ninner is now small (16) such that kernel runs at peek performance ...

INLINE

Attention: A,B are overwritten !!

Barycentric (b0,b1,b2,b3) node order for a VTK Lagrange tetrahedron of degree o (corners, six edges, four faces via the triangle ordering, interior te

Compoutes B <--- B A^{-1} (trans = 'N') Compoutes B <--- B A^{-T} (trans = 'T') Compoutes B <--- B A^{-H} (trans = 'H') // trans

Generate a structured 2D surface mesh in 3D Parameters ---------- quads : bool If True, a quadrilateral mesh is generated. If

1/2 x^T A x + b x + c

Hm-1 * ... * H1 * H0 * m2

micro-kernels for A^t B pa+i * pb+j -> C 3x16

scale every row from a ..

from bla

---------------------------------------------------------------------- the Abaqus deck parser --------------------------------------------------------