Hello Peter,

On May 26 00:20 96, Peter Fitzsimmons of 1:259/414 wrote:

 PF> I don't think the 61440 will ever be backed by physical 
 PF> ram.  I think they are refering to _address space_,  as I 

that was my understanding too. (and thats what i meant in my original
message). but eating 60K for a 4K alloc (and one my think that it is a good
idea too allocate in 4K pages and use suballocation mechanisms on this 4K)
...

 PF> 512mb address space,  but it is not nearly as bad has 
 PF> wasting physical memroy).

whats about MEMMAN=COMMIT? will the memory be back-up'ed by swap-space? i think so.

 PF> I wonder why they made the change?

I wonder IF they made the chance. I wrote a small programm, allocating
1byte R/W nonCommitted. some people compared the OS2MEMO values (allocated,
committed, ...) and they are not different between WARP and WARP Connect.

(i assume that WARP connect means also WARP with a fixpack >= 5. I run
WARP FP5 and it still allocates 64K). 
 WARPED!, Mario

--- Msgedsq/2 2.2e