Intel® MKL 10.3 Bug Fixes

The tables below summarize specific customer issues or feature requests that have been addressed by the indicated product releases. Some items span multiple architectures and/or operating systems and some relate only to a single architecture or operating system.

NOTE: The issues, defects, bug reports, and feature requests described below represent specific issues with specific test cases. An item listed here does not imply that it necessarily applies to your application(s). If your situation does not match the specific test case you may not have experienced the error or problem associated with that update. It is not possible to describe the details of every issue and its specific test case in these tables.
 

Providing a complete description of each item in the list below is impractical. For that reason we ask that you post a message on the Intel MKL user forum, /en-us/forums/intel-math-kernel-library/, and reference this page and the "DPD" number associated with the item of interest, if you have questions regarding the information in these tables.

These tables are updated regularly to correct any errors and omissions.

Intel® MKL 10.3 update 12 (14 Sept 2012)

 

  
DPD200292943Fixed the mkl_?dncsr failure when converting upper triangular dense matrix

Intel® MKL 10.3 update 11 (26 June 2012)

 

  
DPD200283785Fixed the problem of - DFT in MKL with AVX code make the test code hang
DPD200286656Fixed description of rhint of df?integrate1d/df?integrateex1d
DPD2001814253D cluster FFT computes incorrect local buffer size when larger than 2**31
DPD200281980Fixed Conflicting documentation for P?GEMM descriptor in mkl manual
DPD200274859MKL Crash when using dynamic threaded library under Xcode
DPD200275066Description of PARDISO Errors and messages
DPD200262923Pardiso OOC version crashes if no iterative refinement chosen. In-Core version run OK.
DPD200172358Performance degradation of several functions like dgelqf, dgesdd in mkl 10.3.4 compared to mkl 10.1.1
DPD200179713slasd3 and slasd4 performance degradation on Win32
DPD200280082DGELSD returns incorrect MKL ERROR when data contains NaN
DPD200178593Fixed Crash with repeated load and unload of MKL DLL
DPD200285349VSL crashes in a customer application
DPD200283138ddot slow on 8-core 32-bt winxp harpertown without HT
DPD200181490Fixed the problem of SGESVD returning NaNs
DPD200287657Sparse BLAS mkl_dcsrcoo sorts CSR matrix incorrectly for 0-based indexing

 

Intel® MKL 10.3 update 10 (30 April 2012)

 

  
DPD200274946{S|D}STERM description in the Intel MKL reference did not mention that parameter 'd' will be overwritten
DPD200274901PARDISO:division by zero occurs when trying to solve unsymmetric matrices
DPD200272909Trailing underscores ( e.g - "fftwnd_f77_mpi_" ) are missing in the F77 interface wrappers
DPD200180544Performance regression in *SYEV
DPD200270801MT19937/MT2203 generator results when obtained all at once in a single array are different from results obtained in multiple subsets
DPD200274331Performance issue of cblas _?nrm2 on processors supporting Intel Advanced Vector Extensions (Intel AVX)
DPD200276708PARDISO does not solve the diagonal matrix

 

Intel® MKL 10.3 update 9 (18 February 2012)

 

  
DPD200175424IDAMAX can return bad index - (note: can cause heap corruption in other MKL functions. e.g., DGETRF DPD200274016)
DPD200211015Poor scaling of double-precision 2-power 1d FFT on ia32
DPD200267589Updated MKL User Guide to reflect the latest version XCode
DPD200263102Error in local PARDISO version report in 10.3.7 did not change from 10.3.6 - It is still 103000116
DPD200177488DAXPY performance degradation for small sizes when using multiple threads
DPD200225339Performance regression for several functions from 10.3.4 to 10.1.1

 

Intel® MKL 10.3 update 8 (19 December 2011)

 

  
DPD200172281Bug in 1D FFT with DFTI_BACKWARD_SCRAMBLED with N>=16384
DPD200223467Bug in error message for vslSaveStreamF
DPD200231975Heap corruption occurs when calling SetThreadAffinityMask() before any Intel MKL functions (workaround: call mkl_set_num_threads() first)
DPD200261042DftiCommitDescriptor craches for 1D FFT setup when DFTI_NUMBER_OF_TRANSFORMS > 1 and dynamic linking is used

 

Intel® MKL 10.3 update 7 (25 October 2011)

 

  
DPD200225521Conflict from use of MKL_PARDISO as threading domain and module name
DPD200216996Problem with PDTRAN function in case of P=2 and Q=2
DPD200228197PARDISO problem when solving a new system after initial solve with diagonal matrix
DPD200230260PARDISO fails on IA32 for large systems with more than 250 million LU factors

 

Intel® MKL 10.3 update 6 (6 September 2011)

 

  
DPD200218547PARDISO may return #IND values in the solution vector
DPD200219651zgetc2 performance degradation when dynamically linked or on 64-bit architectures
DPD200221246Line 188 in mkl_pblas.h won't compile unless changed to PZTRANC( MKL_INT *m, MKL_INT *n
DPD200221247two unresolved symbols in mkl_scalapack_ilp64.lib(pxerbla.obj)

 

Intel® MKL 10.3 update 5 (4 August 2011)

 

  
DPD200169403Problem with using max value for an upper bound in the trust region solver.
DPD200169895Bug in PARDISO symbolic factorization when solving indefinite non-symmetric double complex matrices
DPD200192242The validity of input values is now checked to prevent the solver from hanging
DPD200210337PARDISO OOC: *.sin and *.sle files remain after execution
DPD200210595mkl_zcsradd produces incorrect results for the array sizes set according to the Manual
DPD200212849DSTEGR function seg faults for many cases when linked dynamically or with LP64 interface
DPD200214225Attempts to convert a CSR matrix to BSR format with mkl_dcsrbsr using JOB(6)=-1 fail with input error message
DPD200214530Run-time problems when mkl_thread_free_buffers() is the first MKL function called on a thread
DPD200218547PARDISO may return #IND values in the solution vector

 

Intel® MKL 10.3 update 4 (12 May 2011)

 

  
DPD200209064mkl_zdnscsr produces incorrect results when estimating memory for conversion of double precision complex dense matrices to the CSR format
DPD200207760Iterative refinement and CGS iterations errantly invoked when partial solving step (Fw, Diag or Bw) is requested in PARDISO
DPD200167242OpenMP settings are ignored by Intel MKL after a single call to DSYRK
DPD200207682Out-of-core (OOC) PARDISO factorizes the first matrix but fails to factorize further matrices in the "same topology" mode
DPD200199308In-core PARDISO generates wrong result with a single test data if number of threads is 32
DPD200167236Syntax error in mklvars.csh
DPD200204522DGESDD returns U and VT outputs that are not orthogonal enough
DPD200168437Performance degradation of VML functions in threaded environment

 

Intel® MKL 10.3 update 3 (29 March 2011)

 

  
DPD200205256ZGESVD in multithreaded MKL produces wrong results
DPD200165165DGETRF returns different results on consecutive runs
DPD200203834Many warnings when building a custom static library based on the Intel MKL library
DPD200196299ScaLAPACK LU test program fails to run on 1024 node cluster
DPD200083045mkl_gmp.h causes compile time error C++ codes
DPD200206509scgemm/dzgemm errors
DPD200199131PDGEMR2D: out of memory error even when using 64-bit libraries
DPD200192508PARDISO is 2x slower than SuperLU for some data
DPD200151246Performance scalability problem for DGEMM on Win32 on 8 threads
DPD200204717Real VML functions outperform the complex VML functions more then 20 times
DPD200165114*TRMV provides wrong results for problems larger than 3000 for multi-threaded cases

 

Intel® MKL 10.3 update 2 (31 January 2011)

 

  
DPD200090962Expanded threading to new cases for multi-dimensional transforms
DPD200133579Solved problem of large 1D descriptors without performance degradation
DPD200198147Calls to blacs_exit before initialization routines no longer causes a crash
DPD200198405VML functions crash or get incorrect results when called from within Linux native threads in ILP64 mode
DPD200199435Performance degradation of VML functions when using native threading and linking DLLs on Windows
DPD200202108Access violation caused by corruption of the PARDISO handle when called multiple times with only a single handle initialization

 

Intel® MKL 10.3 update 1 (3 December 2010)

 

  
DPD200192242Trust Region Solver hangs
DPD200132424Full support for TR single precision solvers
DPD200162063Error in ?LATRD output parameter description in reference manual
DPD200194170Performance degradation in VML functions Add, Sub, Mul in 10.2.5 (and 10.3) for Intel® Pentium® 4 processors on Windows*
DPD200185620DGEQP3 scales poorly on multiple threads

 

Intel® MKL 10.3 (9 November 2010)

 

  
DPD200095498Unclassifiable error in PARDISO when MKL_PARDISO_OOC_MAX_CORE_SIZE=200 is set
DPD200186873clapack_dgbtrf returns the wrong results
DPD200186934mkl_dcsrcoo produces the wrong output if nnz is not initialized explicitly
DPD200187697dss_factor_complex crashed with invalid floating op when /Qtrapuv
DPD200192952Documentation bug in MKL-Lapack95 GESDD
DPD200084190PARDISO OOC will now run parallel code for all supported matrix types
DPD200087174ippGenNumCoresOndie symbol in Intel MKL conflicts with IPP when used simultaneously
DPD200092652Incorrect results for mkl_dcsrcsc when using 0-based indexing
DPD200091197Intel MKL FFT performance deficiency on non-power-two vector sizes
DPD200090409Add Fortran 90 include path in MKL environment setup scripts
DPD200149638Intel MKL doesn't contain CDFT interface in shared libraries
DPD200085280Intel MKL CFFT: Lack of scalability on non power of 2 core-counts
DPD200154738Path variable corrupted with quote characters after running mklvars.bat
DPD200084778Examples for calling PARDISO from Fortran 90

 

 

 
For more complete information about compiler optimizations, see our Optimization Notice.

1 comment

Top
Ying H. (Intel)'s picture

<a href="http://software.intel.com/en-us/articles/intel-mkl-102-fixes-list/">
Intel® MKL 10.2 Bug Fixes
</a><br /><br />

Add a Comment

Have a technical question? Visit our forums. Have site or software product issues? Contact support.