Intel Corporation

Other

Sk Hynix Nand Product Solutions Corp.

Tahoe Research, Ltd.

Rajat Agarwal

Wei P Chen

David M Durham

Kuljit Singh Bains

Karanvir S Grewal

Sergej Deutsch

Kai Cheng

Glenn J Hinton

Raj K Ramanujan

James W Alexander

Siddhartha Chhabra

Bill Nale

Debaleena Das

Mohan J Kumar

Jing Ling

Kjersten E Criss

Christopher J Hughes

Robert L Sankman

Debendra Mallik

Ravindranath Vithal Mahajan

James A McCall

Ioannis T Schoinas

Dheeraj R Subbareddy

Ankireddy Nalamalpu

Md Altaf Hossain

Ravi Prakash Gutala

Wei Wu

Scott J Weber

Ram S Viswanath

David Zimmerman

Jong Soo Park

Aravind Raghavendra Dasu

Brian S Morris

Sandeep B Sane

Sriram Srinivasan

Sandeep Ahuja

Dimitrios Ziakas

Bassam N Coury

Muthukumar Swaminathan

Ruchira Sasanka

Suresh S Chittor

Dennis W Brzezinski

Berkin Akin

Camille C Raad

Taarinya Polepeddi

Bruce A Christenson

Daniel W Liu

C Scott Huddleston

John B Halbert

Murugasamy K Nachimuthu

George Vergis

Alaa R Alameldeen

Theodros Yigzaw

John G Holm

Andrew M Rudoff

Eric W Buddrius

Meenakshisundaram Chinthamani

Chet R Douglas

Andy M Rudoff

Uksong Kang

Jongwon Lee

Ramesh Subashchandrabose

Tiffany J Kasanicky

Chiachen Chou

Roger D Flynn

Sai Prashanth Muralidhara

Anupam Mohanty

Kirk S Yap

Baiju V Patel

Shih-Lien L Lu

Santosh Ghosh

Avinash Sodani

Anatoli Aleksandrovich Bolotov

Ron Gabor

Pete D Vogt

Mikhail I Grinchuk

Howard S David

Zeshan A Chishti

Cong Li

Zhe Wang

Sivakumar Radhakrishnan

Adrian C Moga

Deep K Buch

Suneeta Sah

Subhankar Panda

William Harry Nale

Mariusz Oriol

Sreenivas Mandava

Chong J Zhao

Henry Stracovsky

Vivek Joy Kozhikkottu

Henk G Neefs

Xiaoming Du

Darrell S McGinnis

Timothy Y Kam

Rebecca Z Loop

Malcolm Mandviwalla

Christoph Dobraunig

Hsing-Min Chen

Hussein Alameer

A system can dynamically migrate memory pages from near memory to far memory during runtime. A system basic input output system (BIOS) can program a first memory address space of size P and a second memory address space of size P to a near memory (NM) space of size (N) and a far memory (FM) space of size (M), where P equals N+M. For the first memory address space, the OS can manage the NM space and the FM space as a flat memory space with an address space of size P available. For the second memory address space, the OS can manage the NM space as a NM cache for FM, with an address space of size M available.

Dynamic multilevel memory system

A system includes a multilevel memory such as a two level memory (2LM), where a first level memory acts as a cache for the second level memory. A memory controller or cache controller can detect a cache miss in the first level memory for a request for data. Instead of automatically performing a swap, the controller can determine whether to perform a swap based on a swap policy assigned to a memory region associated with the address of the requested data.

Selective fill for logical control over hardware multilevel memory

A memory subsystem includes multiple memory resources connected in parallel, including a first memory resource and a second memory resource. The memory subsystem can split a portion of data into multiple sub-portions. Split into smaller portions, the system needs fewer ECC (error checking and correction) bits to provide the same level of ECC protection. The portion of data can include N ECC bits for error correction, and the sub-portions can each include a sub-portion of (N−M) ECC bits for error correction. The system can then use M bits of data for non-ECC purposes, such as metadata.

Distribution of error checking and correction (ECC) bits to allocate ECC bits for metadata

A memory subsystem with error checking and scrubbing (ECS) logic on-device on the memory can adapt the rate of ECS operations in response to detection of errors in the memory when the memory device is in automatic ECS mode. The ECS logic can include an indication of rows of memory that have been offlined by the host. The ECS logic can skip the offlined rows in ECS operation counts. The ECS logic can include requests or hints by the host to have ECS operations performed. An internal address generator of the ECS logic can select between generated addresses and the hints. The system can allow a memory controller to detect multibit errors (MBEs) related to a specific address of the associated memory. When the detected MBEs indicate a pattern of errors, the memory controller triggers a row hammer response for the specific address.

Adaptive internal error scrubbing and error handling

Embodiments disclosed herein include electronic packages. In an embodiment, the electronic package comprises, a package substrate, an interposer on the package substrate, a first die cube and a second die cube on the interposer, wherein the interposer includes conductive traces for electrically coupling the first die cube to the second die cube, a die on the package substrate, and an embedded multi-die interconnect bridge (EMIB) in the package substrate, wherein the EMIB electrically couples the interposer to the die.

Interconnect architecture with silicon interposer and EMIB

A two-level main memory in which both volatile memory and persistent memory are exposed to the operating system in a flat manner and data movement and management is performed in cache line granularity is provided. The operating system can allocate pages in the two-level main memory randomly across the first level main memory and the second level main memory in a memory-type agnostic manner, or, in a more intelligent manner by allocating predicted hot pages in first level main memory and predicted cold pages in second level main memory. The cache line granularity movement is performed in a 'swap' manner, that is, a hot cache line in the second level main memory is swapped with a cold cache line in first level main memory because data is stored in either first level main memory or second level main memory not in both first level main memory and second level main memory.

Two-level main memory hierarchy management

A system can predict memory device failure through identification of correctable error patterns based on the memory architecture. The failure prediction can thus account for the circuit-level of the memory rather than the mere number or frequency of correctable errors. A failure prediction engine correlates hardware configuration of the memory device with correctable errors (CEs) detected in data of the memory device to predict an uncorrectable error (UE) based on the correlation.

Uncorrectable memory error prediction

The technology described herein includes a first plurality of bijection diffusion function circuits to diffuse data bits into diffused data bits and store the diffused data bits into a memory; an error correcting code (ECC) generation circuit to generate ECC bits for the data bits; and a second plurality of bijection diffusion function circuits to diffuse the ECC bits into diffused ECC bits and store the diffused ECC bits into the memory.

Message authentication Galois integrity and correction (MAGIC) for lightweight row hammer mitigation

A memory controller having an error checking and correction (ECC) circuitry can detect an error in data being read from memory and correct that error with a retry flow without needing to send another read to the memory for the data. The read data is stored in a read data buffer (RDB) at the memory controller when the read data is received from memory. The memory controller has an error detection path from the RDB to the host and an error correction path. Read data that has no errors can be sent directly to the host. Instead of flushing the RDB in response to the error detection, the memory controller executes a retry flow, where the RDB provides the read data to the error correction path for error correction.

Fast memory ECC error correction

A memory subsystem includes memory devices with space dynamically allocated for improvement of reliability, availability, and serviceability (RAS) in the system. Error checking and correction (ECC) logic detects an error in all or a portion of a memory device. In response to error detection, the system can dynamically perform one or more of: allocate active memory device space for sparing to spare a failed memory segment; write a poison pattern into a failed cacheline to mark it as failed; perform permanent fault detection (PFD) and adjust application of ECC based on PFD detection; or, spare only a portion of a device and leave another portion active, including adjusting ECC based on the spared portion. The error detection can be based on bits of an ECC device, and error correction based on those bits and additional bits stored on the data devices.

Adaptive error correction to improve system memory reliability, availability, and serviceability (RAS)

In one example, a system for managing encrypted memory comprises a processor to store a first MAC based on data stored in system memory in response to a write operation to the system memory. The processor can also detect a read operation corresponding to the data stored in the system memory, calculate a second MAC based on the data retrieved from the system memory, determine that the second MAC does not match the first MAC, and recalculate the second MAC with a correction operation, wherein the correction operation comprises an XOR operation based on the data retrieved from the system memory and a replacement value for a device of the system memory. Furthermore, the processor can decrypt the data stored in the system memory in response to detecting the recalculated second MAC matches the first MAC and transmit the decrypted data to cache thereby correcting memory errors.

Cryptographic system memory management

An apparatus and method to reduce bandwidth and latency associated with probabilistic caches. For example, one embodiment of a processor comprises: a plurality of cores to execute instructions and process data, one or more of the cores to generate a request for a first cache line; a cache controller comprising cache lookup logic to determine a first way of a cache in which to search for the first cache line based on a first set of tag bits comprising one or more bits associated with the first cache line; the cache lookup logic to compare a second set of tag bits of the first cache line with a third set of tag bits of an existing cache line stored in the first way, wherein if the second set of tag bits and the third set of tag bits to not match, then the cache lookup logic to determine that the first cache line is not in the first way and to compare a fourth set of tag bits of the first cache line with a fifth set of tag bits of the existing cache line, wherein responsive to a match between the fourth set of tag bits and the fifth set of tag bits, the cache lookup logic to determine that the first cache line is stored in a second way and to responsively read the first cache line from the second way.

Apparatus and method to reduce bandwidth and latency overheads of probabilistic caches

The technology disclosed herein comprises a processor; a memory to store data and a plurality of error correcting code (ECC) bits associated with the data; and a memory controller coupled to the memory, the memory controller to receive a write request from the processor and, when an access control field is selected in the write request, perform an exclusive OR (XOR) operation on the plurality of ECC bits and a fixed encoding pattern to generate a plurality of encoded ECC bits and store the data and the plurality of encoded ECC bits in the memory.

Efficient security metadata encoding in error correcting code (ECC) memory without dedicated ECC bits

A method comprises generating, for a cacheline, a first tag and a second tag, the first tag and the second tag generated as a function of user data stored and metadata in the cacheline stored in a first memory device, and a multiplication parameter derived from a secret key, storing the user data, the metadata, the first tag and the second tag in the first cacheline of the first memory device; generating, for the cacheline, a third tag and a fourth tag, the third tag and the fourth tag generated as a function of the user data stored and metadata in the cacheline stored in a second memory device, and the multiplication parameter; storing the user data, the metadata, the third tag and the fourth tag in the corresponding cache line of the second memory device; receiving, from a requesting device, a read operation directed to the cacheline; and using the first tag, the second tag, the third tag, and the fourth tag to determine whether a read error occurred during the read operation.

Algebraic and deterministic memory authentication and correction with coupled cacheline metadata

A memory device that performs internal ECC (error checking and correction) can selectively return read data with application of the internal ECC or without application of the internal ECC, in response to different read commands from the memory controller. The memory device can normally apply ECC and return corrected data in response to a normal read command. In response to a retry command, the memory device can return the read data without application of the internal ECC.

Read retry to selectively disable on-die ECC

Growing community of inventors

Beaverton, OR, United States of America

Rajat Agarwal