Asus C621E not posting

[GreatJay23]

Hi all,

I have a WS C621E Sage motherboard that I bought to house my two Xeon Platinum 8124M CPU’s.
I stuck in 12 sticks of Crucial 8gb ram, am powering this with a supernova 1500 watt power supply, and have a WD black blank drive in.

my issue is, when I try to turn the system on, everything comes to life but I have no beep, no post, no video, and the code displayed is “00”.

I have tried only using one CPU, no difference. I heard of some people needing to update the Bios to use these CPU’s so I did that process with a usb drive, no change. I stuck a gpu in just for the heck of it, and no change.

I have a hard time believing both of these CPU’s are dead. But I don’t know what else to do.

Anyone ever have any experiences like this?

Thanks,
Jay

 

[RolloZ170]

Xeon Platinum 8124M is OEM high TDC (255A) processor.
WS C621E Sage supports TDC up to 228A (std, for all non OEMSKUs)
the PCH compares the TDC of the CPU (DMI) with the VRMs ICC_MAX (228A in this case)
if not same or lower, CPU is helt in reset. no x86 code is executed.
i have still not found a way to modify the ICC_MAX on the WS C621E Sage...
i need a board at home for that.

 

[RolloZ170]

I heard of some people needing to update the Bios to use these CPU’s

no. nothing special, it is skylake H0.

 

[RolloZ170]

not working CPUs
P-8124 (SKL B1 50653)
8124M
8175M
8222CL
8251C
8259CL
8272CL
8275CL
and others.

 

[GreatJay23]

not working CPUs
P-8124 (SKL B1 50653)
8124M
8175M
8222CL
8251C
8259CL
8272CL
8275CL
and others.

So these xeons won’t work with my mobo?

 

[GreatJay23]

Xeon Platinum 8124M is OEM high TDC (255A) processor.
WS C621E Sage supports TDC up to 228A (std, for all non OEMSKUs)
the PCH compares the TDC of the CPU (DMI) with the VRMs ICC_MAX (228A in this case)
if not same or lower, CPU is helt in reset. no x86 code is executed.
i have still not found a way to modify the ICC_MAX on the WS C621E Sage...
i need a board at home for that.

I’m tempted to offer shipping you my board, CPU’s, and pay you in the 12 sticks of ram lol. I want to get ECC ram for this but had 12 sticks of non.

 

[Bert]

I cannot understand why a WS motherboard not enough power delivery to support all possible Xeons. This is unbelievable. @RolloZ170 what would be the best lga 3647 or lga 4189 Workstation motherboard?

 

[RolloZ170]

So these xeons won’t work with my mobo?

"not working"
correct, these do not work, there may be others.

 

[RolloZ170]

I’m tempted to offer shipping you my board, CPU’s, and pay you in the 12 sticks of ram lol.

in am located in the EU(germany) - don't need RAM or CPU for this.
if you want to run 2nd Gen scalables you have to buy RDIMM ECC in any case, Cascade Lake don't support UDIMM.

 

[RolloZ170]

I cannot understand why a WS motherboard not enough power delivery to support all possible Xeons. This is unbelievable

ASUS WS C621 Sage supports all regular SKUs.
the VRM is strong enough. it is just the ICC_MAX value.
but indeed, all ASRocks don't have that incompatibility.

This is unbelievable. @RolloZ170 what would be the best lga 3647 or lga 4189 Workstation motherboard?

best ? can't recommend anything at the moment.
GB MD71 supports high TDC by default.
ASRock EP2C621 supports high TDC by default.

 

[RageBone]

Is the sage using the usual rebranded IR35201 ?

 

[RolloZ170]

Is the sage using the usual rebranded IR35201 ?

of course not. it is the older (but imho still best) Primarion PXE1610C (Primarion is overtaken to Infineon)

 

[oneplane]

Looks like you can switch the PXE1610C to page 0 using the default linux HWMon driver, but that doesn't mention anything about writing to the registers or updating the control program. PowerCode does have definitions for this chip, but I'm not sure if it can read back what's currently stored, it seems to only want to push entire programs.

 

[RolloZ170]

Looks like you can switch the PXE1610C to page 0 using the default linux HWMon driver

yes. the PCH can access the VRM controller. the problem is we need to take the Bus for several transfers,
if any other Busmaster does writes, we fail.
i do same with the MCP2221a device, its micro controller is faster and follows the I2C bus rules.

 

[oneplane]

yes. the PCH can access the VRM controller. the problem is we need to take the Bus for several transfers,
if any other Busmaster does writes, we fail.
i do same with the MCP2221a device, its micro controller is faster and follows the I2C bus rules.

Maybe the PSU PMBus header is wired to the same physical bus as the PXE? Otherwise, chip extraction (or VCC isolation) might be the only reasonable way if it's not the same circuit or tries to power the entire board

 

[RolloZ170]

PowerCode does have definitions for this chip, but I'm not sure if it can read back what's currently stored, it seems to only want to push entire programs.

PowerCode supports the design of the whole VRM design. (note this is Dell 7820 with 6+1 VRM config)
PC is made for PXE1610C (amd other Primarion VR controller)

 

[RolloZ170]

Maybe the PSU PMBus header is wired to the same physical bus as the PXE?

maybe. don't know.

Otherwise, chip extraction (or VCC isolation) might be the only reasonable way if it's not the same circuit or tries to power the entire board

you connect Ground, SDA and SCL. the MCP2221a is powered by USB.

 

[RolloZ170]

modify the PXE1610C is easy now(without PowerCode). but we need to connect without soldering to keep warranty alive thought.