Forums
https://software.intel.com/en-us/view/forum-page-default/36938
enJoin the Intel® Parallel Studio XE 2019 Beta Program today
https://software.intel.com/en-us/forums/intel-math-kernel-library/topic/776981
<div class="field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item even" property="content:encoded"><p>Join the <a href="https://software.seek.intel.com/parallel-studio-xe-2019-beta?cid=em-elq-35330&utm_source=elq&utm_medium=email&utm_campaign=35330&elqTrackId=3b825cdfd3a84f6cbba4b4be2b3c3690&elq=74802990a1c346a0962583868573324d&elqaid=35330&elqat=1&elqCampaignId=">Intel® Parallel Studio XE 2019 Beta Program</a> today and—for a limited time—get early access to new features and get an open invitation to tell us what you really think.</p>
<p>We want YOU to tell us what to improve so we can create high-quality software tools that meet your development needs.</p>
<p><a href="https://software.seek.intel.com/LP=19285">Sign Up Now ></a></p>
<p>Top New Features in Intel® Parallel Studio XE 2019 Beta</p>
<ul>
<li>Scale and perform on the path to exascale. Enable greater scalability and improve latency with the latest Intel<sup>®</sup> MPI Library.</li>
<li>Get better answers with less overhead. Focus more fully on useful data, CPU utilization of physical cores, and more using new data-selection support from Intel<sup>®</sup> VTune<sup>™</sup> Amplifier’s Application Performance Snapshot.</li>
<li>Visualize parallelism. Interactively build, validate, and visualize algorithms using Intel® Advisor’s Flow Graph Analyzer.</li>
<li>Stay up-to-date with the latest standards:
<ul>
<li>Expanded C++17 and Fortran 2018 support</li>
<li>Full OpenMP* 4.5 and expanded (partial) support for OpenMP* 5.0 (Preview 2) specification</li>
<li>Python* 3.6 and 2.7</li>
</ul>
</li>
</ul>
<p><strong>New Features in Intel® Math Kernel Library</strong></p>
<ul>
<li>Introduced automatic S/DGEMM JIT capability for small matrix sizes (m,n,k <=16) to improve S/DGEMM performance for Intel® Advanced Vector Extensions 2 (Intel® AVX2) and Intel® Advanced Vector Extensions 512 (Intel® AVX-512) when compiling with MKL_DIRECT_CALL_JIT (threaded usage) or MKL_DIRECT_CALL_SEQ_JIT (sequential usage).</li>
<li>Introduced new functions to JIT (create) optimized S/DGEMM-like matrix multiply kernels for small matrix sizes (m,n,k <=16) for Intel® Advanced Vector Extensions 2 (Intel® AVX2) and Intel® Advanced Vector Extensions 512 (Intel® AVX-512), execute the optimized kernel created using matrices with matching dimensions, and to remove (destroy) the JIT kernel.</li>
<li>Introduce SparseQR functionality</li>
<li>Introduced Extreme{EVD/SVD} functionality</li>
<li>Introduced Multinominal Random Number Generators</li>
</ul>
<p>To learn more, visit <a href="https://software.intel.com/en-us/articles/intel-parallel-studio-xe-2019-beta">Intel<sup>®</sup> Parallel Studio XE 2019 Beta page.</a></p>
<p>Then <a href="https://software.seek.intel.com/LP=19285">sign up</a> to get started.</p>
<p>Note: If you encounter a HTTP ERROR: 404 on the beta registration page, please clear the browser cache and try again.</p>
</div></div></div>Thu, 19 Apr 2018 21:19:22 +0000Shaojuan Z. (Intel)776981 at https://software.intel.comIntel® MKL version 2018 Update 2 is now available
https://software.intel.com/en-us/forums/intel-math-kernel-library/topic/760767
<div class="field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item even" property="content:encoded"><p> </p>
<p>Intel® Math Kernel Library (Intel® MKL) is a highly optimized, extensively threaded, and thread-safe library of mathematical functions for engineering, scientific, and financial applications that require maximum performance.</p>
<p><strong>Intel MKL 2018 Update 2 packages are now ready for download.</strong></p>
<p>Intel MKL is available as part of the <a href="https://software.intel.com/en-us/intel-parallel-studio-xe">Intel® Parallel Studio XE</a> and <a href="https://software.intel.com/en-us/intel-system-studio">Intel® System Studio</a>. Please visit the <a href="https://software.intel.com/en-us/intel-mkl/">Intel® Math Kernel Library Product Page</a>.</p>
<p><strong>Please see What's new in Intel MKL 2018 and in MKL 2018 Update 2 follow this link -</strong> <a href="https://software.intel.com/en-us/articles/intel-math-kernel-library-release-notes-and-new-features">https://software.intel.com/en-us/articles/intel-math-kernel-library-release-notes-and-new-features</a></p>
<p> </p>
</div></div></div>Thu, 22 Mar 2018 01:35:12 +0000Gennady F. (Intel)760767 at https://software.intel.comIntel® MKL 2018 YUM/APT/Conda/Parcel with Cloudera* CDH are available
https://software.intel.com/en-us/forums/intel-math-kernel-library/topic/746150
<div class="field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item even" property="content:encoded"><p>The Intel<sup>®</sup> Math Kernel Library (MKL) 2018 YUM/APT/Conda/ parcel for Cloudera* CDH are available now.</p>
<p>Please follow this link to download and install MKL 2018:</p>
<p>YUM: <a href="https://software.intel.com/en-us/articles/installing-intel-free-libs-and-python-yum-repo">https://software.intel.com/en-us/articles/installing-intel-free-libs-and-python-yum-repo</a></p>
<p>APT: <a href="https://software.intel.com/en-us/articles/installing-intel-free-libs-and-python-apt-repo">https://software.intel.com/en-us/articles/installing-intel-free-libs-and-python-apt-repo</a></p>
<p><a href="https://software.intel.com/en-us/articles/using-intel-distribution-for-python-with-anaconda">conda</a>: <a href="https://software.intel.com/en-us/articles/using-intel-distribution-for-python-with-anaconda">https://software.intel.com/en-us/articles/using-intel-distribution-for-p...</a></p>
<p><a href="https://software.intel.com/en-us/articles/installing-the-intel-distribution-for-python-and-intel-performance-libraries-with-pip-and">pip</a> <a href="https://software.intel.com/en-us/articles/installing-the-intel-distribution-for-python-and-intel-performance-libraries-with-pip-and">https://software.intel.com/en-us/articles/installing-the-intel-distribut...</a></p>
<p><a href="https://software.intel.com/en-us/articles/installing-intel-mkl-cloudera-cdh-parcel">Installing Intel<sup>®</sup> MKL Cloudera* CDH Parcel</a></p>
</div></div></div>Thu, 28 Sep 2017 07:54:00 +0000(name withheld)746150 at https://software.intel.comAnnouncing new tool -- Intel® Math Kernel Library LAPACK Function Finding Advisor
https://software.intel.com/en-us/forums/intel-math-kernel-library/topic/743278
<div class="field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item even" property="content:encoded"><p>The Intel® Math Kernel Library (Intel® MKL) LAPACK domain contains a huge variety of routines. Now, a new tool is provided with a faster method of finding appropriate LAPACK functions in Intel® Math Kernel Library Developer Reference document. This tool would be very useful for Intel® MKL newbies and for users not familiar with LAPACK function naming conventions. By using this tool, users can specify functionality as parameters in drop down lists, descriptions of all functions satisfying the requirements will be shown through this tool. </p>
<p>Please follow this link:<br /><a href="https://software.intel.com/en-us/articles/intel-mkl-function-finding-advisor">Intel® Math Kernel Library LAPACK Function Finding Advisor</a></p>
</div></div></div>Wed, 06 Sep 2017 06:30:56 +0000Fiona Z. (Intel)743278 at https://software.intel.comNNZ coefficients in dss/paridso LU factor
https://software.intel.com/en-us/forums/intel-math-kernel-library/topic/779558
<div class="field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item even" property="content:encoded"><p>Hi,</p>
<p>I was wondering whether there is any way to obtain the number of non-zero coefficients of the matrix factor generated by the MKL function mkl_dss_real? I checked the MKL manual but there seems to be no way to get this number.</p>
<p>Background:</p>
<p>I want to multiply a vector v with a matrix K<sup>-1</sup> , b=K<sup>-1</sup>v, where K is sparse and K<sup>-1</sup> is not constructable. A way to obtain b=K<sup>-1</sup>v is to solve iteratively Kb=v for which I use the mkl_dss solver. In the special setting K is of dimension 2.5Mio x 2.5Mio, is symmetric and positive definite and has 14Mio NNZ coefficients. I understood that the dss_solver uses a LU factorization and subsequently foreward-backward substitution for solving. I also understood that the time complexity of forward/backward substitution is 2o(n<sup>2</sup>). Given the number of NNZ coefficients in K I could make a rough approximation of the number of floating point operations in routine "mkl_dss_solve" and the associated processing time. However, "mkl_dss_solve" needed much more (~x100) processing time. Currenly the only explanation for this observations is that the nnz coefficients in L/U must be much larger than in K.</p>
<p>Any suggestions are welcomed.</p>
<p>Thanks</p>
</div></div></div>Sun, 27 May 2018 05:19:46 +0000may.ka779558 at https://software.intel.comzgetrf/zgetri cannot invert matrix
https://software.intel.com/en-us/forums/intel-math-kernel-library/topic/779457
<div class="field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item even" property="content:encoded"><p>Using code that worked with previous versions (Intel 12.1 OpenMPI 1.4.4 MKL 10.2.5.035 on CentOS 6) of MKL, zgetrf/zgetri cannot invert a matrix that it previously inverted.</p>
<p>Could the newer intel modules be changing how the arguments are accepted for lapack functions?</p>
<p>The pivot index is determined by the output of zgetrf( .. , ipvt , ..) which is then fed into zgetri as input similarly. I believe this is the way to do it.</p>
<p>I have just tested a sample program that only does the zgetrf and zgetri portions and I believe now that this is where the problem lies. I have tested two sample programs, one with a relevant 8x8 matrix that I calculate in my program, and one made-up example 8x8 matrix that I checked has an inverse.</p>
<p>Both programs run fine using the old modules, compiling with mpif90 etc as I do now when I am taking measurements. I get no runtime errors or error outputs from the zgetrf and zgetri subroutines. However, when using the new modules (intel2018), I do get an error output from the zgetri and zgetrf subroutines for both sample matrices. This error output, labeled "INFO" by default, rerturn a value of 1. To save you some searching, INFO should return 0 if successful (zgetrf factorizes the matrix to be inverted, zgetri actually inverts it). However if INFO = i > 0, then U(i,i) = 0 and the matrix is singular. This means that something that I'm doing with the new modules is causing a verifiable invertible matrix to be considered non-invertible, or singular.</p>
<p>Now, how this causes a segfault in the program at large I'm not 100% sure, it doesn't do so in my small sample program. However it seems very likely that since I do not suppress the output of the subroutines when it returns something other than a successful exit value for INFO, that the output of zgetri in this case is garbage that eventually causes some segfault error when it eventually gets multiplied by other matrices and is solved by zgetrs subroutine later on. I can verify that the output of zgetri when INFO=1 is nonsense from my small tests.</p>
<p>So to recap, I have found that using the new modules causes zgetri and zgetrf subroutines to return an error for the same, invertible matrices. This problem may also occur for the subroutine zgetrs, but I haven't tested it. In any case, the return value after the error then probably causes the segfault. The question remains, why is zgetrf/zgetri not working with the new modules?</p>
<p> </p>
<p>Here is sample program (pgrmcheck_cond_zgetri.f90) which reproduces the problem:</p>
<p>! placeholder<br />
!mpiifort -o pgrmcheck.x -O3 pgrmcheck_cond_zgetrf.f90 -i8<br />
!-I${MKLROOT}/include/intel64/ilp64 -I${MKLROOT}/include<br />
!${MKLROOT}/lib/intel64/libmkl_lapack95_ilp64.a -L${MKLROOT}/lib/intel64<br />
!-lmkl_scalapack_ilp64 -lmkl_intel_ilp64 -lmkl_intel_thread -lmkl_core<br />
!-lmkl_blacs_intelmpi_ilp64 -liomp5 -lpthread -lm -ldl</p>
<p> PROGRAM dyncond_check<br />
IMPLICIT NONE<br />
include 'mpif.h'</p>
<p>
complex,DIMENSION(8) :: rightvec,work<br />
integer,dimension(8) ::ipvt<br />
complex,DIMENSION(8,8) :: tmatrixN1<br />
integer:: error1,error2,i,j</p>
<p>
tmatrixN1=0</p>
<p> do i=1,8<br />
do j=1,8</p>
<p> if (j<(i+1)) tmatrixN1(i,j)=i+(j-1)*8<br />
enddo<br />
enddo</p>
<p> call mkl_set_num_threads(1)</p>
<p>
call zgetrf(8,8,tmatrixN1,8,ipvt,error1)<br />
print *,error1<br />
call zgetri(8,tmatrixN1,8,ipvt,work,8,error2)<br />
print *,error2<br />
do i=1,8<br />
do j=1,8<br />
! print *,tmatrixN1(i,j)<br />
enddo<br />
enddo</p>
<p>end program</p>
<p> </p>
<p>Compiled with:</p>
<p>### Command to compile for intel 2018 modules (Not functioning) ###</p>
<p>module load intel/cluster/2018<br />
mpiifort -o pgrmcheck2.x -O3 pgrmcheck_cond_zgetri.f90 -i8 -I${MKLROOT}/include/intel64/ilp64 -I${MKLROOT}/include ${MKLROOT}/lib/intel64/libmkl_lapack95_ilp64.a -L${MKLROOT}/lib/intel64 -lmkl_scalapack_ilp64 -lmkl_intel_ilp64 -lmkl_intel_thread -lmkl_core -lmkl_blacs_intelmpi_ilp64 -liomp5 -lpthread -lm -ldl<br />
./pgrmcheck2.x</p>
<p>###</p>
<p>### Command to compile for (old) intel 12.1 modules (functioning) ###</p>
<p>module load intel/12.1 ompi/1.4.4/intel mkl/10.2.5.035<br />
mpif90 -o pgrmcheck2.x -O3 -r8 pgrmcheck_cond_zgetri.f90 -L/soft/intel/mkl/10.2.1.017/lib/em64t -lmkl_lapack -lmkl_intel_thread -lmkl_core -lguide -lmkl_intel_lp64<br />
./pgrmcheck2.x</p>
<p>###</p>
<p>
Thanks!</p>
<p> </p>
</div></div></div>Thu, 24 May 2018 05:02:53 +0000Swartz, Brent779457 at https://software.intel.comPardiso memory problem with Visual Fortran
https://software.intel.com/en-us/forums/intel-math-kernel-library/topic/779388
<div class="field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item even" property="content:encoded"><p>I use mkl with Visual Fortran Compiler XE 14.0.0.103. </p>
<p>I employ the pardiso function for a program that involves solution of a system of linear equations: [A]{x} = {b}, where [A], {b} are given and we want to find vector {x}. My code has to solve this system of equations repeatedly (for many iterations), for different values of [A], {b}. The values of [A], {b} at each iteration depends on the value of {x} from the previous iteration, so the algorithm is something like this:</p>
<p>Initialize {x}</p>
<p>do i = 1,Niter</p>
<p>[Find [A], {b}, given {x}]</p>
<p>[Solve [A]{x} = {b} and find updated {x}]</p>
<p>end do !i</p>
<p>My code has encountered a SERIOUS problem with memory management. Specifically, I see a continuous increase in the amount of memory used, until my computer runs out of memory and the program aborts. I do not have any dynamic memory allocation in my code (I do not use pointers), so I believe that the problem is due to mkl. </p>
<p>I tried to include the line:</p>
<p>call mkl_free_buffers()</p>
<p>but this did not help in any way. I found some posts in this forum with similar comments, but I did not find anything helpful. Any help on this issue would be greatly appreciated!!</p>
</div></div></div>Tue, 22 May 2018 23:15:56 +0000Ioannis K.779388 at https://software.intel.comCan't start MKL pardiso in Mac OSX!
https://software.intel.com/en-us/forums/intel-math-kernel-library/topic/779348
<div class="field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item even" property="content:encoded"><p>Hi, there is command line for the build an Intel example pardiso_sym_f90.f90</p>
<p>ifort pardiso_sym_f90.f90 -I/opt/intel/mkl/include /opt/intel/mkl/lib/libmkl_core.a /opt/intel/mkl/lib/libmkl_intel_ilp64.a /opt/intel/mkl/lib/libmkl_intel_thread.a /opt/intel/lib/libiomp5.a</p>
<p>Compiling and linking was successful. However, a running output file (a.out) was not a beautiful so, </p>
<p> Reordering completed ... </p>
<p> The following ERROR was detected: -1</p>
<p> The following ERROR on release stage was detected: -1</p>
<p> </p>
<p>Thank you for any help!</p>
<p>Malik.</p>
<p> </p>
<p> </p>
<p> </p>
<p> </p>
<p> </p>
</div></div></div>Tue, 22 May 2018 07:28:01 +0000Malik M.779348 at https://software.intel.comPSGESVD: Illegal parameter 19
https://software.intel.com/en-us/forums/intel-math-kernel-library/topic/779297
<div class="field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item even" property="content:encoded"><p>Hello,</p>
<p>I want to perform SVD on a matrix in parallel(C). But I am facing issues in psgesvd. When I try to perform svd of a matrix of size 4750*4750, I get an error saying that:</p>
<p>"{ 0, 0}: On entry to<br />
PSGESVD parameter number 19 had an illegal value"</p>
<p>Parameter 19 is 'lwork' for which a make a query first which returns the minimum size of 'work' array required. The svd works fine for smaller matrices but big matrices give an error. Can anyone please help me with this as to what is the reason for this error. The value of lwork in this case is 45362800.</p>
<p>When I use serial "sgesvd" I don't get any such issue and the results are perfectly correct. But I am not sure what is causing an issue in this case.</p>
<p>Thanks in advance!</p>
<p>-Shailesh Tripathi</p>
</div></div></div>Sun, 20 May 2018 10:16:02 +0000Tripathi, Shailesh779297 at https://software.intel.comMistake in the mkl_dnn documentation
https://software.intel.com/en-us/forums/intel-math-kernel-library/topic/779224
<div class="field field-name-body field-type-text-with-summary field-label-hidden"><div class="field-items"><div class="field-item even" property="content:encoded"><p>Hi Everybody,</p>
<p>i have been using the intel mkl_dnn library for convolutions and batch normalization and have noticed that there is a simple spelling mistake in the documentation that took me around 2 hours to find in my code.</p>
<p><em>dnnResourceScaleShift</em> | Scale and shift data.</p>
<p><em>dnnResourceScaleShift</em> | Gradient with respect to scale and shift.</p>
<p>The second enum should be <em>dnnResourceDiffScaleShift</em> if i am not mistaken?</p>
<p><a href="https://software.intel.com/en-us/mkl-developer-reference-c-enumerated-types">https://software.intel.com/en-us/mkl-developer-reference-c-enumerated-types</a></p>
<p>Thats all.</p>
</div></div></div>Fri, 18 May 2018 12:54:17 +0000schlieder, lennart779224 at https://software.intel.com