香橙派AI pro 8tops CANN 8.2.RC1 安装MIndSpore2.7.1、2.7.0、2.6.0失败
开发板是香橙派AI pro 8tops (16g)固件版本25.2.0,芯片310B4目前安装 mindspore 2.7.1、2.7.0、2.6.0都没有通过检测2.6.0的报错如下CANN 通过 `./Ascend-cann-kernels-310b_8.2.RC1_linux-aarch64.run –upgrade ` 从8.0.0更新到 8.2.RC1,也安装了 8.2.RC1 的 ke
·
问题描述
开发板是香橙派AI pro 8tops (16g)
固件版本25.2.0,芯片310B4
目前安装 mindspore 2.7.1、2.7.0、2.6.0都没有通过检测
2.7.0的报错如下:
(ms27) HwHiAiUser@orangepiaipro:~/WorkPlace/projects/mindspore$ python -c "import mindspore;mindspore.set_device('Ascend');mindspore.run_check()"
/usr/local/miniconda3/envs/ms27/lib/python3.9/site-packages/numpy/core/getlimits.py:549: UserWarning: The value of the smallest subnormal for <class 'nt64'> type is zero.
setattr(self, word, getattr(machar, word).flat[0])
/usr/local/miniconda3/envs/ms27/lib/python3.9/site-packages/numpy/core/getlimits.py:89: UserWarning: The value of the smallest subnormal for <class 'nu64'> type is zero.
return self._float_to_str(self.smallest_subnormal)
/usr/local/miniconda3/envs/ms27/lib/python3.9/site-packages/numpy/core/getlimits.py:549: UserWarning: The value of the smallest subnormal for <class 'nt32'> type is zero.
setattr(self, word, getattr(machar, word).flat[0])
/usr/local/miniconda3/envs/ms27/lib/python3.9/site-packages/numpy/core/getlimits.py:89: UserWarning: The value of the smallest subnormal for <class 'nu32'> type is zero.
return self._float_to_str(self.smallest_subnormal)
MindSpore version: 2.7.0
[WARNING] DEVICE(10053,dfffe7c87020,python):2025-11-24-16:03:42.964.367 [mindspore/ccsrc/plugin/res_manager/ascend/mem_manager/ascend_memory_adapter.cctialize] Free memory size is less than half of total memory size.Device 0 Device MOC total size:16368312320 Device MOC free size:7432048640 may be othees occupying this card, check as: ps -ef|grep python
MindSpore running check failed.
Hccl set deterministic mode failed! mode is 0 and error flag is 0
----------------------------------------------------
- C++ Call Stack: (For framework developers)
----------------------------------------------------
mindspore/ccsrc/plugin/res_manager/ascend/ascend_res_manager.cc:722 SetDeterministic
2.6.0的报错如下
(ms27) HwHiAiUser@orangepiaipro:~/WorkPlace/projects/mindspore$ python -c "import mindspore;mindspore.set_device('Ascend');mindspore.run_check()"
[WARNING] ME(10204:246289795915808,MainProcess):2025-11-24-16:10:07.180.373 [mindspore/run_check/_check_version.py:324] MindSpore version 2.6.0 and Ascftware package (Ascend Data Center Solution)version 8.2 does not match, the version of software package expect one of ['7.6', '7.7']. Please refer to tinfo on: https://www.mindspore.cn/install
/usr/local/miniconda3/envs/ms27/lib/python3.9/site-packages/numpy/core/getlimits.py:549: UserWarning: The value of the smallest subnormal for <class 'nt64'> type is zero.
setattr(self, word, getattr(machar, word).flat[0])
/usr/local/miniconda3/envs/ms27/lib/python3.9/site-packages/numpy/core/getlimits.py:89: UserWarning: The value of the smallest subnormal for <class 'nu64'> type is zero.
return self._float_to_str(self.smallest_subnormal)
/usr/local/miniconda3/envs/ms27/lib/python3.9/site-packages/numpy/core/getlimits.py:549: UserWarning: The value of the smallest subnormal for <class 'nt32'> type is zero.
setattr(self, word, getattr(machar, word).flat[0])
/usr/local/miniconda3/envs/ms27/lib/python3.9/site-packages/numpy/core/getlimits.py:89: UserWarning: The value of the smallest subnormal for <class 'nu32'> type is zero.
return self._float_to_str(self.smallest_subnormal)
[WARNING] ME(10204:246289795915808,MainProcess):2025-11-24-16:10:12.157.771 [mindspore/run_check/_check_version.py:342] MindSpore version 2.6.0 and "teackage version 8.2 does not match. For details, refer to the installation guidelines: https://www.mindspore.cn/install
[WARNING] ME(10204:246289795915808,MainProcess):2025-11-24-16:10:12.158.098 [mindspore/run_check/_check_version.py:355] Please pay attention to the abog, countdown: 3
[WARNING] ME(10204:246289795915808,MainProcess):2025-11-24-16:10:13.159.352 [mindspore/run_check/_check_version.py:355] Please pay attention to the abog, countdown: 2
[WARNING] ME(10204:246289795915808,MainProcess):2025-11-24-16:10:14.160.495 [mindspore/run_check/_check_version.py:355] Please pay attention to the abog, countdown: 1
[WARNING] ME(10204:246289795915808,MainProcess):2025-11-24-16:10:18.802.366 [mindspore/run_check/_check_version.py:324] MindSpore version 2.6.0 and Ascftware package (Ascend Data Center Solution)version 8.2 does not match, the version of software package expect one of ['7.6', '7.7']. Please refer to tinfo on: https://www.mindspore.cn/install
[WARNING] ME(10204:246289795915808,MainProcess):2025-11-24-16:10:18.803.813 [mindspore/run_check/_check_version.py:342] MindSpore version 2.6.0 and "teackage version 8.2 does not match. For details, refer to the installation guidelines: https://www.mindspore.cn/install
[WARNING] ME(10204:246289795915808,MainProcess):2025-11-24-16:10:18.804.012 [mindspore/run_check/_check_version.py:355] Please pay attention to the abog, countdown: 3
[WARNING] ME(10204:246289795915808,MainProcess):2025-11-24-16:10:19.804.488 [mindspore/run_check/_check_version.py:355] Please pay attention to the abog, countdown: 2
[WARNING] ME(10204:246289795915808,MainProcess):2025-11-24-16:10:20.805.851 [mindspore/run_check/_check_version.py:355] Please pay attention to the abog, countdown: 1
MindSpore version: 2.6.0
[WARNING] DEVICE(10204,dfffcfcc2020,python):2025-11-24-16:10:21.992.706 [mindspore/ccsrc/plugin/res_manager/ascend/mem_manager/ascend_memory_adapter.cctialize] Free memory size is less than half of total memory size.Device 0 Device MOC total size:16368312320 Device MOC free size:6867746816 may be othees occupying this card, check as: ps -ef|grep python
MindSpore running check failed.
Hccl set deterministic mode failed! mode is 0 and error flag is 0
----------------------------------------------------
- C++ Call Stack: (For framework developers)
----------------------------------------------------
mindspore/ccsrc/plugin/res_manager/ascend/ascend_res_manager.cc:521 SetDeterministic
CANN 通过 `./Ascend-cann-kernels-310b_8.2.RC1_linux-aarch64.run –upgrade ` 从8.0.0更新到 8.2.RC1,也安装了 8.2.RC1 的 kernels
run_check() 2.7.1 和 2.7.0时都报错:
Hccl set deterministic mode failed! mode is 0 and error flag is 0
问题解答
CANN换成 8.1.RC1就OK了
昇腾计算产业是基于昇腾系列(HUAWEI Ascend)处理器和基础软件构建的全栈 AI计算基础设施、行业应用及服务,https://devpress.csdn.net/organization/setting/general/146749包括昇腾系列处理器、系列硬件、CANN、AI计算框架、应用使能、开发工具链、管理运维工具、行业应用及服务等全产业链
更多推荐

所有评论(0)