EU Gigabyte Mainboard MC12-LE0 Re1.0 AMD B550 AM4 Ryzen
- Thread starter flobert
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Have LSI 9300-8i running with 0 problems in one of my boards
I can test a LSI 9211-8i on monday, if you don't get it solved over the weekend
D
Deleted member 45030
Guest
There are a bunch from Intertech if you are in europe,
Server Case - Inter-Tech Elektronik Handels GmbH
www.inter-tech.de
keep in mind most standard atx io bezels are just to tall to fit inside 1u.
You might want to look into 2u cases with a better low profile cooler if you can, these 1u coolers that you could get for an additional 10€ with the deal are close to unusable in therms of temperature, i got one of them and tested it with a ryzen 3700x (65w tdp) and it goes to over 90°C within seconds of any load...
No issue what so ever with the 1u cooler with a 3600 in my case. I let prime95 run for hours, I don’t think I saw more than 70C.
Maybe you could have started with giving us the power consumption of your system? 
My system could be similar (you did not list everything): 5700X, 2x 32GB Samsung, BMC57414 (one port linked at 25GbE using a DAC), Optane M10 16GB, beQuiet 400W (about 10 years old), 4x NF-S12A, Dynatron A24 (all fans are running at 40% to keep NIC cool) OS is fresh install of Ubuntu 22.04 server updated to newest packages but 5.15 kernel
Power consumption at idle is slightly below 33W (measured externally).
My system could be similar (you did not list everything): 5700X, 2x 32GB Samsung, BMC57414 (one port linked at 25GbE using a DAC), Optane M10 16GB, beQuiet 400W (about 10 years old), 4x NF-S12A, Dynatron A24 (all fans are running at 40% to keep NIC cool) OS is fresh install of Ubuntu 22.04 server updated to newest packages but 5.15 kernel
Power consumption at idle is slightly below 33W (measured externally).
Sorry was complaining about power usage and forgot to mention it duh.Maybe you could have started with giving us the power consumption of your system?
My system could be similar (you did not list everything): 5700X, 2x 32GB Samsung, BMC57414 (one port linked at 25GbE using a DAC), Optane M10 16GB, beQuiet 400W (about 10 years old), 4x NF-S12A, Dynatron A24 (all fans are running at 40% to keep NIC cool) OS is fresh install of Ubuntu 22.04 server updated to newest packages but 5.15 kernel
Power consumption at idle is slightly below 33W (measured externally).
Mike is idling at 72w.and that is literally everything I have (plus 4 noctua fans but those should be negligible).
Running esxi with only 1 opnsense vm. The ssd is a crucial p2. Psu is a be quiet tfx-300.
The nic is a silicon PE210G4SPI9. That's basically a 8724 pex chip with 2 x520s cramed into 1 card (should only be 10w with one sfp+ cable attached like I have). The cpu is mostly idling and stays at 1% at all times.
That's not how that works. That's not how anything works. And since I'm running exsi I can prove it.(shout out to esxi for using the dumbest possible representation of cpu time usage possible)
Also without the 10G nic it consumes 61w
Attachments
Truth might be in the middle (like most of the time).
8724 pex is officially 5W. But old dual port 10GbE NICs are more than 5W so I am pretty shure your NIC is more than 10W (count of connected ports is only a minor difference in all 10/25GbE server NICs I know)
Apart from that the major thing is: CPU power consumption is not tightly connected to CPU load. There are enormous differences possible depending on type of load and (this is probably the major point here) CPU sleep states! Depending on the type of load sleep states are practically never reached despite CPU load being low. Next thing is difference in OS. I don't have experience with ESXi but I have read multiple times that it practically disables sleep states for performance reasons (latency). This last point could be the major factor here.
8724 pex is officially 5W. But old dual port 10GbE NICs are more than 5W so I am pretty shure your NIC is more than 10W (count of connected ports is only a minor difference in all 10/25GbE server NICs I know)
Apart from that the major thing is: CPU power consumption is not tightly connected to CPU load. There are enormous differences possible depending on type of load and (this is probably the major point here) CPU sleep states! Depending on the type of load sleep states are practically never reached despite CPU load being low. Next thing is difference in OS. I don't have experience with ESXi but I have read multiple times that it practically disables sleep states for performance reasons (latency). This last point could be the major factor here.
I have literally pulled it out and used the built in nic and compared to get the 10-11w value. So I'm sure about that one.
Esxi used to be good at letting things go to higher c states with low power mode enabled (which I have). I'll need to install debian and redo my setup with kvm to test. Good idea.
First thing might be to exchange the SSD (or put a second in) and install plain Ubuntu server just to have a comparison. If that results in a big difference, I agree that it would be worth the effort and be quite interesting to run your VM(s) in a Linux/KVM environment of your choice.
Hi, I'm planning to purchase this motherboard for my Proxmox server and have a few questions:
1. I intend to install a second-generation Ryzen (2600) CPU on this board. Although theoretically not supported, there have been instances where second-generation processors work flawlessly on B550 chipsets. Has anyone successfully done this with this particular board?
2. Regarding the onboard graphics: my CPU lacks an integrated GPU, and I'd like the option to connect a monitor directly to the motherboard's VGA port for troubleshooting, without using a discrete graphics card. Does this motherboard come with integrated graphics similar to what an ASRock Rack motherboard offers? I'm trying to avoid using one of the limited PCI-E slots for a graphics card.
3. I plan to upgrade to a Ryzen 3900X in the future. Is anyone running such a high-performance CPU on this board? My concern is whether the VRM (Voltage Regulator Module) can handle the increased power demand.
4. How is the stability? I've seen posts from a year ago mentioning stability issues, but there seem to be improvements following updates. Can anyone provide current feedback on this?
Thanks in advance for your assistance. Have a great day
1. I intend to install a second-generation Ryzen (2600) CPU on this board. Although theoretically not supported, there have been instances where second-generation processors work flawlessly on B550 chipsets. Has anyone successfully done this with this particular board?
2. Regarding the onboard graphics: my CPU lacks an integrated GPU, and I'd like the option to connect a monitor directly to the motherboard's VGA port for troubleshooting, without using a discrete graphics card. Does this motherboard come with integrated graphics similar to what an ASRock Rack motherboard offers? I'm trying to avoid using one of the limited PCI-E slots for a graphics card.
3. I plan to upgrade to a Ryzen 3900X in the future. Is anyone running such a high-performance CPU on this board? My concern is whether the VRM (Voltage Regulator Module) can handle the increased power demand.
4. How is the stability? I've seen posts from a year ago mentioning stability issues, but there seem to be improvements following updates. Can anyone provide current feedback on this?
Thanks in advance for your assistance. Have a great day
The Ryzen 2600 will be a challenge…
the AST2500 has offers basic VGA output- totally sufficient if not transcoding
3900X: use a decent top blow cooler and set to 65W eco mode
stability is fine
only downside for me: iGPU not working for transcoding, because cannot be activated in BIOS
the AST2500 has offers basic VGA output- totally sufficient if not transcoding
3900X: use a decent top blow cooler and set to 65W eco mode
stability is fine
only downside for me: iGPU not working for transcoding, because cannot be activated in BIOS
Okay, and how do you evaluate iPMI, mainly that's what I'm interested in, because my server is located in a place that's quite far from my home. Have you maybe used PCI bifurcation, I heard it exists, but I also heard that it sometimes causes problems.
IPMI is actually pretty good (better than my Asrock Rack X470D4U, but also AST2500 vs. 2400)
Bifurcation is possible (8x4x4 with APUs, or 4x4x4x4 with regular CPUs)