Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Cache Working Set Reduction by Stack Cache Management

IP.com Disclosure Number: IPCOM000123732D
Original Publication Date: 1999-Apr-01
Included in the Prior Art Database: 2005-Apr-05
Document File: 1 page(s) / 27K

Publishing Venue

IBM

Related People

Olszewski, B: AUTHOR [+3]

Abstract

For most commercial server computing, the performance of computer hardware is highly dependent on the ability of the system caches to contain the workload. A significant fraction of cache accesses are due to data manipulated on the program's stack. These accesses to data cause operations to main memory when the data is not present in the cache. The invention is to have the operating system purge data from the cache when a process is to be undispatched for a significant period of time (i.e. a disk I/O). If the hardware has an instruction that "pushes" modified data from the cache to memory and removes unmodified data from the cache and this action makes the now unused cache entries favored for the next cache use, the operating system reduce cache misses.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 100% of the total text.

Cache Working Set Reduction by Stack Cache Management

   For most commercial server computing, the performance of
computer hardware is highly dependent on the ability of the system
caches to contain the workload.  A significant fraction of cache
accesses are due to data manipulated on the program's stack.  These
accesses to data cause operations to main memory when the data is not
present in the cache.  The invention is to have the operating system
purge data from the cache when a process is to be undispatched for a
significant period of time (i.e. a disk I/O).  If the hardware has an
instruction that "pushes" modified data from the cache to memory and
removes unmodified data from the cache and this action makes the now
unused cache entries favored for the next cache use, the operating
system reduce cache misses.  Since the stack is a contiguous range
of addresses, it is the most obvious target of the cache purging
process.  For systems with both kernel and process stacks, the
operation is applicable to either or both.