POWER CONSERVATION IN VERTICALLY-STRIPED NUCA CACHES
First Claim
1. A method, comprising:
- enabling the operation of a plurality of banks of a non-uniform cache access (NUCA) cache, wherein ways of the cache are vertically distributed across multiple banks of the plurality;
disabling, sequentially, individual banks of the plurality to conserve power of the NUCA cache, wherein the sequential disabling comprises firstly turning off individual banks with the greatest access latencies of banks of the NUCA cache and lastly turning off individual banks with the least access latencies, wherein further the disabling of the individual banks comprises turning off banks of the plurality grouped via discrete power states.
1 Assignment
0 Petitions
Accused Products
Abstract
Embodiments that dynamically conserve power in non-uniform cache access (NUCA) caches are contemplated. Various embodiments comprise a computing device, having one or more processors coupled with one or more NUCA cache elements. The NUCA cache elements may comprise one or more banks of cache memory, wherein ways of the cache are vertically distributed across multiple banks. To conserve power, the computing devices generally turn off groups of banks, in a sequential manner according to different power states, based on the access latencies of the banks. The computing devices may first turn off groups having the greatest access latencies. The computing devices may conserve additional power by turning of more groups of banks according to different power states, continuing to turn off groups with larger access latencies before turning off groups with the smaller access latencies.
-
Citations
20 Claims
-
1. A method, comprising:
-
enabling the operation of a plurality of banks of a non-uniform cache access (NUCA) cache, wherein ways of the cache are vertically distributed across multiple banks of the plurality; disabling, sequentially, individual banks of the plurality to conserve power of the NUCA cache, wherein the sequential disabling comprises firstly turning off individual banks with the greatest access latencies of banks of the NUCA cache and lastly turning off individual banks with the least access latencies, wherein further the disabling of the individual banks comprises turning off banks of the plurality grouped via discrete power states. - View Dependent Claims (2, 3, 4, 5, 6, 7)
-
-
8. An apparatus, comprising:
-
a plurality of banks of a non-uniform cache access (NUCA) cache, wherein ways are vertically distributed across multiple banks of the plurality; a plurality of switches configured to turn off groups of banks of the plurality of banks, wherein banks are aggregated to the groups based on access latencies; and a state selector, coupled to the plurality of switches, to select different power states of the NUCA cache, wherein the state selector is arranged to turn off groups of banks with the larger access latencies before turning off other groups with smaller access latencies. - View Dependent Claims (9, 10, 11, 12, 13, 14, 15)
-
-
16. A computer program product comprising a tangible computer readable storage medium including instructions that, when executed by at least one processor:
-
search ways of a plurality of banks of a non-uniform cache access (NUCA) cache, wherein the ways are vertically distributed across multiple banks of the NUCA cache; and sequentially turn off groups of banks of the plurality of banks, wherein each of the groups comprises banks aggregated based on access latencies of banks to the at least one processor, wherein further the sequence comprises turning off groups with larger access latencies before turning off groups with smaller access latencies. - View Dependent Claims (17, 18, 19, 20)
-
Specification