Intel® Distribution for LINPACK* Benchmark
affinity mask
aligning data, example
architecture support
BLAS
Calling LAPACK, BLAS, and CBLAS Routines from C/C++ Language Environments
Fortran 95 Interfaces to LAPACK and BLAS
OpenMP* Threaded Functions and Problems
calling routines from C
Fortran 95 interface to
OpenMP* threaded routines
C interface to LAPACK, use of
C, calling LAPACK, BLAS, CBLAS from
C/C++, Intel® MKL complex types
calling
Calling BLAS Functions that Return the Complex Values in C/C++ Code
Calling LAPACK, BLAS, and CBLAS Routines from C/C++ Language Environments
BLAS functions from C
CBLAS interface from C
complex BLAS Level 1 function from C
complex BLAS Level 1 function from C++
Fortran-style routines from C
CBLAS interface, use of
Cluster FFT
Managing Performance of the Cluster Fourier Transform Functions
Managing Invalid Input Checking in LAPACKE Functions
Linking with oneMKL Cluster Software
environment variable for
Managing Performance of the Cluster Fourier Transform Functions
Managing Invalid Input Checking in LAPACKE Functions
linking with
managing performance of
Managing Performance of the Cluster Fourier Transform Functions
Managing Invalid Input Checking in LAPACKE Functions
cluster software, Intel® MKL
cluster software, linking with
Linking with oneMKL Cluster Software
Examples of Linking for Clusters
commands
linking examples
Cluster Sparse Solver, linking with
code examples, use of
coding
Coding Tips
Coding Techniques
data alignment
techniques to improve performance
compilation, Intel® MKL version-dependent
compiler run-time libraries, linking with
compiler-dependent function
complex types in C and C++, Intel® MKL
computation results, consistency
computational libraries, linking with
conditional compilation
configuring Eclipse* CDT
Configuring Your Integrated Development Environment to Link with Intel® oneAPI Math Kernel Library
Configuring the Eclipse* IDE CDT to Link with oneMKL
consistent results
conventions, notational
custom shared object
Building Custom Shared Objects
Composing a List of Functions
Specifying Function Names
building
composing list of functions
specifying function names
data alignment, example
denormal number, performance
direct call, to Intel® Math Kernel Library computational kernels
directory structure
High-level Directory Structure
Appendix C: Directory Structure in Detail
high-level
in-detail
dispatch Intel® architectures, configure with an environment variable
dispatch, new Intel® architectures, enable with an environment variable
Eclipse* CDT
Configuring Your Integrated Development Environment to Link with Intel® oneAPI Math Kernel Library
Configuring the Eclipse* IDE CDT to Link with oneMKL
configuring
Configuring Your Integrated Development Environment to Link with Intel® oneAPI Math Kernel Library
Configuring the Eclipse* IDE CDT to Link with oneMKL
Enter index keyword
environment variables
oneMKL-specific Environment Variables for OpenMP Threading Control
Modulefiles to Set Environment Variables
Scripts to Set Environment Variables
Instruction Set Specific Dispatching on Intel® Architectures
MKL_NUM_STRIPES
Managing Behavior of Function Domains with Environment Variables
Managing Behavior of the Intel® oneAPI Math Kernel Library with Environment Variables
Managing Performance of the Cluster Fourier Transform Functions
for threading control
setting for specific architecture and programming interface
Modulefiles to Set Environment Variables
Scripts to Set Environment Variables
to control dispatching for Intel® architectures
to control threading algorithm for ?gemm
to enable dispatching of new architectures
to manage behavior of function domains
to manage behavior of Intel® Math Kernel Library with
to manage performance of cluster FFT
examples, linking
Examples of Linking for Clusters
Linking Examples
for cluster software
general
FFT interface
OpenMP* Threaded Functions and Problems
OpenMP* threaded problems
FFTW interface support
Fortran 95 interface libraries
function call information, enable printing
header files, Intel® MKL
heterogeneity
Overview of the Intel® Distribution for LINPACK* Benchmark
of Intel® Distribution for LINPACK* Benchmark
heterogeneous cluster
Heterogeneous Support in the Intel® Distribution for LINPACK* Benchmark
support by Intel® Distribution for LINPACK* Benchmark for Clusters
high-bandwidth memory, use in Intel® Math Kernel Library
HT technology, configuration tip
ILP64 programming, support for
improve performance, for matrices of small sizes
include files, Intel® MKL
information, for function call , enable printing
installation, checking
Shared Library Versioning
Checking Your Installation
Intel® Distribution for LINPACK* Benchmark
Overview of the Intel® Distribution for LINPACK* Benchmark
heterogeneity of
Intel® Distribution for LINPACK* Benchmark for Clusters
Heterogeneous Support in the Intel® Distribution for LINPACK* Benchmark
heterogeneous support
Intel® Hyper-Threading Technology, configuration tip
Intel® Optimized High Performance Conjugate Gradient Benchmark
getting started
overview
Intel® Threading Building Blocks, functions threaded with
interface
Linking with Fortran 95 Interface Libraries
Using the ILP64 Interface vs. LP64 Interface
Fortran 95, libraries
LP64 and ILP64, use of
interface libraries and modules, Intel® MKL
interface libraries, linking with
kernel, in Intel® Math Kernel Library, direct call to
language interfaces support
language-specific interfaces
Using Language-Specific Interfaces with Intel® oneAPI Math Kernel Library
interface libraries and modules
LAPACK
Calling LAPACK, BLAS, and CBLAS Routines from C/C++ Language Environments
Fortran 95 Interfaces to LAPACK and BLAS
OpenMP* Threaded Functions and Problems
Coding Techniques
C interface to, use of
calling routines from C
Fortran 95 interface to
OpenMP* threaded routines
performance of packed routines
layers, Intel® MKL structure
libraries to link with
Selecting Libraries to Link with
computational
interface
run-time
system libraries
threading
link tool, command line
link-line syntax
linking examples
Examples of Linking for Clusters
Linking Examples
cluster software
general
linking with
Linking with Compiler Run-time Libraries
Linking with Computational Libraries
Linking with Interface Libraries
Linking with System Libraries
Linking with Threading Libraries
compiler run-time libraries
computational libraries
interface libraries
system libraries
threading libraries
linking, quick start
linking, Web-based advisor
LINPACK benchmark
memory functions, redefining
memory management
memory renaming
memory, high-bandwidth, use in Intel® Math Kernel Library
message-passing interface
Using a Custom Message-Passing Interface
Interaction with the Message-passing Interface
custom, usage
Intel® Math Kernel Library interaction with
mixed-language programming
module, Fortran 95
MPI
Using a Custom Message-Passing Interface
Interaction with the Message-passing Interface
custom, usage
Intel® Math Kernel Library interaction with
multi-core performance
notational conventions
number of threads
Changing the Number of OpenMP* Threads at Run Time
Setting the Number of Threads Using an OpenMP* Environment Variable
MKL_DYNAMIC
Setting the Number of OpenMP* Threads
Techniques to Set the Number of Threads
changing at run time
changing with OpenMP* environment variable
Intel® MKL choice, particular cases
setting for cluster
techniques to set
numerically reproducible results
OpenMP* threaded functions
OpenMP* threaded problems
parallel performance
parallelism, of Intel® MKL
performance
Managing Multi-core Performance
Operating on Denormals
multi-core
with denormals
with subnormals
performance improvement, for matrices of small sizes
performance, of Intel® MKL, improve on specific processors
results, consistent, obtaining
results, numerically reproducible, obtaining
ScaLAPACK, linking with
SDL
Dynamically Selecting the Interface and Threading Layer
Using the Single Dynamic Library
Single Dynamic Library
Dynamically Selecting the Interface and Threading Layer
Using the Single Dynamic Library
structure
High-level Directory Structure
Appendix C: Directory Structure in Detail
Layered Model Concept
high-level
in-detail
model
support, technical
supported architectures
syntax, link-line
system libraries, linking with
technical support
thread safety, of Intel® MKL
threaded functions, with Intel® Threading Building Blocks
threading control, Intel® MKL-specific
threading libraries, linking with
unstable output, getting rid of
Vector Mathematics
Setting the Default Mode of Vector Math with an Environment Variable
default mode, setting with environment variable
environment variable to set default mode
verbose mode, of Intel® MKL