US5357623A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Dynamic cache partitioning by modified steepest descent
Assignee: International Business Machines Corp
•
Filing: 1991-11-01
•
Granted: 1994-10-18
•
Expires: 2011-11-01
Abstract
Abstract Dynamic partitioning of cache storage into a plurality of local caches for respective classes of competing processes is performed by a step of dynamically determining adjustments to the cache partitioning using a steepest descent method. A modified steepest descent method allows unpredictable local cache activities prior to cache repartitioning to be taken into account to avoid readjustments which would result in unacceptably small or, even worse, negative cache sizes in cases where a local cache is extremely underutilized. The method presupposes a unimodal distribution of cache misses.
Inventors
- 👤 Igal Megory-Cohen
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-29 19:11
CPC Classifications
G; G06; G06F; G06F12/00; G06F12/02; G06F12/08; G06F12/0802; G06F12/0866; G06F12/0806; G06F12/0842; G06F2212/00; G06F2212/31; G06F2212/311
Claims Analysis
Total Claims: 4
Claim 1
Independent Claim
202 words
In a computer system comprising a CPU, a storage subsystem attaching said CPU, and a cache intercoupling said CPU and subsystem, said cache being initially partitioned into a plurality of local caches, each local cache i being of size Lc i and being subject to a miss ratio of LM i , a method for dynamically adjusting the partition of said local caches referenced by counterpart classes of CPU executed processes, comprising the iterative steps of (a) tracking, for each local cache (i) of current size c i , the number of current references (f i ) to said local cache and the number of current references (m i ) which result in a cache miss, the number of cache misses m i from each of the tracked local caches being distributed over a unimodal surface; and (b) adjusting local cache sizes c i for a weighted miss ratio function (w=ΣF i *M i (c i )) as a vector in the direction of the steepest descent projected on a functional hyperplane representative of a constrained sum of individual cache sizes intersecting the unimodal surface, where F i =f i /Σf i , M i (c i )=m i /f i .
Claim 2
Dependent Claim
202 words
In a computer system comprising a CPU, a storage subsystem attaching said CPU, and a cache intercoupling said CPU and subsystem, said cache being initially partitioned into a plurality of local caches, each local cache i being of size Lc i and being subject to a miss ratio of LM i , a method for dynamically adjusting the partition of said local caches referenced by counterpart classes of CPU executed processes, comprising the iterative steps of (a) tracking, for each local cache (i) of current size c i , the number of current references (f i ) to said local cache and the number of current references (m i ) which result in a cache miss, the number of cache misses m i from each of the tracked local caches being distributed over a unimodal surface; and (b) adjusting local cache sizes c i for a weighted miss ratio function (w=ΣF i *M i (c i )) as a vector in the direction of the steepest descent projected on a functional hyperplane representative of a constrained sum of individual cache sizes intersecting the unimodal surface, where F i =f i /Σf i , M i (c i )=m i /f i .
Claim 3
Dependent Claim
82 words
The method according to claim 1, wherein step (a) includes the sub-steps of (a1) maintaining a control block for each local cache (i), said control block having a first data field for storing f i and a second data field for storing m i ; and (a2) incrementing the content of the first data field on each reference to that local cache and the content of the second data field being incremented each time such a reference results in a cache miss.
Claim 4
Dependent Claim
153 words
The method according to claim 1, wherein step (b) includes the sub-steps of (b1) determining which local caches have sizes c i which are not smaller than respective predetermined sizes (t i ); (b2) for the sub-set of local caches determined in step (b1) only, determining, for a weighted miss ratio function (w=ΣF i *M i (c i )) subject to the constraint that the sum of the individual cache sizes (Σc j , j ranges over the subset) equals the total of the sizes (Σc i ) of local caches of the sub-set, a vector in the direction of the steepest descent projected on a functional hyperplane representative of said constraint; and (b3) terminating step (b) if the cache sizes c j of the sub-set of local caches are not smaller than respective predetermined sizes t i , otherwise repeat steps (b1) to (b3) recursively for a new sub-set of local caches.
References Cited (Backward Citations)
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1997-10-17 | 1998-04-18 | 1998-10-18 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2001-10-17 | 2002-04-18 | 2002-10-18 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2005-10-17 | 2006-04-18 | 2006-10-18 | $7700.00 | Window Closed / Lapsed |