# US 2025/0157561 A1 · Error detection for sram used in a safety-critical domain

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

- Number: US 2025/0157561 A1
- Document ID: US20250157561A1
- Status: Published application
- Kind: A1
- Application number: 18/389,056
- Filed: 2023-11-13
- Published: 2025-05-15
- Earliest claimed: 2023-11-13
- Assignees: Advanced Micro Devices, Inc.; ATI Technologies ULC
- Applicants: Advanced Micro Devices, Inc.; Ati Technologies Ulc
- Inventors: Uma Sankara Rao Balla; Carl Kittredge Wakeland; Kaushal Amolak Sanghai; Balatripura S. Chavali; Andy Sung
- Web page: https://patentgrep.com/patent/US20250157561A1

## CPC

- G11C 29/42: using error correcting codes [ECC] or parity check
- G11C 29/52: Protection of memory contents; Detection of errors in memory contents
- G11C 2029/0409: {Online test}
- G11C: STATIC STORES (semiconductor memory devices H10B)

## Abstract

A system on a chip (SOC) includes a critical domain including components configured to perform critical operations and a non-critical domain including components configured to perform non-critical operations. To help perform such operations, the critical domain and non-critical domain share a static random-access memory (SRAM) that includes a first subset of memory banks assigned to the critical domain and a second subset of memory banks assigned to the non-critical domain. The SOC further includes a memory scrubbing circuitry configured to sequentially check each memory bank of the SRAM for errors. To this end, the memory scrubbing circuitry is configured to check a respective memory bank for errors each time an event trigger occurs by implementing one or more error correction codes.

## Claims

### Claim 1

A processing system comprising:
a random-access memory (RAM) shared between a first domain and a second domain of a system on a chip (SOC), the first domain including a first set of components of the SOC and the second domain including a second set of components of the SOC; and
a finite state machine configured to check a first bank of the RAM for one or more memory errors concurrently with read and write access to the first bank of the RAM being disabled for the first set of components of the first domain and the second set of components of the second domain.

### Claim 2 (depends on 1)

The processing system of claim 1, wherein the finite state machine is configured to check the first bank of the RAM for one or more errors by implementing an error correction code.

### Claim 3 (depends on 1)

The processing system of claim 1, wherein the first domain is powered by a first power rail and the second domain is powered by a second power rail that is different from the first power rail.

### Claim 4 (depends on 1)

The processing system of claim 1, wherein the finite state machine is configured to:
check a second bank of the RAM based on a predetermined time interval elapsing after the finite state machine has checked the first bank of the RAM for one or more memory errors, wherein the second bank is a next sequential memory bank of the RAM as indicated by a memory address associated with the RAM.

### Claim 5 (depends on 1)

The processing system of claim 1, wherein the finite state machine is configured to:
concurrently with checking the first bank of the RAM for one or more memory errors, receive a redirect request indicating a memory error in a second bank of the RAM; and
in response to receiving the redirect request, check the second bank of the RAM for one or more errors.

### Claim 6 (depends on 1)

The processing system of claim 1, wherein the finite state machine is configured to: based on detecting a multi-bit memory error in the first bank, generate an interrupt.

### Claim 7 (depends on 1)

The processing system of claim 1, wherein the finite state machine is configured to:
concurrently with read and write access to a first bank of the RAM being disabled for the first set of components of the first domain and the second set of components of the second domain and in response to detecting a single-bit memory error in the first bank, write corrected data into the first bank.

### Claim 8 (depends on 7)

The processing system of claim 7, wherein the finite state machine is configured to:
provide, to a memory controller, a request indicating that read and write access to the first bank of the RAM is to be disabled based on a predetermined period of time elapsing.

### Claim 9

A method, comprising:
reading, by a finite state machine, data from a first bank of a random-access memory (RAM) shared between a first domain and a second domain of a system on a chip (SOC) concurrently with read and write access to the first bank of the RAM being disabled for one or more components of the first domain of the SOC and one or more components of the second domain of the SOC; and
checking, at the finite state machine, the data from the first bank for one or more memory errors.

### Claim 10 (depends on 9)

The method of claim 9, wherein checking the data from the first bank for one or more memory errors includes implementing an error correction code.

### Claim 11 (depends on 9)

The method of claim 9, wherein the first domain is powered by a first power rail and the second domain is powered by a second power rail that is different from the first power rail.

### Claim 12 (depends on 9)

The method of claim 9, further comprising:
checking a second bank of the RAM for one or more memory errors based on a predetermined time interval elapsing after the finite state machine has checked the first bank of the RAM for one or more memory errors, wherein the second bank is a next sequential memory bank of the RAM as indicated by a memory address associated with the RAM.

### Claim 13 (depends on 9)

The method of claim 9, further comprising:
concurrently with checking the first bank of the RAM for one or more memory errors, receiving, at the finite state machine, a redirect request indicating a memory error in a second bank of the RAM; and
in response to receiving the redirect request, checking a second bank of the RAM for one or more errors.

### Claim 14 (depends on 9)

The method of claim 9, further comprising:
based on detecting a multi-bit memory error in the first bank, generating an interrupt.

### Claim 15 (depends on 9)

The method of claim 9, further comprising:
concurrently with read and write access to a first bank of the RAM being disabled for the one or more components of the first domain and the one or more components of the second domain and in response to detecting a single-bit memory error in the first bank, writing corrected data into the first bank.

### Claim 16 (depends on 15)

The method of claim 15, further comprising:
providing, to a memory controller, a request indicating that read and write access to the first bank of the RAM is to be disabled based on a predetermined period of time elapsing.

### Claim 17

A system comprising:
a non-critical domain comprising a central processing unit;
a critical domain powered independently from the non-critical domain;
a random access memory (RAM) shared between the non-critical domain and the critical domain, the RAM including a plurality of banks; and
a finite state machine configured to:
sequentially check each bank of the plurality of banks for one or more memory errors.

### Claim 18 (depends on 17)

The system of claim 17, wherein the finite state machine is configured to wait a predetermined time interval after checking each bank of the plurality of banks for one or more memory errors.

### Claim 19 (depends on 17)

The system of claim 17, wherein a first subset of the plurality of banks is assigned to the critical domain and a second subset of the plurality of banks is assigned to the non-critical domain.

### Claim 20 (depends on 17)

The system of claim 17, wherein the finite state machine is configured to, based on detecting a memory error in a bank of the plurality of banks, write corrected data to the bank, wherein write access to the bank is blocked for components of the critical domain.
