Atlas 300I(3000) Chip Query Error

Hi,

I have some problem related to Atlas 300I(Model 3000) card. I completed the first installation step including OS dependency, NPU driver, NPU firmware etc. and I passed the every step without faced any problem but at the end of the installation I faced that I can see only 1 Ascend 310 Chip with “npu-smi info” command but acording to whitepapera, the card has 4 Ascend 310 Chip. It is prety strange.

Here are some queries and their outputs as follows:

• Server Model = TaiShan 2280

• OS Info = Ubuntu-18.04.6 LTS

• Atlas 300I Card Info = Atlas300I - Model 3000

• NPU Driver Vesion = 21.0.2

• NPU Firmware Version = 1.78.23.33.230

• NPU Driver&Firmware Download Link = https://support.huawei.com/enterprise/zh/software/253276357-ESW2000387090

• lspci | grep d100 query output =

000a:11:00.0 Processing accelerators: Huawei Technologies Co., Ltd. Device d100 (rev 20)

• npu-healthcheck.sh 31 query output =

miniD 1 to MCU iic channel status: Fault miniD 3 to MCU iic channel status: OK MCU to miniD 0 iic channel status: OK MCU to miniD 1 iic channel status: OK MCU to miniD 2 iic channel status: OK MCU to miniD 3 iic channel status: OK slot6_card6 health state is Fault Product Type                   : Atlas 300I Model 3000 2021-12-16 09:39:20 slot6-chip0 query slot6-chip0 pci_switch failed 8

• npu-smi info query output =

As you see, I can see only 1 Ascend 310 Chip. I searched this error on "ascend/modelzoo gitee" page, "bbs.huaweicloud" and "Ascend Developer Zone" group but I didn't find any solution related to this.

So, what can I do at this stage? I look forward to your help in this issue.
If you have any idea to figure it out this problem, I would be happy to hear.

Have a nice day.

Best,

Kubilay