# US 2004/0205295 A1 · Apparatus and method to share a cache memory

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2004/0205295 A1
- Document ID: US20040205295A1
- Status: Published application
- Kind: A1
- Application number: 10/411,725
- Filed: 2003-04-11
- Published: 2004-10-14
- Earliest claimed: 2003-04-11
- Assignees: Intel Corporation
- Inventors: Dennis M. O'Connor; Michael W. Morrow
- Web page: https://patentgrep.com/patent/US20040205295A1

## CPC

- G06F 12/084: with a shared cache
- G06F 12/0866: for peripheral storage systems, e.g. disk cache
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Briefly, in accordance with an embodiment of the invention, an apparatus and method to share a cache memory is provided. The method may include dynamically partitioning a cache memory to cache data for one or more active clients during operation of the cache memory, wherein the number of active clients varies during operation of the cache memory. The apparatus may include a control device coupled to a cache memory to dynamically partition the cache memory.

## Claims

### Claim 1

A method to share a cache memory, comprising:
partitioning a cache memory to cache data for one or more clients, wherein the number of clients varies over time.

### Claim 2 (depends on 1)

The method of claim 1, wherein partitioning includes dynamically partitioning a cache memory to cache data for at least two active clients during operation of the cache memory, wherein the number of active clients varies during operation of the cache memory between one active client to N active clients.

### Claim 3 (depends on 2)

The method of claim 2, wherein N is a power of two.

### Claim 4 (depends on 2)

The method of claim 2, further including preserving data of active clients after varying the number of active clients.

### Claim 5 (depends on 2)

The method of claim 2, wherein partitioning includes partitioning the cache memory equally between the at least two active clients.

### Claim 6 (depends on 2)

The method of claim 2, further comprising:
assigning a first client identifier to a first active client of the at least two active clients;
assigning a second client identifier to a second active client of the at least two active clients;
requesting information for the first active client by sending an address from the first active client;
comparing a value of a first portion of the address to a value of a tag stored at a location in the cache memory corresponding to a second portion of the address, wherein the tag includes either the first client identifier or the second client identifier; and
indicating a cache miss if the value of the first portion of the address is not equal to the value of the tag; and
responding to the cache miss by allocating a cache line of the cache memory with data from another higher latency memory, wherein allocating includes storing the first client identifier in a tag of the cache line.

### Claim 7 (depends on 1)

The method of claim 1, wherein partitioning includes:
assigning a client identifier to a first active client of the one or more clients; and
accessing the cache memory using an address having a tag portion, a set index portion, and an offset portion, wherein accessing includes accessing the cache memory using the client identifier and the set index portion of the address.

### Claim 8 (depends on 7)

The method of claim 7, wherein the client identifier is combined with the set index portion of the address so that the client identifier is the high bit of the set index.

### Claim 9 (depends on 7)

The method of claim 7, further comprising:
comparing the tag portion of the address to a stored tag located at a cache line in the cache memory corresponding to the set index portion and the client identifier, wherein the stored tag includes the client identifier; and
indicating a cache hit if a value of the stored tag is equal to a value of the tag portion of the address.

### Claim 10 (depends on 1)

The method of claim 1, wherein partitioning includes:
partitioning a cache memory to cache data for one or more active clients, wherein the number of active clients varies over time;
allowing use of the whole cache memory to one active client of the one or more active clients if the number of active clients is one;
allowing use of a first half of the cache memory by a first active client of the one or more active clients, if the number of active clients is two; and
allowing use of a second half of the cache memory by a second active client of the one or more active clients, if the number of active clients is two.

### Claim 11 (depends on 10)

The method of claim 10, further comprising:
populating the first half of the cache memory with data for use by the first active client;
populating the second half of the cache memory with data for use by the second active client;
transitioning the second active client from active to inactive operation; and
preserving the data in the first half of the cache memory for use by the first active client after transitioning.

### Claim 12 (depends on 1)

The method of claim 1, wherein partitioning includes:
assigning a client identifier to a first client of the one or more clients; and
allocating a cache line of the cache memory with data from another higher latency memory, wherein allocating includes storing the client identifier in a tag of the cache line.

### Claim 13

An apparatus, comprising:
a cache memory; and
a control device coupled to the cache memory, wherein the control device is adapted to allow use of the cache memory by one or more clients, wherein the number of clients varies over time.

### Claim 14 (depends on 13)

The apparatus of claim 13, wherein the number of clients varies between one client and N clients.

### Claim 15 (depends on 13)

The apparatus of claim 13, wherein the one or more clients are one or more transaction initiators.

### Claim 16 (depends on 13)

The apparatus of claim 13, wherein the control device is adapted to allow access to a first portion of the cache memory by a first active client of the one or more clients and adapted to allow access to a second different portion of the cache memory by a second active client of the one or more clients.

### Claim 17 (depends on 16)

The apparatus of claim 16, wherein the first portion is a first half of the cache memory and wherein the second portion is the second half of the cache memory.

### Claim 18 (depends on 16)

The apparatus of claim 16, wherein the control device is adapted to assign a first client identifier corresponding to the first active client and adapted to assign a second client identifier corresponding to the second active client; and
wherein the control device is adapted to equally divide use of the cache memory between the first and second active clients during operation of the cache memory.

### Claim 19 (depends on 13)

The apparatus of claim 13,
wherein the control device is adapted to assign a client identifier to a first active client of the one or more clients;
wherein the control device is adapted to receive an address from the first active client; and
wherein the control device is adapted to use the client identifier and a first portion of the address to allow the first client to access a first portion of the cache memory.

### Claim 20 (depends on 19)

The method of claim 19,
wherein the control device is adapted to compare a value of the second portion of the address and a value of the tag stored at a cache line located at a location in the cache memory corresponding to the first portion of the address and the client identifier;
wherein the control device is adapted to indicate a cache miss if the value of the second portion of the address is not equal to the value of the tag; and
wherein the control device is adapted to respond to the cache miss by allocating a cache line of the cache memory with data from another higher latency memory and wherein allocating includes storing the client identifier in a tag of the cache line.

### Claim 21 (depends on 13)

The apparatus of claim 13, wherein the cache memory is a direct mapped cache memory.

### Claim 22

An article comprising a storage medium having stored thereon instructions, that, when executed by a computing platform, result in: partitioning a cache memory to cache data for one or more clients, wherein the number of clients varies over time.

### Claim 23 (depends on 22)

The article of claim 22, wherein the instructions, when executed, further result in: preserving data of active clients of the one or more clients after varying the number of active clients.

### Claim 24 (depends on 22)

The article of claim 22, wherein if the number of clients is at least two clients, partitioning includes partitioning the cache memory equally between the at least two clients.

### Claim 25

A system, comprising:
a wireless transceiver;
a cache memory coupled to the wireless transceiver; and
a control device coupled to the cache memory, wherein the control device is adapted to allow use of the cache memory by one or more clients, wherein the number of clients varies over time.

### Claim 26 (depends on 25)

The system of claim 25, further comprising a dipole antenna coupled to the wireless transceiver.

### Claim 27 (depends on 25)

The system of claim 25, wherein the control device is adapted to allow access to a first portion of the cache memory by a first active client of the one or more clients and adapted to allow access to a second different portion of the cache memory by a second active client of the one or more clients.
