多卡训练报错，有时报，有时不报：

--------------------------------------
C++ Traceback (most recent call last):
--------------------------------------
0 paddle::framework::SignalHandle(char const*, int)
1 paddle::platform::GetCurrentTraceBackString[abi:cxx11]()

----------------------
Error Message Summary:
----------------------
FatalError: `Segmentation fault` is detected by the operating system.
[TimeInfo: *** Aborted at 1624505994 (unix time) try "date -d @1624505994" if you are using GNU date ***]
[SignalInfo: *** SIGSEGV (@0x1378) received by PID 516 (TID 0x7fc68c1de700) from PID 4984 ***]

/mnt
[INFO]: train job failed! train_ret: 1