[求助] 如何设定GAUSS可用的RAM?

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

各位GAUSS前辈，

我现用GAUSS6.0，我的RAM从512M增到4G,处理数据时都被说‘no sufficient memory'。

而根据被储存的经过处理的数据文件来看，RAM用的可能不到400M.

请问那里/怎样可以设定GAUSS可用的RAM?

请帮忙！！！

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏0 回帖

关键词：GAUSS Aus RAM USS sufficient 求助 GAUSS RAM

我刚刚在网上找到下面关于GAUSS RAM的说法，如果真是这样是不是表示我除了改CODE没别的可想了？

2.3 Memory management

In the early days of GAUSS, efficient memory management was often crucial to getting a program running well. The amount of memory used by GAUSS could be varied by the user to make appropriate use of scarce resources. However, this is much less of an issue in modern computers, and from version 4.0 for Windows, GAUSS no longer gives you the option to manage memory directly. Instead it relies on the more efficient memory-management facilities of the operating system.

This does not mean you may ignore the issue of effective programming skills. It is suprisingly easy to run out of memory when doing complex operations on large matrices. For a more detailed discussion see the section on code refinements.

请看官方 User Guide；版本也有关。

以下是引用gemini69在2008-7-22 2:34:00的发言：
请看官方 User Guide；版本也有关。

Thank you but could you give a bit more details?

So far I have looked into file gauss.cfg and another source file system.src.

In gauss.cfg I didnt find any parameters related to memory size.

In system.src where gauss manages the variable 'maxvector', I changed it to 10 billion and didnt get much improvements.

Could you tell me exactly how can I do it?

Thanks!

以下是引用gemini69在2008-7-22 2:34:00的发言：
请看官方 User Guide；版本也有关。

I have tried changing the parameters in two files:

system.dec and system.src

for the max elements that can be allowed in one matrix, even I set it to 1 billion, the result of my code does not change.

It still stopped at the same stage and reported 'no sufficient memory'.

Could anybody just tell me how should I deal with it or it is impossible at all unless I change my code?

Thanks in advance!

Well, you are in the right way to solve the memory usage problem.
It is, however, not the only way to figure that out, according to the official.

Good luck

From official FAQ,
----------------------------------------------------------------------------------

Q：   I have large programs that take a long time to run--sometimes they run out of memory unexpectedly. What can I do?

Ans：   We suggest Windows NT, 2000 or XP for running long jobs that use a lot of memory.

Q：   My problem has large matrices, how large can they be?

Ans：   GAUSS uses all available memory in the computer and thus the primary limitation is the amount of RAM. Another limitation is the maximum addressable size on the computer. A 32-bit computer can maximally address about 2.2 gigabytes. The limitation in GAUSS in this case would be about one-sixth of 2.2 gigabytes, or about 360 megabytes because of temporary copies of matrices made during operations within a statement. A single element of a matrix is a double precision number taking 8 bytes. Thus a computer with more than 2.2 gigabytes can handle matrices of about 6000 by 6000. When matrices are that large in a problem, there could be serious slowdowns. The time involved in a matrix multiply, for example, is a cubic function of the size of the matrix. A single matrix multiply involving a 6000 by 6000 matrix would take about 30 hours to compute on one of today's PCs.

Usually only datasets come in sizes that large, and efficient code can be written to handle datasets larger than even 2.2 gigabytes by leaving the data on the disk and using special functions for accessing the data on the disk. Efficient methods will also be needed for computations. For example, rather than compute a moment matrix with a matrix multiply use GAUSS's moment function which computes only the lower left portion reducing the problem to a quadratic from a cubic function of size. And there is also a moment function, called moment, that can compute a moment matrix from a GAUSS dataset on the disk.