Using MKL, the Intel Math Kernel Library (v11.0.0.079) - ICHEC
Using MKL, the Intel Math Kernel Library (v11.0.0.079) - ICHEC
Using MKL, the Intel Math Kernel Library (v11.0.0.079) - ICHEC
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
•Interface LayerThis layer provides matching between <strong>the</strong> compiled code of an application and <strong>the</strong> threading/computationalcomponents of <strong>the</strong> library. This layer:•provides an LP64 interface.•access to <strong>MKL</strong> ILP64.•deals with <strong>the</strong> way in which different compilers return function values.•a way to map single precision names to double precision names in applications that employ ILP64programming model (e.g. Cray-style naming).Interface Layer Librarieslibmkl_intel_lp64 libmkl_intel_ilp64 libmkl_gf_lp64 libmkl_gf_ilp64 libmkl_intel_sp2dp•Threading LayerThis layer helps <strong>the</strong> threaded <strong>MKL</strong> to co-operate with compiler level threading. This also provides <strong>the</strong>sequential version layer.Threading Layer Librarieslibmkl_intel_thread libmkl_gnu_thread libmkl_pgi_thread libmkl_sequential•Computational LayerThis is <strong>the</strong> heart of <strong>MKL</strong> and has only one variant for any processor/operating system family, Thecomputational layer accommodates multiple architectures through identification of <strong>the</strong> architecture orarchitectural feature and chooses <strong>the</strong> appropriate binary code at execution. <strong>Intel</strong> <strong>MKL</strong> may be thought of as<strong>the</strong> large computational layer that is unaffected by different computational environments. Then, as it has noRTL requirements, RTLs refer not to <strong>the</strong> computational layer but to one of <strong>the</strong> layers above it: <strong>the</strong> Interfacelayer or Threading layer. The most likely case is matching <strong>the</strong> threading layer with <strong>the</strong> RTL layer.Computational Layer Librarieslibmkl_avx libmkl_core libmkl_def libmkl_mc/mc3 libmkl_scalapack_lp64/ilp64libmkl_vml_def/p4n/mc/mc2/mc3/avx/cmptlibmkl_cdft_core2