Re: Workspace size

[John Darrington]

On Fri, Jul 25, 2008 at 09:35:58PM -0700, Ben Pfaff wrote:
     > 2.  My interpretation of how the Chicago company treats this setting
     >     is different from what we use it for.   If we're going to emulate
     >     their semantics, then it ought to do something like:
     >
     >     struct rlimit limit = {x, RLIM_INFINITY};
     >     setrlimit (RLIMIT_AS, &limit);
     >
     >     whereas we currently use it to set an upper limit on the number of
     >     cases stored in memory, which is also useful, but perhaps we
     >     should change the name of the current setting?
     
     I don't understand your interpretation of SET WORKSPACE.  The v16
     manual says: 
     
         WORKSPACE allocates more memory for some procedures when you
         receive a message indicating that the available memory has
         been used up or indicating that only a given number of
         variables can be processed... WORKSPACE allocates workspace
         memory in kilobytes for some procedures that allocate only
         one block of memory.
     
     Isn't limiting the number of cases stored in memory a reasonable
     approximation of that behavior?

Hmm. I thought I remembered it described somewhat differently.  Maybe
I was thinking of MXMEMORY.

Also on the subject of WORKSPACE, what do people think about the
default value?  Currently PSPP uses 4096KB, the v14 manual says 6148
is their default.    I think that both these are far too low for a
modern machine. 64MB would be more reasonable IMO.

J'

-- 
PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See http://pgp.mit.edu or any PGP keyserver for public key.

signature.asc
Description: Digital signature