Functions | |
| __device__ void | cuSU2plaq (Complex_f *u11t, Complex_f *u12t, Complex_f *Sigma11, Complex_f *Sigma12, unsigned int *iu, const unsigned int i, const unsigned short mu, const unsigned short nu) |
| Calculates the SU2 plaquette. | |
| template<typename T> | |
| __device__ void | cuByGenLeft (T a[nc], const unsigned short gen) |
| Multiply leaf (or part of one) by generator from left. | |
| template<typename T> | |
| __device__ void | cuByGenRight (T a[nc], const unsigned short gen) |
| Multiply leaf (or part of one) by generator from right. | |
| template<typename T> | |
| __device__ void | Half_Leaf (complex< T > Leaves[nc], complex< T > *u11t, complex< T > *u12t, complex< T > a[nc], unsigned int *iu, unsigned int *id, const unsigned int i, const unsigned short mu, const unsigned short nu, const unsigned short leaf) |
| Calculates the first half of the leaf for a clover term. We split it so that the force term can reuse the first half of the leaf. | |
| template<typename T> | |
| __device__ void | Leaf (complex< T > *u11t, complex< T > *u12t, complex< T > Leaves[nc], unsigned int *iu, unsigned int *id, unsigned int i, const unsigned short mu, const unsigned short nu, const unsigned short leaf) |
| Calculates a leaf for a clover term. | |
| template<typename T> | |
| __device__ void | cuGLeft (T out[4], const T G[2], const T X[4]) |
| Multiplies \( X_{\mu\nu}\) by a gauge field from the left. | |
| template<typename T> | |
| __device__ void | cuGRight (T out[4], const T G[2], const T X[4]) |
| Multiplies \( X_{\mu\nu}\) by a gauge field from the right. | |
| template<typename T> | |
| __device__ void | cuGSandwich (T out[4], T tmp[4], const T Gl[2], const T X[4], const T Gr[2]) |
| Multiplies \( X_{\mu\nu}\) by a gauge field from the left and the right. | |
| __device__ void | GetBilinear (Complex_f Z[nc *nc], Bilinear_a Xmn, unsigned int ind) |
| Loads the compacted bilinear form into a \(2\times2\) complex valued matrix. | |
| template<typename T, unsigned int bsize> | |
| __device__ void | warpReduce_sum (volatile T *sdata, const unsigned int tid) |
| Performs a warp reduction for sum. | |
CUDA Device code.