libflame revision_anchor
Functions
FLA_Her2k_ln_blk_var3.c File Reference

(r)

Functions

FLA_Error FLA_Her2k_ln_blk_var3 (FLA_Obj alpha, FLA_Obj A, FLA_Obj B, FLA_Obj beta, FLA_Obj C, fla_her2k_t *cntl)
 

Function Documentation

◆ FLA_Her2k_ln_blk_var3()

FLA_Error FLA_Her2k_ln_blk_var3 ( FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B,
FLA_Obj  beta,
FLA_Obj  C,
fla_her2k_t cntl 
)
14{
15 FLA_Obj AT, A0,
16 AB, A1,
17 A2;
18
19 FLA_Obj BT, B0,
20 BB, B1,
21 B2;
22
23 FLA_Obj CTL, CTR, C00, C01, C02,
24 CBL, CBR, C10, C11, C12,
25 C20, C21, C22;
26
27 dim_t b;
28
30 FLA_Cntl_sub_scalr( cntl ) );
31
32 FLA_Part_2x1( A, &AT,
33 &AB, 0, FLA_TOP );
34
35 FLA_Part_2x1( B, &BT,
36 &BB, 0, FLA_TOP );
37
38 FLA_Part_2x2( C, &CTL, &CTR,
39 &CBL, &CBR, 0, 0, FLA_TL );
40
41 while ( FLA_Obj_length( AT ) < FLA_Obj_length( A ) ){
42
44
46 /* ** */ /* ** */
47 &A1,
48 AB, &A2, b, FLA_BOTTOM );
49
51 /* ** */ /* ** */
52 &B1,
53 BB, &B2, b, FLA_BOTTOM );
54
55 FLA_Repart_2x2_to_3x3( CTL, /**/ CTR, &C00, /**/ &C01, &C02,
56 /* ************* */ /* ******************** */
57 &C10, /**/ &C11, &C12,
58 CBL, /**/ CBR, &C20, /**/ &C21, &C22,
59 b, b, FLA_BR );
60
61 /*------------------------------------------------------------*/
62
63 /* C10 = C10 + A1 * B0' */
65 alpha, A1, B0, FLA_ONE, C10,
66 FLA_Cntl_sub_gemm1( cntl ) );
67
68 /* C21 = C21 + B2 * A1' */
70 alpha, B2, A1, FLA_ONE, C21,
71 FLA_Cntl_sub_gemm2( cntl ) );
72
73 /* C11 = C11 + A1 * B1' + B1 * A1' */
75 alpha, A1, B1, FLA_ONE, C11,
76 FLA_Cntl_sub_her2k( cntl ) );
77
78 /*------------------------------------------------------------*/
79
81 A1,
82 /* ** */ /* ** */
83 &AB, A2, FLA_TOP );
84
86 B1,
87 /* ** */ /* ** */
88 &BB, B2, FLA_TOP );
89
90 FLA_Cont_with_3x3_to_2x2( &CTL, /**/ &CTR, C00, C01, /**/ C02,
91 C10, C11, /**/ C12,
92 /* ************** */ /* ****************** */
93 &CBL, /**/ &CBR, C20, C21, /**/ C22,
94 FLA_TL );
95
96 }
97
98 return FLA_SUCCESS;
99}
FLA_Error FLA_Gemm_internal(FLA_Trans transa, FLA_Trans transb, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, FLA_Obj beta, FLA_Obj C, fla_gemm_t *cntl)
Definition FLA_Gemm_internal.c:16
FLA_Error FLA_Her2k_internal(FLA_Uplo uplo, FLA_Trans trans, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, FLA_Obj beta, FLA_Obj C, fla_her2k_t *cntl)
Definition FLA_Her2k_internal.c:16
FLA_Error FLA_Scalr_internal(FLA_Uplo uplo, FLA_Obj alpha, FLA_Obj A, fla_scalr_t *cntl)
Definition FLA_Scalr_internal.c:16
FLA_Obj FLA_ONE
Definition FLA_Init.c:18
FLA_Error FLA_Cont_with_3x3_to_2x2(FLA_Obj *ATL, FLA_Obj *ATR, FLA_Obj A00, FLA_Obj A01, FLA_Obj A02, FLA_Obj A10, FLA_Obj A11, FLA_Obj A12, FLA_Obj *ABL, FLA_Obj *ABR, FLA_Obj A20, FLA_Obj A21, FLA_Obj A22, FLA_Quadrant quadrant)
Definition FLA_View.c:304
FLA_Error FLA_Part_2x2(FLA_Obj A, FLA_Obj *A11, FLA_Obj *A12, FLA_Obj *A21, FLA_Obj *A22, dim_t mb, dim_t nb, FLA_Quadrant quadrant)
Definition FLA_View.c:17
FLA_Error FLA_Cont_with_3x1_to_2x1(FLA_Obj *AT, FLA_Obj A0, FLA_Obj A1, FLA_Obj *AB, FLA_Obj A2, FLA_Side side)
Definition FLA_View.c:428
FLA_Error FLA_Repart_2x1_to_3x1(FLA_Obj AT, FLA_Obj *A0, FLA_Obj *A1, FLA_Obj AB, FLA_Obj *A2, dim_t mb, FLA_Side side)
Definition FLA_View.c:226
FLA_Error FLA_Part_2x1(FLA_Obj A, FLA_Obj *A1, FLA_Obj *A2, dim_t mb, FLA_Side side)
Definition FLA_View.c:76
dim_t FLA_Obj_length(FLA_Obj obj)
Definition FLA_Query.c:116
FLA_Error FLA_Repart_2x2_to_3x3(FLA_Obj ATL, FLA_Obj ATR, FLA_Obj *A00, FLA_Obj *A01, FLA_Obj *A02, FLA_Obj *A10, FLA_Obj *A11, FLA_Obj *A12, FLA_Obj ABL, FLA_Obj ABR, FLA_Obj *A20, FLA_Obj *A21, FLA_Obj *A22, dim_t mb, dim_t nb, FLA_Quadrant quadrant)
Definition FLA_View.c:142
dim_t FLA_Determine_blocksize(FLA_Obj A_unproc, FLA_Quadrant to_dir, fla_blocksize_t *cntl_blocksizes)
Definition FLA_Blocksize.c:234
unsigned long dim_t
Definition FLA_type_defs.h:71
int i
Definition bl1_axmyv2.c:145
Definition FLA_type_defs.h:159

References FLA_Cont_with_3x1_to_2x1(), FLA_Cont_with_3x3_to_2x2(), FLA_Determine_blocksize(), FLA_Gemm_internal(), FLA_Her2k_internal(), FLA_Obj_length(), FLA_ONE, FLA_Part_2x1(), FLA_Part_2x2(), FLA_Repart_2x1_to_3x1(), FLA_Repart_2x2_to_3x3(), FLA_Scalr_internal(), and i.

Referenced by FLA_Her2k_ln().