topic Re: AUX/Main EFUSE Regulator 1 (10h)) Tesla GPU in ProLiant Servers (ML,DL,SL)

AUX/Main EFUSE Regulator 1 (10h)) Tesla GPU

Gflex78 — Fri, 17 Mar 2023 09:42:03 GMT

Hi,

We have one M3G77A - HPE DL380 Gen9 E5-2620v3 SP8040TV EU Svr with 2x CPU

We have one 754383-001 1400 Watts PSU with firmware 1.10

We have 2 x 719076-B21 HP DL380 G9 Primary 2 Slot GPU Ready Riser Kit installed

We installed HP nVidia Tesla P100 16GB PCIe 3.0 x16 Q0E21A in the GPU riser

We installed the 719082-B21 HP DL380 GEN9 GRAPHICS ENABLEMENT KIT

We plugged in 1 HP DL380 Gen9/8 G9/8 GPU Power Cable 25cm 1x 10-pin -- 1x 8-pin 803403-001 in one PCI riser and one in the other. And we plugged in the other side of the cable into the cable in the picture (don’t know the part number) and then into the GPU.

When booting the server we receive the error message

Server Critical Fault (Service Information: Power On Fault, System Board, AUX/Main EFUSE Regulator 1 (10h))

We also tried using the HP 60cm / 70cm GPU Power Cable DL380 DL560 DL580 G8 G9 10-Pin to 2x 6-Pin + 2x 2-Pin but we still receive the error message when we boot the server.

Unfortunately, the server is out of warranty so we cannot create a support case.

What are we doing wrong?

Does anyone know how to make the Tesla P100 work with the server above?

There is really not a lot of documentation on how to do this.

Re: AUX/Main EFUSE Regulator 1 (10h)) Tesla GPU

Sebasbin — Fri, 17 Mar 2023 09:39:54 GMT

Hi,

->I would suggest to configure the thermal settings to "maximum cooling " in BIOS and verify -https://support.hpe.com/hpesc/public/docDisplay?docId=sf000046358en_us&docLocale=en_US

->Also test the server without installing the NVIDIA Tesla P100 GPU and verify if the Server Critical Fault error is still reporting .

->If the error is reported without installing the NVIDIA Tesla P100,this may be the issue is with system board .

->If the issue still persists then suggesting logging a support case is with HPE for further investigation including log analysis

Re: AUX/Main EFUSE Regulator 1 (10h)) Tesla GPU

Gflex78 — Mon, 20 Mar 2023 05:58:56 GMT

@Sebasbin

Hi,

There are no errors without the Tesla card

The cooling is set to max.

Is there a way to log a case without support, post warranty?

Is there a way to read the SUM log files?