Rate and give feedback:
Huawei uses machine translation combined with human proofreading to translate this document to different languages in order to help you better understand the content of this document.
Note: Even the most advanced machine translation cannot match the quality of professional translators.
Huawei shall not bear any responsibility for translation accuracy and it is recommended that you refer to the English document (a link for which has been provided).
Hot Reset
Function
The npu-smi set -t reset -i id -c chip_id [–m 1] command is used to reset a specific chip.
After the NPU is reset successfully, wait for at least 20s before querying information about this chip or other chips.
Syntax
npu-smi set -t reset -i id -c chip_id [–m 1]
Parameters
Parameter |
Description |
|---|---|
id |
Device ID. The NPU ID obtained by running the npu-smi info -l command is the device ID. |
chip_id |
Chip ID. If there is only one chip, the chip ID is 0. |
Restrictions
- Before performing a hot reset, all services must be stopped on the NPUs.
- The hot reset command must be run as the root user on a physical machine. If the IP address is configured as a non-root user on the physical machine, in a container, or on an VM, an error is reported.
- In the preceding command, [-m 1] indicates an optional parameter. If this parameter is used, the PCIe card is reset in in-band mode. If this parameter is not used, the PCIe card is reset in out-of-band mode. Resetting the PCIe card in out-of-band mode is only supported in 20.2.0 and later versions.
- If the is used together the Atlas 800 inference server (model 3010), in-band reset of the standard card can be used in 20.2.0.SPC100 and later versions.
Example
# Perform a hot reset on NPU 2 in in-band mode.
npu-smi set -t reset -i 2 -c 0 -m 1
Resetting a standard PCIe card or npu chip during service running may cause system hang or abnormal reset.
Are you sure you want to continue resetting?(Y/N)
n
Status : Fail
Message : User aborts reset.
npu-smi set -t reset -i 2 -c 0 -m 1
Resetting a standard PCIe card or npu chip during service running may cause system hang or abnormal reset.
Are you sure you want to continue resetting?(Y/N)
y
Message : resetting ...
Status : OK
Message : Reset chip successfully
# Perform a hot reset on NPU 2 in out-of-band mode.
npu-smi set -t reset -i 2 -c 0
Resetting a standard PCIe card or npu chip during service running may cause system hang or abnormal reset.
Are you sure you want to continue resetting?(Y/N)
n
Status : Fail
Message : User aborts reset.
npu-smi set -t reset -i 2 -c 0
Resetting a standard PCIe card or npu chip during service running may cause system hang or abnormal reset.
Are you sure you want to continue resetting?(Y/N)
y
device 2 is rebooting, it takes about 60s
Message : resetting ...
Status : OK
Message : Reset chip successfully