那位了解这个系统错误？之一。

标红的部分我也出现了。文中提到的22.5k 是指22500个proc 过程吗？

怎样解决这个问题呢?是不是只有拆程序了。我有个proc mean可以用计数的data 步代替就会解决这个问题呢？我差不多要运行67*243=16281次这个proc。

Paulo:
The


ERROR:  Read Access Violation In Task ( DMS Process ]
Exception occurred at (676A64FC)
Task Traceback
ERROR: Generic critical error.

usually means that a process cannot allocate enough memory to finish a task.
Running each regression as a separate batch process clears memory in between
process.

While batch processing may make estimating >22.5K regression models
possible, I have to wonder why you would need to do that. It sounds similar
to a regression model with a supposedly independent categorical variable of
22.5K states. What one would do with 22.5 regression lines (planes,
hyperplanes) eludes me. As I have said in another context, time and
resources constraints in today's computing environment should raise
questions about whether the benefit of information and knowledge that you
are obtaining from data (and, in this case, estimates from data) offset the
processing and storage costs.
Sig

-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Paulo
Alves
Sent: Monday, January 31, 2005 1:46 PM
To: SAS-L@LISTSERV.UGA.EDU
Subject: SAS keeps crashing...


Hi.

I have to run several sets of regressions, each set 1500 companies and 5
years = 7.500 regressions...

And SAS always crashes after I run it sometimes just stops, others gives the
following error:

ERROR:  Read Access Violation In Task ( DMS Process ]
Exception occurred at (676A64FC)
Task Traceback
ERROR: Generic critical error.

Any help?


Cheers,

PA

果然是含金量很高的问题

好无聊啊,回归那么多次,多想想有没有实际意义

batch模式试试

我真想看看你那数据是啥数据。
让人很怀疑你做这个回归的意义。

我没做回归啊！把我的问题仔细看看

其实把所有log output 关掉就可以了。这个问题我已经解决，也不需要改程序。