# US 2025/0094358 A1 · Methods and apparatus to reduce bank pressure using aggressive write merging

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

- Number: US 2025/0094358 A1
- Document ID: US20250094358A1
- Status: Published application
- Kind: A1
- Application number: 18/966,252
- Filed: 2024-12-03
- Published: 2025-03-20
- Earliest claimed: 2019-05-24
- Assignees: Texas Instruments Incorporated
- Applicants: Texas Instruments Incorporated
- Inventors: Naveen Bhoria; Timothy David Anderson; Pete Michael Hippleheuser
- Web page: https://patentgrep.com/patent/US20250094358A1

## CPC

- G06F 12/128: adapted to multidimensional cache systems, e.g. set-associative, multicache, multiset or multilevel
- G06F 9/3001: {Arithmetic instructions}
- G06F 9/30043: {LOAD or STORE instructions; Clear instruction}
- G06F 9/30047: {Prefetch instructions; cache control instructions}
- G06F 9/546: {Message passing systems or structures, e.g. queues}
- G06F 11/1064: {in cache or content addressable memories}
- G06F 12/0215: {with look ahead addressing means}
- G06F 12/0238: {Memory management in non-volatile memory, e.g. resistive RAM or ferroelectric memory}
- G06F 12/0292: {using tables or multilevel address translation means (G06F12/023 takes precedence; address translation in virtual memory systems G06F12/10)}
- G06F 12/0802: Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches
- G06F 12/0804: with main memory updating (G06F12/0806 takes precedence)
- G06F 12/0806: Multiuser, multiprocessor or multiprocessing cache systems
- G06F 12/0811: with multilevel cache hierarchies
- G06F 12/0815: Cache consistency protocols
- G06F 12/082: {Associative directories (G06F12/0822 takes precedence)}
- G06F 12/0853: Cache with multiport tag or data arrays
- G06F 12/0855: Overlapped cache accessing, e.g. pipeline (G06F12/0846 takes precedence)
- G06F 12/0864: using pseudo-associative means, e.g. set-associative or hashing
- G06F 12/0884: Parallel mode, e.g. in parallel with main memory or CPU
- G06F 12/0888: using selective caching, e.g. bypass
- G06F 12/0891: using clearing, invalidating or resetting means
- G06F 12/0895: of parts of caches, e.g. directory or tag array
- G06F 12/0897: with two or more cache hierarchy levels (with multilevel cache hierarchies G06F12/0811)
- G06F 12/1027: using associative or pseudo-associative address translation means, e.g. translation look-aside buffer [TLB]
- G06F 12/12: Replacement control
- G06F 12/121: using replacement algorithms
- G06F 12/126: with special data handling, e.g. priority of data or instructions, handling errors or pinning
- G06F 12/127: using additional replacement algorithms
- G06F 13/1605: {based on arbitration (arbitration in handling access to a common bus or bus system G06F13/36)}
- G06F 13/1642: {with request queuing}
- G06F 13/1673: {using buffers}
- G06F 13/1689: {Synchronisation and timing concerns (synchronisation on a memory bus G06F13/4234)}
- G06F 15/8069: {using a cache}
- G11C 5/066: {Means for reducing external access-lines for a semiconductor memory clip, e.g. by multiplexing at least address and data signals}
- G11C 7/10: Input/output [I/O] data interface arrangements, e.g. I/O data control circuits, I/O data buffers
- G11C 7/1015: {Read-write modes for single port memories, i.e. having either a random port or a serial port}
- G11C 7/106: {Data output latches}
- G11C 7/1075: {for multiport memories each having random access ports and serial ports, e.g. video RAM}
- G11C 7/1078: {Data input circuits, e.g. write amplifiers, data input buffers, data input registers, data input level conversion circuits}
- G11C 7/1087: {Data input latches}
- G11C 7/222: {Clock generating, synchronizing or distributing circuits within memory device}
- G11C 29/42: using error correcting codes [ECC] or parity check
- G11C 29/44: Indication or identification of errors, e.g. for repair
- G11C 29/52: Protection of memory contents; Detection of errors in memory contents
- G11C 29/72: {with optimized replacement algorithms}
- G11C 29/76: {using address translation or modifications}
- G06F 2212/1016: Performance improvement
- G06F 2212/1021: Hit rate improvement
- G06F 2212/1024: Latency reduction
- G06F 2212/1041: Resource optimization
- G06F 2212/1044: Space efficiency improvement
- G06F 2212/301: In special purpose processing node, e.g. vector processor
- G06F 2212/454: Vector or matrix data
- G06F 2212/603: of operating mode, e.g. cache mode or local memory mode
- G06F 2212/6032: Way prediction in set-associative cache
- G06F 2212/6042: Allocation of cache space to multiple users or processors
- G06F 2212/608: Details relating to cache mapping
- G06F 2212/62: Details of cache specific to multiprocessor cache arrangements
- G11C 29/4401: {for self repair}
- G11C 2029/0409: {Online test}
- G11C 2029/0411: {Online error correction}
- Y02D 10/00: Energy efficient computing, e.g. low power processors, power management or thermal management
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G11C: STATIC STORES (semiconductor memory devices H10B)
- Y02D: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE

## Abstract

Methods, apparatus, systems and articles of manufacture to reduce bank pressure using aggressive write merging are disclosed. An example apparatus includes a first cache storage; a second cache storage; a store queue coupled to at least one of the first cache storage and the second cache storage and operable to: receive a first memory operation; process the first memory operation for storing the first set of data in at least one of the first cache storage and the second cache storage; receive a second memory operation; and prior to storing the first set of data in the at least one of the first cache storage and the second cache storage, merge the first memory operation and the second memory operation.

## Claims

### Claim 1

A device that includes:
a cache memory; and
a cache queue coupled to the cache memory that includes:
a first pipeline stage;
a second pipeline stage that precedes the first pipeline stage; and
a circuit coupled to the first pipeline stage and the second pipeline stage and configured to:
compare a first address associated with a first write operation of a first set of data with a second address associated with a second write operation of a second set of data; and
determine, based on comparison of the first address with the second address, whether to inhibit writing of the first set of data to the cache memory and cause the first set of data to be merged with the second set of data within the cache queue to produce a third set of data.

### Claim 2 (depends on 1)

The device of claim 1, wherein the cache queue includes:
a third pipeline stage that precedes the second pipeline stage; and
a second circuit configured to:
compare the first address associated with the first write operation with a third address associated with a third write operation of a fourth set of data; and
determine, based on comparison of the first address with the third address, whether to inhibit writing of the first set of data to the cache memory and cause the first set of data to be merged with the fourth set of data within the cache queue to produce a fifth set of data.

### Claim 3 (depends on 1)

The device of claim 1, wherein:
the first pipeline stage includes error correcting code (ECC) generation circuitry configured to provide a set of ECC data for the first set of data; and
the first pipeline stage is configured to cause the first set of data and the set of ECC data to be stored in the cache memory.

### Claim 4 (depends on 1)

The device of claim 1, wherein:
the second pipeline stage includes read-modify-write merge circuitry; and
the circuit is coupled to the read-modify-write merge circuitry to provide the third set of data to the read-modify-write merge circuitry.

### Claim 5 (depends on 1)

The device of claim 1, wherein:
the second pipeline stage includes arithmetic circuitry; and
the circuit is coupled to the arithmetic circuitry to provide the third set of data to the arithmetic circuitry.

### Claim 6 (depends on 1)

The device of claim 1, wherein:
the second pipeline stage includes atomic compare-and-swap circuitry; and
the circuit is coupled to the atomic compare-and-swap circuitry to provide the third set of data to the atomic compare-and-swap circuitry.

### Claim 7 (depends on 1)

The device of claim 1, wherein the circuit is configured to cause the first set of data to be merged with the second set of data such that a portion of the first set of data is overwritten by a portion of the second set of data.

### Claim 8 (depends on 1)

The device of claim 1, wherein the cache memory is a level-one (L1) cache memory.

### Claim 9 (depends on 1)

The device of claim 1, further comprising a main cache memory and a victim cache memory, wherein the cache memory is either the main cache memory or the victim cache memory.

### Claim 10

A device comprising:
a main cache memory;
a victim cache memory; and
a main cache queue coupled to the main cache memory and that includes:
a first pipeline stage;
a second pipeline stage that precedes the first pipeline stage; and
a first circuit coupled to the first pipeline stage and the second pipeline stage and configured to:
determine, based on an address associated with a first set of data and an address associated with a second set of data that are received for writing, whether to:
cause the first set of data to be written to the main cache memory; or
inhibit writing of the first set of data to the main cache memory and cause the first set of data to be merged with the second set of data to produce a third set of data.

### Claim 11 (depends on 10)

The device of claim 10, further comprising:
a victim cache queue coupled to the victim cache memory and that includes:
a third pipeline stage;
a fourth pipeline stage that precedes the third pipeline stage; and
a second circuit coupled to the third pipeline stage and the fourth pipeline stage and configured to:
determine, based on an address associated with a fourth set of data and an address associated with a fifth set of data that are received for writing, whether to:
cause the fourth set of data to be written to the victim cache memory; or
inhibit writing of the fourth set of data to the victim cache memory and cause the fourth set of data to be merged with the fifth set of data to produce a sixth set of data.

### Claim 12 (depends on 10)

The device of claim 10, wherein the main cache queue includes:
a third pipeline stage that precedes the second pipeline stage; and
a second circuit coupled to the first pipeline stage and the third pipeline stage and configured to:
determine, based on the address associated with the first set of data and an address associated with a fourth set of data that is received for writing, whether to inhibit writing of the first set of data to the main cache memory and cause the first set of data to be merged with the fourth set of data to produce a fifth set of data.

### Claim 13 (depends on 10)

The device of claim 10, wherein:
the first pipeline stage includes error correcting code (ECC) generation circuitry configured to provide a set of ECC data for the first set of data; and
the first pipeline stage is configured to cause the first set of data and the set of ECC data to be stored in the main cache memory.

### Claim 14 (depends on 10)

The device of claim 10, wherein the first circuit is configured to cause the first set of data to be merged with the second set of data such that a portion of the first set of data is overwritten by a portion of the second set of data.

### Claim 15 (depends on 10)

The device of claim 10, wherein the main cache memory is a level-one (L1) main cache memory, and the victim cache memory is an L1 victim cache memory.

### Claim 16

A method comprising:
receiving, by a cache queue, sets of data for writing to a cache memory that include a first set of data and a second set of data, wherein the cache queue includes a first pipeline stage and a second pipeline stage preceding the first pipeline stage;
based on the first set of data being in the first pipeline stage and the second set of data being in the second pipeline stage, comparing a first address associated with the first set of data with a second address associated with the second set of data; and
determining, based on comparison of the first address with the second address, whether to inhibit writing of the first set of data to the cache memory and cause the first set of data to be merged with the second set of data within the cache queue to produce a third set of data.

### Claim 17 (depends on 16)

The method of claim 16, wherein:
the sets of data include a fourth set of data; and
the method comprises:
based on the first set of data being in the first pipeline stage and the fourth set of data being in a third pipeline stage that precedes the second pipeline stage, comparing the first address with a third address associated with the fourth set of data; and
determining, based on comparison of the first address with the third address, whether to inhibit writing of the first set of data to the cache memory and cause the first set of data to be merged with the fourth set of data within the cache queue to produce a fifth set of data.

### Claim 18 (depends on 17)

The method of claim 17 further comprising generating error correcting code (ECC) data for the first set of data, wherein determining whether to inhibit writing of the first set of data to the cache memory includes determining whether to inhibit writing of the ECC data to the cache memory.

### Claim 19 (depends on 17)

The method of claim 17, wherein the cache queue is a level-one (L1) cache queue.

### Claim 20 (depends on 17)

The method of claim 17 wherein the cache memory is a victim cache memory.
