IP Library Granted Patent US 12,223,207
Granted Patent B2
US 12,223,207 · App. 18/487,955 · Granted Feb 11, 2025

Memory component having internal read-modify-write operation

Inventors: Frederick A. Ware (Los Altos Hills, CA); Thomas Vogelsang (Mountain View, CA)
Assignee: Rambus Inc.
G06F3/0659G06F3/0604G06F3/0619G06F3/0638G06F3/0673G06F11/1076G11C7/1006G11C7/1009G11C7/1087G11C7/109G11C7/1093G11C29/023G11C29/028G11C2029/0411G11C2207/107
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 12,223,207
App. No.
18/487,955
Granted
Feb 11, 2025
Kind
B2
Abstract

An memory component includes a memory bank and a command interface to receive a read-modify-write command, having an associated read address indicating a location in the memory bank and to either access read data from the location in the memory bank indicated by the read address after an adjustable delay period transpires from a time at which the read-modify-write command was received or to overlap multiple read-modify-write commands. The memory component further includes a data interface to receive write data associated with the read-modify-write command and an error correction circuit to merge the received write data with the read data to form a merged data and write the merged data to the location in the memory bank indicated by the read address.

Claims (53)

1. A memory device comprising:

a first memory die; and

a second memory die stacked with the first memory die, wherein each of the first memory die and the second memory die comprise:

one or more memory banks;

a command interface coupled to the one or more memory banks, the command interface to receive, from a memory controller, a read-modify-write command having an associated read address indicating a location in the one or more memory banks and to access read data from the location in the one or more memory banks indicated by the read address after an adjustable delay period transpires from a time at which the read-modify-write command was received;

a data interface coupled to the one or more memory banks, the data interface to receive write data associated with the read-modify-write command; and

an error correction circuit coupled to the data interface and to the one or more memory banks, the error correction circuit to merge the received write data with the read data to form a merged data and write the merged data to the location in the one or more memory banks indicated by the read address.

2. The memory device of claim 1 , wherein the adjustable delay period is adjustable in view of at least one of a control signal received at a delay element in the memory device, a value specified in the read-modify-write command, or a value maintained in a control register in the memory device, and wherein increasing the adjustable delay period increases a rate at which successive read-modify-write commands are processed.

3. The memory device of claim 1 , wherein each of the first memory die and the second memory die further comprise:

a plurality of memory bank pairs; and

a dedicated error correction circuit associated with each of the plurality of memory bank pairs, wherein a read-modify-write command directed toward each memory bank pair is processed on the dedicated error correction circuit in parallel, such that error detection and correction for a read-modify-write operation for one bank of a memory bank pair proceeds in a staggered fashion with an other bank of the memory bank pair.

4. The memory device of claim 1 , wherein the adjustable delay period comprises two column cycle intervals of a system clock, and wherein the error correction circuit is further to:

decode the read data and encode the merged data within a same column cycle interval of the system clock, wherein a period of time that transpires when decoding the read data and encoding the merged data within a same column cycle interval is less than a time that transpires when decoding and encoding separately.

5. The memory device of claim 1 , wherein the error correction circuit is further to:

decode the read data prior to merging the read data with the received write data, wherein to decode the read data, the error correction circuit to correct single bit errors in the read data using an error syndrome in the read data.

6. The memory device of claim 1 , wherein to merge the received write data with the read data to form the merged data, the error correction circuit is to combine the write data and the read data using a data mask received with the write data.

7. The memory device of claim 1 , wherein the error correction circuit is further to:

encode the merged data prior to writing the merged data to the location in the memory bank, wherein to encode the merged data, the error correction circuit to generate an error syndrome for the merged data.

8. A memory device comprising:

a first memory die; and

a second memory die stacked with the first memory die, wherein each of the first memory die and the second memory die comprise:

one or more memory banks; and

processing logic coupled to the one or more memory banks, wherein the processing logic is configured to:

receive, from a memory controller, a read-modify-write command having an associated read address indicating a location in the one or more memory banks;

access read data from the location in the one or more memory banks indicated by the read address after an adjustable delay period transpires from a time at which the read-modify-write command was received;

receive write data associated with the read-modify-write command;

merge the received write data with the read data to form a merged data; and

write the merged data to the location in the one or more memory banks indicated by the read address.

9. The memory device of claim 8 , wherein the adjustable delay period is adjustable in view of at least one of a control signal received at a delay element in the memory device, a value specified in the read-modify-write command, or a value maintained in a control register in the memory device, and wherein increasing the adjustable delay period increases a rate at which successive read-modify-write commands are processed.

10. The memory device of claim 8 , wherein each of the first memory die and the second memory die further comprise:

a plurality of memory bank pairs; and

a dedicated error correction circuit associated with each of the plurality of memory bank pairs, wherein a read-modify-write command directed toward each memory bank pair is processed on the dedicated error correction circuit in parallel, such that error detection and correction for a read-modify-write operation for one bank of a memory bank pair proceeds in a staggered fashion with an other bank of the memory bank pair.

11. The memory device of claim 8 , wherein the adjustable delay period comprises two column cycle intervals of a system clock, and wherein the processing logic is further configured to:

decode the read data and encode the merged data within a same column cycle interval of the system clock, wherein a period of time that transpires when decoding the read data and encoding the merged data within a same column cycle interval is less than a time that transpires when decoding and encoding separately.

12. The memory device of claim 8 , wherein the processing logic is further configured to:

decode the read data prior to merging the read data with the received write data, wherein to decode the read data, the processing logic is to correct single bit errors in the read data using an error syndrome in the read data.

13. The memory device of claim 8 , wherein to merge the received write data with the read data to form the merged data, the processing logic is to combine the write data and the read data using a data mask received with the write data.

14. The memory device of claim 8 , wherein the processing logic is further configured to:

encode the merged data prior to writing the merged data to the location in the memory bank, wherein to encode the merged data, the processing logic is to generate an error syndrome for the merged data.

15. A method of operation in a memory device, the memory comprising:

receiving, from a memory controller, a read-modify-write command having an associated read address indicating a location in one or more memory banks of a first memory die stacked with a second memory die;

accessing read data from the location in the one or more memory banks indicated by the read address after an adjustable delay period transpires from a time at which the read-modify-write command was received;

receiving write data associated with the read-modify-write command;

merging the received write data with the read data to form a merged data; and

writing the merged data to the location in the one or more memory banks indicated by the read address.

16. The method of claim 15 , wherein the adjustable delay period is adjustable in view of at least one of a control signal received at a delay element in the memory device, a value specified in the read-modify-write command, or a value maintained in a control register in the memory device, and wherein increasing the adjustable delay period increases a rate at which successive read-modify-write commands are processed.

17. The method of claim 15 , wherein the adjustable delay period comprises two column cycle intervals of a system clock, and wherein the method further comprises:

decoding the read data and encode the merged data within a same column cycle interval of the system clock, wherein a period of time that transpires when decoding the read data and encoding the merged data within a same column cycle interval is less than a time that transpires when decoding and encoding separately.

18. The method of claim 15 , further comprising:

decoding the read data prior to merging the read data with the received write data, wherein decoding the read data comprises correcting single bit errors in the read data using an error syndrome in the read data.

19. The method of claim 15 , wherein merging the received write data with the read data to form the merged data comprises combining the write data and the read data using a data mask received with the write data.

20. The method of claim 15 , further comprising

encoding the merged data prior to writing the merged data to the location in the memory bank, wherein encoding the merged data comprises generating an error syndrome for the merged data.

Continuity (7)
Continuation 17824665 · May 25, 2022
Continuation 17247167 · Dec 2, 2020
Continuation 16371345 · Apr 1, 2019
Continuation 15990211 · May 25, 2018
Continuation 15022176
Provisional Application 61881869 · Sep 24, 2013
Related Publication 20240118837A1 · Apr 11, 2024
References Cited (47)
US 5535406A · Kolchinsky · 1996 [cited by applicant]
US 6141272A · Van de Graaff et al. · 2000 [cited by applicant]
US 6546476B1 · Gillingham · 2003 [cited by applicant]
US 7475271B2 · Oner · 2009 [cited by applicant]
US 7505356B2 · Ware et al. · 2009 [cited by applicant]
US 7908443B2 · Hillier, III et al. · 2011 [cited by applicant]
US 8050134B2 · Ware et al. · 2011 [cited by applicant]
US 8397129B2 · Schreck et al. · 2013 [cited by applicant]
US 8909860B2 · Rao · 2014 [cited by applicant]
US 9983830B2 · Ware et al. · 2018 [cited by applicant]
US 10248358B2 · Ware et al. · 2019 [cited by applicant]
US 10860253B2 · Ware · 2020 [cited by examiner]
US 11822822B2 · Ware · 2023 [cited by examiner]
US 20020015344A1 · Leung et al. · 2002 [cited by applicant]
US 20040183808A1 · Radke et al. · 2004 [cited by applicant]
US 20050273564A1 · Lakshmanamurthy et al. · 2005 [cited by applicant]
US 20060072366A1 · Ware et al. · 2006 [cited by applicant]
US 20060080589A1 · Holm et al. · 2006 [cited by applicant]
US 20060112321A1 · Leung · 2006 [cited by applicant]
US 20060123322A1 · Leung et al. · 2006 [cited by applicant]
US 20060195631A1 · Rajamani · 2006 [cited by applicant]
US 20070266300A1 · Ito et al. · 2007 [cited by applicant]
US 20080195894A1 · Schreck et al. · 2008 [cited by applicant]
US 20090276587A1 · Moyer et al. · 2009 [cited by applicant]
US 20100290150A1 · Sumi et al. · 2010 [cited by applicant]
US 20100306623A1 · Hsueh et al. · 2010 [cited by applicant]
US 20110191655A1 · Schreck et al. · 2011 [cited by applicant]
US 20120063235A1 · Yoon et al. · 2012 [cited by applicant]
US 20130194879A1 · Perego et al. · 2013 [cited by applicant]
US 20130246750A1 · Moyer et al. · 2013 [cited by applicant]
US 20140344653A1 · Flinsbaugh et al. · 2014 [cited by applicant]
US 20150378740A1 · Moyer · 2015 [cited by examiner]
US 20160231962A1 · Ware · 2016 [cited by examiner]
CN 102314950A · 2012 [cited by applicant]
CN 102714061A · 2012 [cited by applicant]
CN 102934171A · 2013 [cited by applicant]
M. Lou, Y. Cui, R. Pei, X. Zhang and L. Wu, “A fault-tolerant SDRAM controller based on a dynamically reconfigurable shift register chain,” 2014 IEEE International Conference on Progress in Informatics and Computing, Sh… [cited by examiner]
CN Notification of the First Office Action with Mail Date Nov. 23, 2022 re: CN Appln. No. 201910411592.9. 9 Pages. (w/translation). [cited by applicant]
CN Office Action With Mail Date Oct. 12, 2018 re: CN Appln. No. 201480042800.8. 7 Pages (With Translation). [cited by applicant]
CN Office Action with Mail Date Feb. 5, 2018 re: CN Appln. No. 201480042800.8. 13 Pages. (W/Translation). [cited by applicant]
Huang, Ting-Lu, “Fast and Fair Mutual Exclusion for Shared Memory Systems”, Proceedings 19th IEEE International Conference on Distributed Computing Systems (Cat. No. 99CB37003), Austin, TX, 1999, pp. 224-231. 9 pages. [cited by applicant]
Jian, Xun et al., “Parity Helix: Efficient Protection for Single-Dimensional Faults in Multi-dimensional Memory Systems”, 2016 IEEE International Symposium on High Performance Computer Architecture (HPCA), Barcelona, Sp… [cited by applicant]
Liu, Lei et al., “An Efficient LDPC Encoder Based on Block-Column Cycle Structure for CMMB”, 2013 IEEE 4th International Conference on Software Engineering and Service Science, Beijing, 2013, pp. 511-514 (Year: 2013). 4… [cited by applicant]
Noguchi, Hiroki et al., “4Mb STT-MRAM-Based Cache with Memory-Access-Aware Power Optimization and Write-Verify-Write / Read-Modify-Write Scheme”, 2016 IEEE International Solid-State Circuits Conference (ISSCC), Session … [cited by applicant]
Notification Concerning Transmittal of International Preliminary Report on Patentability with mail date of Apr. 7, 2016 re Int'l. Appln. No. PCT/US14/057040. 12 Pages. [cited by applicant]
PCT International Search Report and Written Opinion dated Dec. 29, 2014 in International Application No. PCT/US2014/057040. 13 pages. [cited by applicant]
Y.-H. Park, et al., “Bit-wise read-compare-write scheme for low power read-modify-write DRAM operation”, Electronics Letters, vol. 38, No. 2, pp. 62-63, Jan. 17, 2002. 2 Pages. [cited by applicant]