IP Library › Granted Patent US 12,430,256
Granted Patent B2
US 12,430,256 · App. 18/215,115 · Granted Sep 30, 2025

Unloaded cache bypass

Inventor: Emanuele Confalonieri (Segrate, IT)
Assignee: Micron Technology, Inc.
G06F12/0888G06F2212/60
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,430,256
App. No.
18/215,115
Granted
Sep 30, 2025
Kind
B2
Abstract

Systems, apparatuses, and methods related to a memory controller for unloaded cache bypass are described. An example memory controller can be coupled to a memory device. The example memory controller can include a cache. The cache can include a cache sequence controller configured to determine a quantity of a pending cache look-up operations, determine the quantity satisfies an unloaded bypass threshold, and cause performance of a bypass memory operation that bypasses the cache and accesses the memory device.

Claims (43)

1. An apparatus, comprising:

a memory device; and

a memory controller coupled to the memory device and configured to receive a host read request, the memory controller comprising:

a cache;

a switch coupled to the memory device; and

a cache sequence controller coupled to the cache and the switch, configured to:

responsive to receiving the host read request, determine that a quantity of pending cache look-up operations in a queue of the cache satisfies a variable unloaded bypass threshold, wherein the variable unloaded bypass threshold is varied based on a cache hit rate of cache look-up operations associated with the cache; and

cause performance of a bypass read operation that bypasses the cache and accesses the memory device via the switch responsive to determining the quantity satisfies the variable unloaded bypass threshold.

2. The apparatus of claim 1 , wherein the memory controller is configured to cause performance of a cache look-up operation associated with the cache substantially contemporaneously with the performance of the bypass read operation.

3. The apparatus of claim 2 , wherein performance of the cache look-up operation results in a cache hit, and wherein the memory controller is further to discard data read from the memory device in association with the bypass read operation.

4. The apparatus of claim 2 , wherein performance of the cache look-up operation results in a cache miss.

5. The apparatus of claim 4 , wherein the cache miss is a clean cache miss, and wherein the cache sequence controller is further to cache-in data read from the memory device in association with the bypass read operation.

6. The apparatus of claim 4 , wherein the cache miss is a dirty cache miss, and wherein:

the cache sequence controller is further to cache-in data read from the memory device in association with the bypass read operation; and

the memory controller is to cause performance of a cache write-back operation to write data from the cache to the memory device.

7. The apparatus of claim 1 , wherein memory controller includes metric logic including counters, and wherein the counters comprise a cache hit counter, write hit counter, or a combination thereof.

8. The apparatus of claim 1 , wherein the cache further comprises a cache memory and a cache controller.

9. The apparatus of claim 1 , wherein the quantity of pending cache look-up operations satisfies the variable unloaded bypass threshold when the quantity of pending cache look-up operations is less than or equal to the variable unloaded bypass threshold.

10. A system, comprising:

a memory device; and

a memory controller coupled to the memory device and configured to receive a host read request, the memory controller comprising:

a cache;

a switch coupled to the memory device; and

a cache sequence controller coupled to the cache and the switch, configured to:

responsive to receiving the host read request:

determine, based on collected metrics, that a quantity of pending cache look-up operations in a queue of the cache satisfies a variable unloaded bypass threshold, wherein the variable unloaded bypass threshold is varied based on a cache hit rate of cache look-up operations associated with the cache;

cause performance of a cache look-up operation; and

in parallel with performance of the cache look-up operation, cause initiation of a bypass read operation that bypasses the cache and accesses the memory device via the switch to read data corresponding to the host read request from the memory device.

11. The system of claim 10 , wherein the cache is configured to operate in accordance with a First-in First-out (FIFO) cache replacement policy.

12. The system of claim 10 , wherein the cache sequence controller is to cause performance of a plurality of bypass read operations when the quantity of the pending cache look-up operations in the queue is less than or equal to the variable unloaded bypass threshold.

13. The system of claim 10 , further comprising an exit threshold that is less than the variable unloaded bypass threshold, and wherein the cache sequence controller is to abstain from performance of a bypass read operation until the quantity of the pending look-up operations is less than the exit threshold.

14. The system of claim 10 , wherein:

the variable unloaded bypass threshold is decreased based on the cache hit rate being above a target hit rate; and

the variable unloaded bypass threshold is increased based on the cache hit rate being below a target hit rate.

15. A method, comprising:

determining, by a memory controller coupled to a memory device, a quantity of pending cache look-up operations in a queue associated with a cache responsive to receiving a host access request;

determining the quantity of pending cache look-up operations is less than or equal to an unloaded bypass threshold that is variable based on a cache hit rate of cache look-up operations associated with the cache; and

responsive to determining the quantity of pending cache look-up operations is less than or equal to the unloaded bypass threshold:

causing performance of a bypass read operation that bypasses the cache and accesses the memory device via a switch of the memory controller that is coupled to the memory device; and

causing performance of a cache look-up operation associated with the cache substantially contemporaneously with the performance of the bypass read operation.

16. The method of claim 15 , wherein data associated with the performance of the bypass read operation is selectively utilized based on a type of result of the cache look-up operation.

17. The method of claim 15 , wherein the unloaded bypass threshold is equal to a non-zero quantity of pending cache look-up operations.

18. The method of claim 15 , wherein the unloaded bypass threshold is equal to zero pending cache look-up operations.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jun 28, 2023
From: CONFALONIERI, EMANUELE
To: MICRON TECHNOLOGY, INC.
Reel/Frame 064095/0231 →
Continuity (2)
Provisional Application 63356378 · Jun 28, 2022
Related Publication 20230418755A1 · Dec 28, 2023
References Cited (15)
US 4858111A · Steps · 1989 [cited by examiner]
US 6470425B1 · Yamashiroya · 2002 [cited by examiner]
US 11409601B1 · Brewer et al. · 2022 [cited by applicant]
US 12130739B2 · Mandal · 2024 [cited by examiner]
US 20040064648A1 · Brown · 2004 [cited by examiner]
US 20070113019A1 · Beukema · 2007 [cited by examiner]
US 20070226418A1 · Mizuno · 2007 [cited by examiner]
US 20130262767A1 · Lih · 2013 [cited by examiner]
US 20170357600A1 · Moon · 2017 [cited by applicant]
US 20190042451A1 · Leone · 2019 [cited by examiner]
US 20190079877A1 · Gaur et al. · 2019 [cited by applicant]
US 20200174938A1 · Shergill · 2020 [cited by examiner]
US 20210034539A1 · Roberts · 2021 [cited by examiner]
Rathi, Nitin, et al. “Cache bypassing and checkpointing to circumvent data security attacks on sttram.” arXiv preprint arXiv: 1603.06227 (2016). (Year: 2016). [cited by examiner]
Xie, Xiaolong, et al. “Coordinated static and dynamic cache bypassing for GPUs.” 2015 IEEE 21st International Symposium on High Performance Computer Architecture (HPCA). IEEE, 2015. (Year: 2015). [cited by examiner]