Complex matrix multiplication operations with data pre-conditioning in a high performance computing architecture

Eichenberger, Alexandre E; Gschwind, Michael K; Gunnels, John A

Title: Complex matrix multiplication operations with data pre-conditioning in a high performance computing architecture

Abstract

Mechanisms for performing a complex matrix multiplication operation are provided. A vector load operation is performed to load a first vector operand of the complex matrix multiplication operation to a first target vector register. The first vector operand comprises a real and imaginary part of a first complex vector value. A complex load and splat operation is performed to load a second complex vector value of a second vector operand and replicate the second complex vector value within a second target vector register. The second complex vector value has a real and imaginary part. A cross multiply add operation is performed on elements of the first target vector register and elements of the second target vector register to generate a partial product of the complex matrix multiplication operation. The partial product is accumulated with other partial products and a resulting accumulated partial product is stored in a result vector register.

Inventors:: Eichenberger, Alexandre E; Gschwind, Michael K; Gunnels, John A

Issue Date:: 2014-02-11

Research Org.:: International Business Machines Corp., Armonk, NY (United States)

Sponsoring Org.:: USDOE

OSTI Identifier:: 1119675

Patent Number(s):: 8650240

Application Number:: 12/542,324

Assignee:: International Business Machines Corporation (Armonk, NY)

Patent Classifications (CPCs):: G - PHYSICS G06 - COMPUTING G06F - ELECTRIC DIGITAL DATA PROCESSING

Show more

G - PHYSICS G06 - COMPUTING G06F - ELECTRIC DIGITAL DATA PROCESSING
G06F17/16 - Matrix or vector computation {, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
G06F9/30036 - {Instructions to perform operations on packed data, e.g. vector operations}
G06F9/30109 - {having multiple operands in a single register}
G06F9/30043 - {LOAD or STORE instructions
G06F9/30014 - {with variable precision}
G06F9/30032 - {Movement instructions, e.g. MOVE, SHIFT, ROTATE, SHUFFLE}

Show less

DOE Contract Number:: B554331

Resource Type:: Patent

Country of Publication:: United States

Language:: English

Subject:: 97 MATHEMATICS AND COMPUTING

Citation Formats


                    Eichenberger, Alexandre E, Gschwind, Michael K, and Gunnels, John A. Complex matrix multiplication operations with data pre-conditioning in a high performance computing architecture.  United States: N. p., 2014. 
        Web.

Copy to clipboard


                    Eichenberger, Alexandre E, Gschwind, Michael K, & Gunnels, John A. Complex matrix multiplication operations with data pre-conditioning in a high performance computing architecture.  United States.

Copy to clipboard


                    Eichenberger, Alexandre E, Gschwind, Michael K, and Gunnels, John A. Tue .  
        "Complex matrix multiplication operations with data pre-conditioning in a high performance computing architecture".  United States.  https://www.osti.gov/servlets/purl/1119675.

Copy to clipboard


                    
@article{osti_1119675,

  title        = {Complex matrix multiplication operations with data pre-conditioning in a high performance computing architecture},

  author       = {Eichenberger, Alexandre E and Gschwind, Michael K and Gunnels, John A},

  abstractNote = {Mechanisms for performing a complex matrix multiplication operation are provided. A vector load operation is performed to load a first vector operand of the complex matrix multiplication operation to a first target vector register. The first vector operand comprises a real and imaginary part of a first complex vector value. A complex load and splat operation is performed to load a second complex vector value of a second vector operand and replicate the second complex vector value within a second target vector register. The second complex vector value has a real and imaginary part. A cross multiply add operation is performed on elements of the first target vector register and elements of the second target vector register to generate a partial product of the complex matrix multiplication operation. The partial product is accumulated with other partial products and a resulting accumulated partial product is stored in a result vector register.},

  doi          = {},

  journal      = {},
number       = ,

  volume       = ,

  place        = {United States},

  year         = {2014},

  month        = {2}

}

Copy to clipboard

Patent:

Save / Share:

Export Metadata

Save to My Library

Works referenced in this record:

Adaptive Strassen and ATLAS's DGEMM: a fast square-matrix multiply for modern high-performance systems
conference, January 2005

D'Alberto, P.; Nicolau, A.
Eighth International Conference on High-Performance Computing in Asia-Pacific Region (HPCASIA'05)
https://doi.org/10.1109/HPCASIA.2005.18

High performance software on Intel Pentium Pro processors or Micro-Ops to TeraFLOPS
conference, January 1997

Greer, Bruce; Henry, Greg
Proceedings of the 1997 ACM/IEEE conference on Supercomputing (CDROM) - Supercomputing '97, p. 1-13
https://doi.org/10.1145/509593.509639

Similar Records in DOE Patents and OSTI.GOV collections:

Matrix multiplication operations with data pre-conditioning in a high performance computing architecture

Patent Eichenberger, Alexandre E; Gschwind, Michael K; Gunnels, John A

Mechanisms for performing matrix multiplication operations with data pre-conditioning in a high performance computing architecture are provided. A vector load operation is performed to load a first vector operand of the matrix multiplication operation to a first target vector register. A load and splat operation is performed to load an element of a second vector operand and replicating the element to each of a plurality of elements of a second target vector register. A multiply add operation is performed on elements of the first target vector register and elements of the second target vector register to generate a partial productmore » « less
Full Text Available
Matrix multiplication operations using pair-wise load and splat operations

Patent Eichenberger, Alexandre E.; Gschwind, Michael K.; Gunnels, John A.; ...

Mechanisms for performing a matrix multiplication operation are provided. A vector load operation is performed to load a first vector operand of the matrix multiplication operation to a first target vector register. A pair-wise load and splat operation is performed to load a pair of scalar values of a second vector operand and replicate the pair of scalar values within a second target vector register. An operation is performed on elements of the first target vector register and elements of the second target vector register to generate a partial product of the matrix multiplication operation. The partial product is accumulatedmore » « less
Full Text Available
System and method for integrating and accessing multiple data sources within a data warehouse architecture

Patent Musick, Charles R [Castro Valley, CA]; Critchlow, Terence [Livermore, CA]; Ganesh, Madhaven [San Jose, CA]; ...

A system and method is disclosed for integrating and accessing multiple data sources within a data warehouse architecture. The metadata formed by the present method provide a way to declaratively present domain specific knowledge, obtained by analyzing data sources, in a consistent and useable way. Four types of information are represented by the metadata: abstract concepts, databases, transformations and mappings. A mediator generator automatically generates data management computer code based on the metadata. The resulting code defines a translation library and a mediator class. The translation library provides a data representation for domain specific knowledge represented in a data warehouse,more » « less
Full Text Available
Cloud object store for archive storage of high performance computing data using decoupling middleware

Patent Bent, John M.; Faibish, Sorin; Grider, Gary

Cloud object storage is enabled for archived data, such as checkpoints and results, of high performance computing applications using a middleware process. A plurality of archived files, such as checkpoint files and results, generated by a plurality of processes in a parallel computing system are stored by obtaining the plurality of archived files from the parallel computing system; converting the plurality of archived files to objects using a log structured file system middleware process; and providing the objects for storage in a cloud object storage system. The plurality of processes may run, for example, on a plurality of compute nodes.more » « less
Full Text Available
Accessing protected data by a high-performance computing cluster

Patent Barnes, Peter; Fedor-Thurman, Daniel; Halliday, Kyle D.

A data protection system is provided that allows applications to access protected data in a way that restricts applications from outputting to unauthorized targets any unprotected data derived from the protected data and that ensures that the applications do not have access to a key that allows access to the unprotected data. The data protection system provides a policy server that may execute on a service node of a high performance computing system and a data encryption process that may execute on each compute node that is allocated to an application or batch job. The policy server maintains policies ofmore » « less
Full Text Available

Similar Records

Title: Complex matrix multiplication operations with data pre-conditioning in a high performance computing architecture

Abstract

Citation Formats

Adaptive Strassen and ATLAS's DGEMM: a fast square-matrix multiply for modern high-performance systems conference, January 2005

High performance software on Intel Pentium Pro processors or Micro-Ops to TeraFLOPS conference, January 1997

Adaptive Strassen and ATLAS's DGEMM: a fast square-matrix multiply for modern high-performance systems
conference, January 2005

High performance software on Intel Pentium Pro processors or Micro-Ops to TeraFLOPS
conference, January 1997