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 11.0 update 3 (28 Mar 2013)
| DPD200335246 | SVD generates segfault in multithreading mode - http://software.intel.com/en-us/articles/svd-multithreading-bug-in-mkl |
| DPD200325627 | Several tests in SSTS crashed at HSW 64bits - rootcaused to BLAS |
| DPD200326851 | Stack corruption in mkl_dcscmv when input matrix is a 1x2 sparce matrix |
| DPD200241328 | Degraded HPCC DGEMM bm performance on 11/12 MPI ranks/socket on IVB 12-core node |
| DPD200316407 | The DFL routines should return the relevant error code in the case of violation of boundary conditions. |
| DPD200316016 | Fixed mkl_scsrgemv error with one test data |
| DPD200240289 | Fixed MKL Pardiso crash with single precision complex matrix |
| DPD200316391 | Fixed segmentation problem of mkl_dcsrmultcsr with transposed case (transa='t') |
| DPD200319887 | Fixed mkl_scscmv crash when #threads >= 2 |
| DPD200322909 | Fixed NEi - Pardiso giving wrong results when the hybrid mode and #threads > 1 |
| DPD200319693 | Fixed - Pardiso F90 API doesn’t support vector solution and right-hand side vectors like 2D arrays |
| DPD200317411 | vslLoadStreamF (from v.11.0 u1) produces Segmentation fault while reading the stream which was saved with 10.1 version MKL |
| DPD200323601 | Access Violation in mkl_vml_mc3 with latest MKL (dfdInterpolate1D) |
| DPD200316524 | xerbla and mkl_progress are missing from libmkl_rt.dll |
| DPD200311017 | cluster FFTW wrapper problem |
Intel® MKL 11.0 update 2 (1 Feb 2012)
| DPD200303268 | mkl_dcsmm performance slower than reference openMP code on Sandy Bridge |
| DPD200236783 | Failure of Poisson solver for some rectangular meshes |
| DPD200305099 | IALENV returning wrong results on 64 bit systems |
| DPD200307634 | DS: mkl_ddnscsr I/ILP64 behavior |
| DPD200297382 | PARDISO crashes in OOC mode with CGS preconditioning |
| DPD200237461 | Fixed slower performance of mkl_dcsmm compared to openMP code on Intel® Xeon Phi™ coprocessor |
| DPD200237753 | Fixed Multi-threaded DSS crashes on Windows |
| DPD200303385 | Fixed the problem of Pardiso crashing randomly when trying to solve a particular SLAE |
| DPD200300091 | Fixed the failure of OOC PARDISO 32-bit mode - *** Error in PARDISO (reordering_phase) error_num= 18 |
| DPD200298622 | Implemented the new function that will return the maximum amount of memory consumed by mkl’s routines during computations stages |
| DPD200236450 | Added mkl_realloc and mkl_calloc to interface layer |
| DPD200291920 | Fixed PDGEMM error with some problem size |
| DPD200206100 | The functions ?(or/un)gqr are now threaded |
| DPD200306358 | Fixed the problem of dss_reorder crash on win7, 64 bit |
Intel® MKL 11.0 update 1 (4 Oct 2012)
| DPD200233034 | cblas_sgemm parallelized for tall and thin matrices on AVX systems |
| DPD200177279 | Fixed DTRSM performance degradation for greater than 40 threads on 80 core system |
| DPD200290222 | Fixed signal SIGTRAP Trace/breakpoint trap |
| DPD200234867 | SGEMV single thread performance degradation for AVX is Fixed |
| DPD200292360 | Provided a version of SHOC GEMM that matches performance of WAMM GEMM |
| DPD200237177 | RCI ISS functions were missing on MIC. They have been implemented now |
| DPD200295622 | fixed sgemm performance issue within multiple process and 16 threading |
| DPD200235448 | fixed the problem of mkl_mem_stat() returning non zero values after freeing all memory |
| DPD200297326 | Included ?lascl2 function definition available into header file (mkl_lapack.h) |
Intel® MKL 11.0 (4 Sept 2012)
| DPD200270265 | When PARDISO runs out of memory it should report an error code but shouldnt stop/hang/crash |
| DPD200288598 | Fixed bug in Akima cubic spline implementation.Fixed incorrect secants values calculations for points x(2) and x(n-1) of the partition x(1), x(2), … x(n). |
| DPD200288851 | Fixed the problem of constrained bounds with TR Solver |
| DPD200292388 | Fixed the problem of Pardiso CNR returning different results when SSE4.1 ISA is used |
| DPD200293577 | Fixed the error "MKL FATAL ERROR: Cannot load <mkl-thread-layer>" when launching the test ( dgesv ) on MIC system with 2 cards |
| DPD200180020 | Fixed Pardiso failure to factorize complex symmetric system |