Locator: BANK 0
        Type: DDR5
        Type Detail: Synchronous
        Speed: 5600 MT/s
        Manufacturer: Crucial Technology
        Part Number: CT48G56C46S5.M16B1
        Rank: 2
        Configured Memory Speed: 5200 MT/s
        Minimum Voltage: 1.1 V
        Maximum Voltage: 1.1 V
        Configured Voltage: 1.1 V
        Memory Technology: DRAM
        Memory Operating Mode Capability: Volatile memory
        Firmware Version: Not Specified
        Module Manufacturer ID: Bank 6, Hex 0x9B
        Module Product ID: Unknown
        Memory Subsystem Controller Manufacturer ID: Unknown
        Memory Subsystem Controller Product ID: Unknown
        Non-Volatile Size: None
        Volatile Size: 48 GB
        Cache Size: None
        Logical Size: None

Handle 0x0029, DMI type 17, 92 bytes
Memory Device
        Array Handle: 0x0027
        Error Information Handle: Not Provided
        Total Width: 64 bits
        Data Width: 64 bits
        Size: 48 GB
        Form Factor: SODIMM
        Set: None
        Locator: Controller1-ChannelA-DIMM0
        Bank Locator: BANK 0
        Type: DDR5
        Type Detail: Synchronous
        Speed: 5600 MT/s
        Manufacturer: Crucial Technology
        Part Number: CT48G56C46S5.M16B1
        Rank: 2
        Configured Memory Speed: 5200 MT/s
        Minimum Voltage: 1.1 V
        Maximum Voltage: 1.1 V
        Configured Voltage: 1.1 V
        Memory Technology: DRAM
        Memory Operating Mode Capability: Volatile memory
        Firmware Version: Not Specified
        Module Manufacturer ID: Bank 6, Hex 0x9B
        Module Product ID: Unknown
        Memory Subsystem Controller Manufacturer ID: Unknown
        Memory Subsystem Controller Product ID: Unknown
        Non-Volatile Size: None
        Volatile Size: 48 GB
        Cache Size: None
        Logical Size: None
Click to expand...
have you checked you logs for any errors?
 
J

JaxJiang

Member
Jan 10, 2023
34
54
18
HomelabArchitect said:
@wadup I'm also having trouble getting vPro working. I can connect via a browser on port 16993 but can't get MeshCommander (using mesh-mini) to connect. It just times out. I'm using IP:16693 for hostname and Digest / TLS for Auth. Any ideas? The vPro page in the browser doesn't show Remote Control.
Click to expand...
Hi.
1. Make sure you already enable MEBX in bios.
2. Make sure you enable vpro setting: "Network Active"
3. Make sure your management port is I226-LM(the rj45 next to SFP+)
Website access, suggestion input by your hand, https://[you-ip-address]:16993
By default the ip address of you manangement is same with you dhcp ip addr. In some case, if no any dhcp package let vpro get a ip. You need setting a ip by manually.
 
S

skipper ohms

Member
Jan 24, 2024
35
23
8
BlueChris said:
All the Dell 240w power supplies thati found are 19.5v... the machine needs 19v
Click to expand...
19.5v and 20v are fine, they are basically the same. DC powered devices accept a small voltage range because voltages drop when the device draws power. As long as you stay within that range it's fine. However, it's always important to have fire alarm/suppression/escape tools close by regardless of what you're doing.

Make sure you get the right size barrel connector on the brick because most barrel adapters are rated for minimal amps. Pretty sure MS-01 uses a 5.5mm x 2.5mm DC barrel, but I don't have one in front of me yet so ymmv.

The SlimQ 240w is nice because it has a 5525 barrel port plus two USB PD ports so you can power your hungry mini off the barrel and a couple not so hungry devices using USB PD to DC cables.
 
  • Like
Reactions: NerdAshes, OJO360 and LiKenun
H

HomelabArchitect

New Member
Feb 20, 2024
2
0
1
JaxJiang said:
Hi.
1. Make sure you already enable MEBX in bios.
2. Make sure you enable vpro setting: "Network Active"
3. Make sure your management port is I226-LM(the rj45 next to SFP+)
Website access, suggestion input by your hand, https://[you-ip-address]:16993
By default the ip address of you manangement is same with you dhcp ip addr. In some case, if no any dhcp package let vpro get a ip. You need setting a ip by manually.
Click to expand...
I did all of that. I can access the webpage but there is no KVM/Remote Desktop option
vPro.png
 
A

adman_c

Active Member
Feb 14, 2016
271
144
43
Chicago
nicoska said:
So, after I little bit more digging, I tried to disable Efficiency core (via BIOS).
So now my Proxmox host run on only "12" core, but at least seems [finger crossed!] stable.
View attachment 34820

I'm able to run the Windows 11 VM which caused the crash before.

I'll keep you posted.

Nico
Click to expand...
Did you install the most recent intel microcode? That supposedly helped with some issues with big/little cores and proxmox. Someone posted a craft computing YouTube link with it earlier in this thread.
 
  • Like
Reactions: GreenAvacado, OJO360 and skipper ohms
F

FreakyBigFoot

New Member
Feb 4, 2024
14
2
3
I'm absolutely loving my 13900H with my QNAP TL-D800S so far!

I installed the PCIE card in just a couple minutes. I'm also using 96GB of RAM from the first comment on this thread :)

So far everything has been great. I'm running unraid (I still need to actually hang this rack)
 

Attachments

H

homelabenthusiast

Member
Feb 19, 2024
30
32
18
homelabenthusiast said:
I also have MS-01(13900H) running proxmox with 96GB memory from crucial like yours (just not kit bought 2 48GB separately because it was cheaper) and i also have the same problem but my server behaves normally. I do not experience random reboots or anything. I just see the errors. I tried memtest86+, tried unplugging external devices and I tried running single ram (tried each ram) and not running any vm at all. But i still see the errors still. I also tried updaating and dowgrading the kernel but it made it just worse! Like i saw Memory failure every second! That was teribble! Running kernel 5.15.136-1-pve i see the error only every minute. Maybe someone else can help? For me it would be enough to see any reponse that you are running MS-01 with 96GB crucial and it is working without the hardware or memory error. I also want to know if you are seeing the errors. Just want to know exactly if this is normal behavior or not. I would be really happy for any help. :) I see this:

View attachment 34805
Click to expand...
@JaxJiang is this a normal behavior for this setup? Could you please verify. Maybe you have any solutions or suggestion? Thanks you
 
T

the-last-englishman

Member
Aug 20, 2023
81
43
18
FreakyBigFoot said:
I'm absolutely loving my 13900H with my QNAP TL-D800S so far!

I installed the PCIE card in just a couple minutes. I'm also using 96GB of RAM from the first comment on this thread :)

So far everything has been great. I'm running unraid (I still need to actually hang this rack)
Click to expand...
What is that server rack, I want one :)
 
N

nicoska

Member
Feb 8, 2023
55
10
8
nicoska said:
So, after I little bit more digging, I tried to disable Efficiency core (via BIOS).
So now my Proxmox host run on only "12" core, but at least seems [finger crossed!] stable.
View attachment 34820

I'm able to run the Windows 11 VM which caused the crash before.

I'll keep you posted.

Nico
Click to expand...
So, no reboot or crash so far.
Can anyone explain to me if running it with E-cores disabled is a real under power unit? I mean performance wise.

Thank you
Nico
 
B

BlueChris

Active Member
Jul 18, 2021
119
40
28
52
Athens-Greece
Jaxjiang replied here for the hardware problem of the memory and some other info also.

Seems no biggy but it seems there is a newer hardware revision of MS-01, i hope he will reply which is which to know what revision we have.
With the above info in mind, maybe this is the reason that some cards refuse to work in the STH tests and to other people so a tape must be inserted in the specific PCI-E smbus pins to circumvent the problem
 
F

Fazio

New Member
Dec 20, 2022
28
21
3
Anyone is able to retrieve and control the fan values from Debian/Proxmox host on MS-01?
 
W

wadup

Member
Feb 13, 2024
98
66
18
skipper ohms said:
19.5v and 20v are fine, they are basically the same. DC powered devices accept a small voltage range because voltages drop when the device draws power. As long as you stay within that range it's fine. However, it's always important to have fire alarm/suppression/escape tools close by regardless of what you're doing.

Make sure you get the right size barrel connector on the brick because most barrel adapters are rated for minimal amps. Pretty sure MS-01 uses a 5.5mm x 2.5mm DC barrel, but I don't have one in front of me yet so ymmv.

The SlimQ 240w is nice because it has a 5525 barrel port plus two USB PD ports so you can power your hungry mini off the barrel and a couple not so hungry devices using USB PD to DC cables.
Click to expand...
Minisforum support said MS-01 barrel port is 5.5mm x 2.5mm.
 
  • Like
Reactions: NerdAshes, FreakyBigFoot, OJO360 and 1 other person
A

adman_c

Active Member
Feb 14, 2016
271
144
43
Chicago
nicoska said:
So, no reboot or crash so far.
Can anyone explain to me if running it with E-cores disabled is a real under power unit? I mean performance wise.

Thank you
Nico
Click to expand...
The internets says that an E core has approximately 60% of the performance of a P core. So by disabling those 8 E cores you're losing performance in the neighborhood of 4-5 P cores. So you're definitely getting a lot less performance than you paid for. Your issues sound very much like the microcode issue described here by Craft Computing. The solution is described there as well. Credit to @GreenAvacado for linking the vid earlier in the thread.

I'm very interested in hearing whether the microcode update improves your experience because I plan to run proxmox on mine and losing the E-cores (or experiencing persistent stability issues) would definitely ruin a lot of the appeal of this box for me.
 
Last edited:
  • Like
Reactions: GreenAvacado
W

wadup

Member
Feb 13, 2024
98
66
18
FreakyBigFoot said:
I'm absolutely loving my 13900H with my QNAP TL-D800S so far!

I installed the PCIE card in just a couple minutes. I'm also using 96GB of RAM from the first comment on this thread :)

So far everything has been great. I'm running unraid (I still need to actually hang this rack)
Click to expand...
MS-01 Is such a good Unraid server coupled with your Qnap Expansion which is really smart. QNAP with MS-01 hardware would be many thousands of dollars and a lot bigger footprint.
 
S

stich86

Member
May 24, 2023
40
12
8
Ok I got also AMT to works, still need to test KVM.

just two questions:

- under Proxmox, if using same Ethernet card that runs both management and AMT, using Linux bridge, start flooding the dmesof address spoofing (because AMT and Proxmox share same MAC address but different IP). Any solution on this?

- still under Proxmox (based on Debian 12 with kernel 6.2), a power off doesn’t shutdown the unit. Any suggestion?

thanks
 
N

nicoska

Member
Feb 8, 2023
55
10
8
adman_c said:
The internets says that an E core has approximately 60% of the performance of a P core. So by disabling those 8 E cores you're losing performance in the neighborhood of 4-5 P cores. So you're definitely getting a lot less performance than you paid for. Your issues sound very much like the microcode issue described here by Craft Computing. The solution is described there as well. Credit to @GreenAvacado for linking the vid earlier in the thread.

I'm very interested in hearing whether the microcode update improves your experience because I plan to run proxmox on mine and losing the E-cores (or experiencing persistent stability issues) would definitely ruin a lot of the appeal of this box for me.
Click to expand...
As I said earlier in the thread, I've already applied the latest microcode update :/
1708535870947.png
 
You must log in or register to reply here.
Top Bottom