re : swap cache

Jason Chaw (jchaw@acm.org)
Mon, 21 Dec 1998 22:49:16 +0800

Today all workstations and servers have alot of memory, diskspace, bandwidth
(memory bus, disk I/O bus, etc). Hence the efficacy of compressed virtual
memory(VM) is no longer realised. As Rik pointed out, the main problem is
drive seeks. We should minimize on drive seeks. To merely compress the pages
of virtual memory is of little use. Disk reads / writes will not be
optimized.

For systems of low memory, The situation will be worse performance as the
limited memory resources will be further strained to accomodate the data
structures of the compression algorithm. Likewise for CPU resources.
The idea is to implement a VM which is optimized for swap in/out of the swap
partition. This will require simple and efficient algorithms.

We have to understand why we use VM. The system is running on borrowed time.

[JASON@CHAW]

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/