Intel Corporation

Sun Microsystems, Inc.

Tahoe Research, Ltd.

Srikanth T Srinivasan

Haitham H Akkary

Shigeki Tomishima

Jonathan D Pearce

David B Sheffield

Rajesh Sundaram

Matthew C Merten

Ravi Rajwar

Chetan Chauhan

Jeffrey J Cook

Jawad Basit Khan

Justin M Deinlein

Andrey Ayupov

Christopher B Wilkerson

Dz-Ching Ju

Ammon J Christiansen

Abhijit Davare

Deborah T Marr

Sin Sim Tan

Mark Dechene

Anton Alexandrovich Sorokin

Bambang Sutanto

Stephan Jourdan

Eriko Nurvitadhi

Wei Wu

Partha P Tirumalai

Robert S Chappell

Rabin Sugumar

Lily Pao Looi

Bruce Querbach

Asit K Mishra

Tong Li

Desmond A Kirkpatrick

Steven Burns

Chia Yin Kevin Lai

Amit Gandhi

Zeev Sperber

Shih-Lien L Lu

Richard L Coulson

Herbert H Hum

Sriram R Vangal

James Hadley

Yury N Ilin

Christine E Wang

Sean Mirkes

Alexandre J Farcy

Anant S Deval

Sivakumar Radhakrishnan

Bruce A Tennant

Selim Bilgin

Ho-Seop Kim

Jaewoong Sim

Dmitry Petrov

Guy Patkin

Choon Yip Soo

Debbie Marr

Sagi Lahav

Rahul R Kulkarni

Thomas Mullins

Rishiraj A Bheda

Stephan J Jourdan

Technologies for performing in-memory macro operations include a memory having a media access circuitry connected to a memory media. The media access circuitry is to receive a request to perform an in-memory macro operation indicative of a set of multiple in-memory operations. The media access circuitry is also to perform, in response to the request, the in-memory macro operation on data present in the memory media.

Technologies for performing macro operations in memory

An apparatus and method for performing efficient, adaptable tensor operations. For example, one embodiment of a processor comprises: front end circuitry to schedule matrix operations responsive to a matrix multiplication instruction; a plurality of lanes to perform parallel execution of the matrix operations, wherein a lane comprises an arithmetic logic unit to multiply a block of a first matrix with a block of a second matrix to generate a product and to accumulate the product with a block of a third matrix, and wherein the matrix blocks are to be stored in registers within the lane; and broadcast circuitry to broadcast one or more invariant matrix blocks to at least one of different registers within the lane and different registers across different lanes.

Apparatus and method for adaptable and efficient lane-wise tensor processing

Technologies for providing multiple tier memory media management include a memory having a media access circuitry connected to a memory media. The media access circuitry is to receive a request to perform an in-memory compute operation. Additionally, the media access circuitry is to read, in response to the request, data from a memory media region of the memory media, write the read data into a compute media region of the memory, perform, on the data in the compute media region, the in-memory compute operation, write, to the memory media region, resultant data indicative of a result of performance of the in-memory compute operation.

Technologies for providing multiple tier memory media management

In one embodiment, an apparatus includes: a plurality of registers; a first instruction queue to store first instructions; a second instruction queue to store second instructions; a program order queue having a plurality of portions each associated with one of the plurality of registers, each of the portions having entries to store a state of an instruction, the state comprising an encoding of a use of the register by the instruction and a source instruction queue for the instruction; and a dispatcher to dispatch for execution the first and second instructions from the first and second instruction queues based at least in part on information stored in the program order queue, to manage instruction dependencies between the first instructions and the second instructions. Other embodiments are described and claimed.

System, apparatus and method for program order queue (POQ) to manage data dependencies in processor having multiple instruction queues

Technologies for provisioning error-corrected data for use in in-memory compute operations include a memory that includes a memory media having multiple memory partitions and media access circuitry coupled to the memory media. The media access circuitry is to receive a request to perform an in-memory compute operation on data from the memory media. The request specifies a memory partition of the memory media in which the data is located. The media access circuitry reads the data from the memory partition. The media access circuitry performs error correction on the read data to produce error-corrected read data and stores the error-corrected read data in a temporary buffer for access by one or more in-memory compute operations, in addition to the requested in-memory compute operation.

Technologies for providing ECC pre-provisioning and handling for cross-point memory and compute operations

In one embodiment, an apparatus includes: a plurality of execution lanes to perform parallel execution of instructions; and a unified symbolic store address buffer coupled to the plurality of execution lanes, the unified symbolic store address buffer comprising a plurality of entries each to store a symbolic store address for a store instruction to be executed by at least some of the plurality of execution lanes. Other embodiments are described and claimed.

System, apparatus and method for symbolic store address generation for data-parallel processor

Technologies for providing adaptive memory media management include media access circuitry connected to a memory media. The media access circuitry is to receive a request to perform at least one memory access operation to be managed by the media access circuitry. The media access circuitry is further to manage the requested at least one memory access operation, including disabling a memory controller in communication with the media access circuitry from managing the memory media while the at least one requested memory access operation is performed.

Technologies for providing adaptive memory media management

In one embodiment, a reservation station of a processor includes: a plurality of first lanes having a plurality of entries to store information for instructions having in-order dependencies; a variable latency tracking table including a second plurality of entries to store information for instructions having a variable latency; and a scheduler circuit to access a head entry of the plurality of first lanes to schedule, for execution on at least one execution unit, at least one instruction from the head entry of at least one of the plurality of first lanes. Other embodiments are described and claimed.

System, apparatus and method for a hybrid reservation station for a processor

An apparatus and method for efficiently processing invariant operations on a parallel execution engine. For example, one embodiment of a processor comprises: a plurality of parallel execution lanes comprising execution circuitry and registers to concurrently execute a plurality of threads; front end circuitry coupled to the plurality of parallel execution lanes, the front end circuitry to arrange the threads into parallel execution groups and schedule operations of the threads to be executed across the parallel execution lanes, wherein the front end circuitry is to dynamically evaluate one or more variables associated with the operations to determine if one or more conditionally invariant operations will be invariant across threads of a parallel execution group and/or across the parallel execution lanes; a scheduler of the front end circuitry to responsively schedule a shared thread upon a determination that a conditionally invariant operation will be invariant across threads of a parallel execution group and/or across the parallel execution lanes.

Apparatus and method for gang invariant operation optimizations using dynamic evaluation

Technologies for providing a scalable architecture to efficiently perform compute operations in memory include a memory having media access circuitry coupled to a memory media. The media access circuitry is to access data from the memory media to perform a requested operation, perform, with each of multiple compute logic units included in the media access circuitry, the requested operation concurrently on the accessed data, and write, to the memory media, resultant data produced from execution of the requested operation.

Technologies for providing a scalable architecture for performing compute operations in memory

Technologies for providing multiple levels of error correction include a memory that includes media access circuitry coupled to a memory media. The media access circuitry is to read data from the memory media. Additionally, the media access circuitry is to perform, with an error correction logic unit located in the media access circuitry, error correction on the read data to produce error-corrected data.

Technologies for providing multiple levels of error correction

A processor and method are described for a multi-level reservation station. For example, one embodiment of an apparatus comprises: execution circuitry comprising a plurality of functional units to execute a plurality of operations; a reservation station comprising a plurality of entries to store a corresponding plurality of operations to be executed on one or more of the functional units, the reservation station comprising: a first RS level to hold a first subset of the plurality of operations which are ready for execution by one or more functional units or which are expected to be ready for execution by the functional units; a second RS level to hold a second subset of the plurality of operations which are not expected to be ready for execution by the functional units; operation evaluation circuitry to evaluate operations in the first RS level and, responsive to identifying one or more operations which are not expected to be ready for execution, to cause the one or more operations to be moved from the first RS level to the second RS level.

Method and apparatus for a multi-level reservation station with instruction recirculation

An apparatus and method for offloading iterative, parallel work to a data parallel cluster. For example, one embodiment of a processor comprises: a host processor to execute a primary thread; a data parallel cluster coupled to the host processor over a high speed interconnect, the data parallel cluster comprising a plurality of execution lanes to perform parallel execution of one or more secondary threads related to the primary thread; and a data parallel cluster controller integral to the host processor to offload processing of the one or more secondary threads to the data parallel cluster in response to one of the cores executing a parallel processing call instruction from the primary thread.

Apparatus and method for a high throughput parallel co-processor and interconnect with low offload latency

In one embodiment, a cache memory includes: a plurality of data banks, each of the plurality of data banks having a plurality of entries each to store a portion of a cache line distributed across the plurality of data banks; and a plurality of tag banks decoupled from the plurality of data banks, wherein a tag for a cache line is to be assigned to one of the plurality of tag banks. Other embodiments are described and claimed.

Gather-scatter cache architecture having plurality of tag and data banks and arbiter for single program multiple data (SPMD) processor

An apparatus and method for data parallel single program multiple data (SPMD) execution. For example, one embodiment of a processor comprises: instruction fetch circuitry to fetch instructions of one or more primary threads; a decoder to decode the instructions to generate uops; a data parallel cluster (DPC) to execute microthreads comprising a subset of the uops, the DPC further comprising: a plurality of execution lanes to perform parallel execution of the microthreads; an instruction decode queue (IDQ) to store the uops prior to execution; and a scheduler to evaluate the microthreads based on associated variables including instruction pointer (IP) values, the scheduler to gang microthreads into fragments for parallel execution on the execution lanes based on the evaluation.

Architecture and method for data parallel single program multiple data (SPMD) execution

Growing community of inventors

Durham, NC, United States of America

Srikanth T Srinivasan