Installing the Firmware (*.rpm)
If an NPU 1.X.X package is used for installation, the firmware package is integrated into the driver package. During the driver installation or upgrade, the firmware is automatically installed or upgraded. For details, see Installing the Driver (*.rpm).
If NPU 20.X.X or later packages are used, install the driver package and then the firmware package. For details about how to install the driver, see Installing the Driver (*.rpm). This section uses the A300-3000-npu-firmware_<version>.rpm firmware package as an example to describe how to install the .rpm firmware package on the host. Replace the package name with the actual one.
Precautions
- The .rpm package applies only to EulerOS scenarios. Do not use it in other environments.
- The installation path of the .rpm package cannot be changed. You need to install the .rpm package according to the normal process. The --force command of the .rpm package cannot be used for installation and upgrade.
- The .rpm package and .run package are at the same level and are incompatible. That is, the .rpm package cannot be installed after the .run package is installed, or the .run package cannot be installed after the .rpm package is installed.
- Some internal functions of the Ascend 310 can be accessed and executed only by the HwHiAiUser user. The HwHiAiUser user is automatically created upon the first installation.
- The installation takes effect after a restart. If you need to reinstall the driver after the uninstallation, restart the system before the installation.
- If the .rpm package of NPU 1.X.X has been installed, uninstall it before installing the .rpm package of NPU 20.X.X or later.
- During the installation or upgrade of a software package, do not reset or power off the host or device. Otherwise, device boot or upgrade fails. If this problem occurs, resolve it by referring to Failed to Start the Device Due to Reset or Power-off.
- During capacity expansion or spare part replacement, run the npu-smi info -t board -i id command to query the version number. You are advised to use the same firmware version for all Atlas 300I inference cards on the server. If the firmware versions are different, upgrade the firmware first. Otherwise, the server may fail to identify the Atlas 300I inference cards, or the Atlas 300I inference cards are lost during service running. For details, see A Server Fails to Identify an Atlas 300I Inference Card Due to Inconsistent Firmware Versions.
Procedure
- Upload the driver package obtained from Obtaining Software Packages to any directory (for example, /opt) in Linux.
- Use PuTTY to log in to the OS CLI of the server where the Atlas 300I inference card (model 3000) is located. For details, see Logging In to the CLI Using PuTTY over a Network Port.
- Run the following command to switch to the root user:
su - root
- Run the following command in the directory where the software package is stored to install the firmware package on the host:
rpm -ivh A300-3000-npu-firmware_<version>.rpm
If the HwHiAiUser user does not exist during the installation, the system automatically creates it.
- Run the reboot command to restart the host OS for the firmware version to take effect.
- Run the d-agent -v command to check whether the firmware is successfully installed.
If information similar to the following is displayed, the installation is successful. Otherwise, the installation fails. Contact Huawei technical support.
driver: 20.1.0 firmware: 1.76.T20.0.B200
The logs generated during the installation are recorded in the /var/log/ascend_seclog/ascend_install.log file. You can run the vim /var/log/ascend_seclog/ascend_install.log command to open the log file.
- Upgrade the chip firmware of the Atlas 300I inference card (model 3000) after the firmware is successfully installed.
- Run the d-agent -b 0000:91:00.0 command to upgrade the firmware of a single chip.
- If no service process is running on all Atlas 300I inference card (model 3000) chips, run the d-agent -a command to upgrade the firmware of all chips.
- You can run the lspci | grep d100 command to view the chip ID of the Atlas 300I inference card (model 3000).
- Upgrading the firmware of a single chip takes about 30s. The firmware upgrade duration of all chips is subject to the actual situation. You can query the firmware version after the firmware upgrade is complete.
Run the d-agent -g 0000:91:00.0 command to check the firmware version after upgrade.If the version is the same as the target version, the upgrade is successful. Otherwise, the upgrade fails. Contact Huawei technical support.
nve:1.76.20.0.B200 xloader:1.76.20.0.B200