Thread

Vectorization Potential Seedup Calculation

Architecture:          x86_64 (ivybridge with 8 cores)
Compiler Version:   icc 15.0

How does the compiler calculate estimated potential speedup for a loop in vector report? How to find cache sizes? And How does aligned and unaligned access affect potential speedup value? can you please explain with regarding to report below?

Fortran module containing interfaces only - offload programming

What is the way to write a Fortran source file that contains only one module and that only has interface declarations to routines that are callable as either host or offload.

I tried:

module foo
  interface
    subroutine foobar
    !dir$ attributes offload : mic ::foobar
    end subroutine foobar
...
  end interface
end module foo

When I subsequently USE foo elsewhere, I found I must still add:

!dir$ attributes offload : mic :: foobar

In the routine that performs the call

mkl_ddiamv example

I have some problems with using mkl_ddiamv (language C). Can anybody show me a code example of using mkl_ddiamv with initialization of all the input formal parametres? It is desirable to use the number of diagonals more than one. For example, 3-diagonals with offset=[-1 0 1] and size 5 by 5 with all elements are ones.Multiplicated vectors are ones too. In different sources the matrix in special diag format is treated as two-dimensional array or one-dimensional array with row-by-row storage.

I'm back DPD200361383

I am building a Fortran .DLL that contains offloads to MIC.

In the project there is a module containing interfaces to non-exported subroutines and functions. mod_interfaces.f90 defining module mod_interfaces, with interfaces only.

One of the examples are

    subroutine games_check(crames)
    !DIR$ ATTRIBUTES OFFLOAD : mic :: games_check
        use dyn_array_crames
        implicit none
        type(CRAMES_t) :: crames
    end subroutine games_check

That is in the interface file. The source file containing the subroutine has:

Memory retention on Xeon Phi in OpenMP 4.0

Hello,

some time ago, memory retention between offloads on Xeon Phi in OpenMP 4.0 was supported only within a lexically structured scope  ( https://software.intel.com/en-us/forums/topic/520644#comment-1796781 )

Have things changed as of Parallel Studio 2015? Can we offload data from one function and re-use from another?

Thanks!

Andrey

Intel® Inspector Sample Applications

Intel® Inspector is a dynamic memory and threading error checking tool for users developing serial and multithreaded applications on Windows* and Linux* operating systems. This topic discusses sample applications included with the Intel Inspector.
  • Desenvolvedores
  • Linux*
  • Microsoft Windows* (XP, Vista, 7)
  • Microsoft Windows* 8.x
  • C#
  • C/C++
  • Fortran
  • Intel® Parallel Studio XE
  • Intel® Parallel Studio XE Cluster Edition
  • Intel® Parallel Studio XE Professional Edition
  • Intel® System Studio
  • Intel® Inspector
  • Computação paralela
  • Thread
  • Intel® Inspector Glossary

    Intel® Inspector is a dynamic memory and threading error checking tool for users developing serial and multithreaded applications on Windows* and Linux* operating systems.This page defines key terminology used in the product.
  • Desenvolvedores
  • Linux*
  • Microsoft Windows* (XP, Vista, 7)
  • Microsoft Windows* 8.x
  • C#
  • C/C++
  • Fortran
  • Intel® Parallel Studio XE
  • Intel® Parallel Studio XE Cluster Edition
  • Intel® Parallel Studio XE Professional Edition
  • Intel® System Studio
  • Intel® Inspector
  • Computação paralela
  • Thread
  • bug: warning #469 is hidden when function overloaded

    Hello,

    When compiling example below, you will get warning:

    #469: temporary used for initial value of reference to non-const (anachronism)

    and when another overload is added (see comment) warning disappears, but compiler still choose "BOOL" overload instead of "OBJECT". Note that disabling microsoft extensions (/Za) brings warning back.

    We are using x64 compiler for visual studio 2013, version 15.0.1.148 build 20141023 on windows 7. 

    Assine o Thread