Micron Technology Incorporated

Intel Corporation

Sean Stephen Eilert

Ameen D Akel

Kenneth Marion Curewitz

Samuel E Bradshaw

Dmitri A Yudanov

Justin M Eno

Shivam Swami

Shivasankar Gunasekaran

Hongyu Wang

Mark P Leinwander

Kunal R Parekh

Glen Earl Hush

Christopher Joseph Bueb

William A Melton

Aliasger Tayeb Zaidy

Shekoufeh Qawami

Rodney R Rozman

Jared E Hulbert

Sivagnanam Parthasarathy

Hernan Adolfo Castro

Helena Caminal

Chris Bueb

Donald M Morgan

Richard Edward Fackenthal

John Rudelic

Sai Krishna Mylavarapu

Bryce D Cook

Robert Melcher

Brian Hirano

Sampath K Ratnam

Glenn J Hinton

David Andrew Roberts

Paul D Ruby

Todd A Marquart

Patrick Alan Estep

Gerard A Kreifels

Rich Fackenthal

Alec W Smidt

Sunil R Atri

David Dressler

John Egler

Peter Leung

Systems, methods and apparatuses to throttle network communications for memory as a service. In one exemplary embodiment, a computing device can borrow an amount of random access memory of the lender device over a communication connection between the lender device and the computing device. The computing device can allocate virtual memory to applications running in the computing device, and configure at least a portion of the virtual memory to be hosted on the amount of memory loaned by the lender device to the computing device. The computing device can throttle data communications used by memory regions in accessing the amount of memory over the communication connection according to the criticality levels of the contents stored in the memory regions.

Throttle memory as a service based on connectivity bandwidth and contents priority levels

Systems, methods and apparatuses of Artificial Neural Network (ANN) applications implemented via Memory as a Service (MaaS) are described. For example, a computing system can include a computing device and a remote device. The computing device can borrow memory from the remote device over a wired or wireless network. Through the borrowed memory, the computing device and the remote device can collaborate with each other in storing an artificial neural network and in processing based on the artificial neural network. Some layers of the artificial neural network can be stored in the memory loaned by the remote device to the computing device. The remote device can perform the computation of the layers stored in the borrowed memory on behalf of the computing device. When the network connection degrades, the computing device can use an alternative module to function as a substitute of the layers stored in the borrowed memory.

Memory as a service for artificial neural network (ANN) applications

An example memory sub-system includes: a plurality of bank groups, wherein each bank group comprises a plurality of memory banks; a plurality of row buffers, wherein two or more row buffers of the plurality of row buffers are associated with each memory bank; a cache comprising a plurality of cache lines; a processing logic communicatively coupled to the plurality of bank groups and the plurality of row buffers, the processing logic to perform operations comprising: receiving an activate command specifying a row of a memory bank of the plurality of memory banks; fetching data from the specified row to a row buffer of the plurality of row buffers; and copying the data to a cache line of the plurality of cache lines.

Memory device with on-die cache

A wafer-on-wafer bonded memory and logic device can enable high bandwidth transmission of data directly between a memory die and a logic die. Memory devices can be formed on a first wafer. First metal pads can be formed on the first wafer and coupled to the memory devices. The memory devices can be tested via the first metal pads. The first metal pads can be removed from the first wafer. Subsequently, second metal pads on the first wafer can be bonded, via a wafer-on-wafer bonding process, to third metal pads on a second wafer. Each memory device on the first wafer can be aligned with and coupled to a respective logic device on the second wafer.

Testing memory of wafer-on-wafer bonded memory and logic

A memory device includes a memory die bonded to a logic die. A logic die that is bonded to a memory die via a wafer-on-wafer bonding process can receive signals indicative of input data from a global data bus of the memory die and through a bond of the logic die and memory die. The logic die can also receive signals indicative of kernel data from local input/output (LIO) lines of the memory die and through the bond. The logic die can perform a plurality of operations at a plurality of vector-vector (VV) units utilizing the signals indicative of input data and the signals indicative of kernel data.

Signal routing between memory die and logic die for performing operations

Systems, apparatuses, and methods related to data reconstruction based on queue depth comparison are described. To avoid accessing the 'congested' channel, a read command to access the “congested” channel can be executed by accessing the other relatively “idle” channels and utilize data read from the “idle” channels to reconstruct data corresponding to the read command.

Read replacement via data reconstruction based on access patterns

Methods, apparatuses, and systems for in-or near-memory processing are described. Strings of bits (e.g., vectors) may be fetched and processed in logic of a memory device without involving a separate processing unit. Operations (e.g., arithmetic operations) may be performed on numbers stored in a bit-parallel way during a single sequence of clock cycles. Arithmetic may thus be performed in a single pass as numbers are bits of two or more strings of bits are fetched and without intermediate storage of the numbers. Vectors may be fetched (e.g., identified, transmitted, received) from one or more bit lines. Registers of a memory array may be used to write (e.g., store or temporarily store) results or ancillary bits (e.g., carry bits or carry flags) that facilitate arithmetic operations. Circuitry near, adjacent, or under the memory array may employ XOR or AND (or other) logic to fetch, organize, or operate on the data.

Methods for performing processing-in-memory operations, and related systems

Systems, methods and apparatuses to provide memory as a service are described. For example, a borrower device is configured to: communicate with a lender device; borrow an amount of memory from the lender device; expand memory capacity of the borrower device for applications running on the borrower device, using at least the local memory of the borrower device and the amount of memory borrowed from the lender device; and service accesses by the applications to memory via communication link between the borrower device and the lender device.

Inter operating system memory services over communication network connections

Methods, systems, and devices for programmable metadata and related operations are described. A method may include receiving signaling that indicates a set of rules for transitions of states of metadata at a memory device storing the metadata. The memory device may receive a command from a host device associated with a set of data after receiving the set of rules. The memory device may transition metadata associated with the set of data stored at the memory device from a first state to a second state based in part on the set of rules and the command. The memory device may execute the command received from the host device.

Programmable metadata

Methods, systems, and devices for redundant computing across planes are described. A device may perform a computational operation on first data that is stored in a first plane that includes content-addressable memory cells. The first data may be representative of a set of contiguous bits of a vector. The device may perform, concurrent with performing the computational operation on the first data, the computational operation on second data that is stored in a second plane. The second data may be representative of the set of contiguous bits of the vector. The device may read from the first plane and write to the second plane, third data representative of a result of the computational operation on the first data.

Redundant computing across planes

Systems, methods, and apparatus related to data recovery in memory devices. In one approach, a memory device encodes stored data. The memory device reads a codeword from a storage media and determines that a number errors in the codeword exceeds an error correction capability of the memory device. The errors are due, for example, to one or more stuck bits. In response to this determination, one or more data patterns are written to the storage media at the same address from which the codeword is read. The data patterns are read to identify bit locations of the stuck bits. The identified locations are used to correct bit errors of the read codeword that correspond to the identified locations. The corrected code word is sent to a host device (e.g., which requested data from the memory device using a read command).

Data recovery system for memory devices

A system having multiple devices that can host different versions of an artificial neural network (ANN) as well as different versions of a feature dictionary. In the system, encoded inputs for the ANN can be decoded by the feature dictionary, which allows for encoded input to be sent to a master version of the ANN over a network instead of an original version of the input which usually includes more data than the encoded input. Thus, by using the feature dictionary for training of a master ANN there can be reduction of data transmission.

Feature dictionary for bandwidth enhancement

A memory device may be used to implement a Bloom filter. In some examples, the memory device may include a memory array to perform a multiply-accumulate operation to implement the Bloom filter. The memory device may store multiple portions of a reference genetic sequence in the memory array and compare the portions of the reference genetic sequence to a read sequence in parallel by performing the multiply-accumulate operation. The results of the multiply-accumulate operation between the read sequence and the portions of the reference genetic sequence may be used to determine where the read sequence aligns to the reference sequence.

Sequence alignment with memory arrays

A memory device includes a memory die bonded to a logic die via a wafer-on-wafer bond. A controller of the memory device that is coupled to the memory die can activate a row of the memory die. Responsive to activating the row, a sense amplifier stripe of the memory die can latch a first plurality of signals. A transceiver can route a second plurality of signals from the sense amplifier stripe to the logic die.

Signal routing between memory die and logic die

Methods, systems, and devices for memory accessing with auto-precharge are described. For example, a memory system may be configured to support an activate with auto-precharge command, which may be associated with a memory device opening a page of memory cells, latching respective logic states stored by the memory cells at a row buffer, writing logic states back to the page of memory cells, and maintaining the latched logic states at the row buffer (e.g., while maintaining power to latches of the row buffer, after closing the page of memory cells, while the page of memory cells is closed).

Growing community of inventors

Penryn, CA, United States of America

Sean Stephen Eilert