Merge pull request !483 from 冯学峰/myr0.601

Merge pull request !479 from 夏易凡/r0.6

Merge pull request !480 from LiHongzhang/fix_unpack

    three-dimensional to two-dimensional when autoupdate is open

Merge pull request !467 from LiHongzhang/bugfixes

fix npu timeout mechanism

sum not used and not available to others

add new field of success

Merge pull request !469 from 冯学峰/myr0.601

Merge pull request !472 from 黄伟锋/r0.6

Merge pull request !474 from 潘慧/r0.6

Merge pull request !476 from luopengting/r0.6

1. New a thread to load detail info. Loading detail info takes too much time,
so the summary list and lineage can not be loaded timely.
2. Add a status for DetailCacheManager to indicate it is INIT, LOADING or DONE.
3. Update UT/ST.

fix hardware issue: 1. Hard to distinguish between specific states for health and available;2.Cannot uncheck after selecting a single cpu at U;3.its better not display ascend usage in hardware resource for gpu env

Merge pull request !465 from 夏易凡/0710master

Merge pull request !463 from 黄伟锋/r0.6

Merge pull request !461 from LiHongzhang/npu_timeout

Merge pull request !454 from LiHongzhang/fix_timeout

Merge pull request !453 from LiHongzhang/v060

!451 use forkserver multiprocess context to avoid forking child process with locked stream resource(eg stdout)
Merge pull request !451 from wenkai/wk5_fix_log_stream_flush_stuck

Merge pull request !450 from LiHongzhang/no_readable

use forkserver multiprocess context to avoid forking child process with locked stream resource(eg stdout)

Merge pull request !446 from luopengting/fix_lineage

!442 convert the object of `NAN`、｀-INF｀and ｀INF｀ to string so that browser can view these tensor data
Merge pull request !442 from wangshuide/wsd_handle_special_tensor

Merge pull request !447 from wangshuide/wsd_limit_tensor_count

Merge pull request !441 from yuximiao/yuximiao_profiler

Merge pull request !445 from wenkai/wk4_0721_ci_pickle_error_fix

!443 Optimize number of max processes used for computing and limit number of thread running load_data_in_thread()
Merge pull request !443 from wenkai/perf_opt2_0721

Optimize number of max processes used for computing and limit number of thread running load_data_in_thread()

For max processes used for computing, we need to make sure every summary directory has a process to load data. We also should not use too many processes to avoid system problems (eg. out of memory). So we calc the max processes cnt in _calc_default_max_processes_cnt

For _load_data_in_thread_wrapper, because self._load_data_in_thread() will create process pool when loading files, we can not afford to run multiple self._load_data_in_thread() simultaneously. So we use a lock to make sure that only one self._load_data_in_thread() is running.

!440 Change the mindinsight multiprocessing computing code to use a unified manager, add new  features
Merge pull request !440 from wenkai/pref_opt_0720_1cp1