Duplicate superDeviceId (EI0014)

Symptom

The CANN logs contain the keyword "superDeviceId[***] in superPod[***]is already exist," as shown below:

[ERROR] HCCL(169030,alltoall_test):2025-10-23-16:28:59.392.635 [topoinfo_exchange_agent.cc:695] [169030][InitGroupStage][RanktableCheck]devices have same superDeviceId[0x3000000] in superPod[super_pod_id_0]. Current device info: serverId[127.10.0.1], rankId[0], group[hccl_world_group]. Another device info: rankId[1].

Possible Cause

superDeviceId indicates the physical ID of a device on the Atlas A3 training product/Atlas A3 inference product in the SuperPoD system, which is the unique. HCCL detects that devices with the same superDeviceId exist on a SuperPoD during consistency verification. As a result, the verification fails. You can run the npu-smi command to query the value of superDeviceId.

npu-smi info -t spod-info -i id -c chip_id
  • id: device ID, which is the NPU ID obtained by running the npu-smi info -l command.
  • chip_id: chip ID, which is obtained by running the npu-smi info -m command.

In the command output, SDID indicates the value of superDeviceId.

Possible causes are as follows:
  • The hardware configuration is incorrect.
  • The logical SuperPoD is configured with several physical SuperPoDs using the HCCL_LOGIC_SUPERPOD_ID environment variable, causing duplicate superDeviceId.

Solution

Modify the hardware configuration or correct the HCCL_LOGIC_SUPERPOD_ID environment variable to avoid duplicate superDeviceId of devices on a SuperPoD.