libflame revision_anchor
Functions
FLA_Trmm_rln.h File Reference

(r)

Go to the source code of this file.

Functions

FLA_Error FLA_Trmm_rln_blk_var1 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, fla_trmm_t *cntl)
 
FLA_Error FLA_Trmm_rln_blk_var2 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, fla_trmm_t *cntl)
 
FLA_Error FLA_Trmm_rln_blk_var3 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, fla_trmm_t *cntl)
 
FLA_Error FLA_Trmm_rln_blk_var4 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, fla_trmm_t *cntl)
 
FLA_Error FLA_Trmm_rln_unb_var1 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B)
 
FLA_Error FLA_Trmm_rln_unb_var2 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B)
 
FLA_Error FLA_Trmm_rln_unb_var3 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B)
 
FLA_Error FLA_Trmm_rln_unb_var4 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B)
 

Function Documentation

◆ FLA_Trmm_rln_blk_var1()

FLA_Error FLA_Trmm_rln_blk_var1 ( FLA_Diag  diagA,
FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B,
fla_trmm_t cntl 
)
16{
17 FLA_Obj ATL, ATR, A00, A01, A02,
18 ABL, ABR, A10, A11, A12,
19 A20, A21, A22;
20
21 FLA_Obj BL, BR, B0, B1, B2;
22
23 dim_t b;
24
25 FLA_Part_2x2( A, &ATL, &ATR,
26 &ABL, &ABR, 0, 0, FLA_TL );
27
28 FLA_Part_1x2( B, &BL, &BR, 0, FLA_LEFT );
29
30 while ( FLA_Obj_length( ATL ) < FLA_Obj_length( A ) ){
31
33
34 FLA_Repart_2x2_to_3x3( ATL, /**/ ATR, &A00, /**/ &A01, &A02,
35 /* ************* */ /* ******************** */
36 &A10, /**/ &A11, &A12,
37 ABL, /**/ ABR, &A20, /**/ &A21, &A22,
38 b, b, FLA_BR );
39
40 FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, /**/ &B1, &B2,
41 b, FLA_RIGHT );
42
43 /*------------------------------------------------------------*/
44
45 /* B1 = B1 * tril( A11 ); */
47 alpha, A11, B1,
48 FLA_Cntl_sub_trmm( cntl ) );
49
50 /* B1 = B1 + B2 * A21; */
52 alpha, B2, A21, FLA_ONE, B1,
53 FLA_Cntl_sub_gemm( cntl ) );
54
55 /*------------------------------------------------------------*/
56
57 FLA_Cont_with_3x3_to_2x2( &ATL, /**/ &ATR, A00, A01, /**/ A02,
58 A10, A11, /**/ A12,
59 /* ************** */ /* ****************** */
60 &ABL, /**/ &ABR, A20, A21, /**/ A22,
61 FLA_TL );
62
63 FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, B1, /**/ B2,
64 FLA_LEFT );
65
66 }
67
68 return FLA_SUCCESS;
69}
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_Trmm_internal(FLA_Side side, FLA_Uplo uplo, FLA_Trans transa, FLA_Diag diag, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, fla_trmm_t *cntl)
Definition FLA_Trmm_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_1x3_to_1x2(FLA_Obj *AL, FLA_Obj *AR, FLA_Obj A0, FLA_Obj A1, FLA_Obj A2, FLA_Side side)
Definition FLA_View.c:475
FLA_Error FLA_Part_1x2(FLA_Obj A, FLA_Obj *A1, FLA_Obj *A2, dim_t nb, FLA_Side side)
Definition FLA_View.c:110
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
FLA_Error FLA_Repart_1x2_to_1x3(FLA_Obj AL, FLA_Obj AR, FLA_Obj *A0, FLA_Obj *A1, FLA_Obj *A2, dim_t nb, FLA_Side side)
Definition FLA_View.c:267
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_1x3_to_1x2(), FLA_Cont_with_3x3_to_2x2(), FLA_Determine_blocksize(), FLA_Gemm_internal(), FLA_Obj_length(), FLA_ONE, FLA_Part_1x2(), FLA_Part_2x2(), FLA_Repart_1x2_to_1x3(), FLA_Repart_2x2_to_3x3(), FLA_Trmm_internal(), and i.

Referenced by FLA_Trmm_rln().

◆ FLA_Trmm_rln_blk_var2()

FLA_Error FLA_Trmm_rln_blk_var2 ( FLA_Diag  diagA,
FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B,
fla_trmm_t cntl 
)
14{
15 FLA_Obj ATL, ATR, A00, A01, A02,
16 ABL, ABR, A10, A11, A12,
17 A20, A21, A22;
18
19 FLA_Obj BL, BR, B0, B1, B2;
20
21 dim_t b;
22
24 FLA_Cntl_sub_scal( cntl ) );
25
26 FLA_Part_2x2( A, &ATL, &ATR,
27 &ABL, &ABR, 0, 0, FLA_TL );
28
29 FLA_Part_1x2( B, &BL, &BR, 0, FLA_LEFT );
30
31 while ( FLA_Obj_length( ATL ) < FLA_Obj_length( A ) ){
32
34
35 FLA_Repart_2x2_to_3x3( ATL, /**/ ATR, &A00, /**/ &A01, &A02,
36 /* ************* */ /* ******************** */
37 &A10, /**/ &A11, &A12,
38 ABL, /**/ ABR, &A20, /**/ &A21, &A22,
39 b, b, FLA_BR );
40
41 FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, /**/ &B1, &B2,
42 b, FLA_RIGHT );
43
44 /*------------------------------------------------------------*/
45
46 /* B0 = B0 + B1 * A10; */
49 FLA_Cntl_sub_gemm( cntl ) );
50
51 /* B1 = B1 * tril( A11 ); */
53 FLA_ONE, A11, B1,
54 FLA_Cntl_sub_trmm( cntl ) );
55
56 /*------------------------------------------------------------*/
57
58 FLA_Cont_with_3x3_to_2x2( &ATL, /**/ &ATR, A00, A01, /**/ A02,
59 A10, A11, /**/ A12,
60 /* ************** */ /* ****************** */
61 &ABL, /**/ &ABR, A20, A21, /**/ A22,
62 FLA_TL );
63
64 FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, B1, /**/ B2,
65 FLA_LEFT );
66
67 }
68
69 return FLA_SUCCESS;
70}
FLA_Error FLA_Scal_internal(FLA_Obj alpha, FLA_Obj A, fla_scal_t *cntl)
Definition FLA_Scal_internal.c:16

References FLA_Cont_with_1x3_to_1x2(), FLA_Cont_with_3x3_to_2x2(), FLA_Determine_blocksize(), FLA_Gemm_internal(), FLA_Obj_length(), FLA_ONE, FLA_Part_1x2(), FLA_Part_2x2(), FLA_Repart_1x2_to_1x3(), FLA_Repart_2x2_to_3x3(), FLA_Scal_internal(), FLA_Trmm_internal(), and i.

Referenced by FLA_Trmm_rln().

◆ FLA_Trmm_rln_blk_var3()

FLA_Error FLA_Trmm_rln_blk_var3 ( FLA_Diag  diagA,
FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B,
fla_trmm_t cntl 
)
14{
15 FLA_Obj BT, B0,
16 BB, B1,
17 B2;
18
19 dim_t b;
20
21 FLA_Part_2x1( B, &BT,
22 &BB, 0, FLA_TOP );
23
24 while ( FLA_Obj_length( BT ) < FLA_Obj_length( B ) ){
25
27
29 /* ** */ /* ** */
30 &B1,
31 BB, &B2, b, FLA_BOTTOM );
32
33 /*------------------------------------------------------------*/
34
35 /* B1 = B1 * tril( A ); */
37 alpha, A, B1,
38 FLA_Cntl_sub_trmm( cntl ) );
39
40 /*------------------------------------------------------------*/
41
43 B1,
44 /* ** */ /* ** */
45 &BB, B2, FLA_TOP );
46
47 }
48
49 return FLA_SUCCESS;
50}
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

References FLA_Cont_with_3x1_to_2x1(), FLA_Determine_blocksize(), FLA_Obj_length(), FLA_Part_2x1(), FLA_Repart_2x1_to_3x1(), FLA_Trmm_internal(), and i.

Referenced by FLA_Trmm_rln().

◆ FLA_Trmm_rln_blk_var4()

FLA_Error FLA_Trmm_rln_blk_var4 ( FLA_Diag  diagA,
FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B,
fla_trmm_t cntl 
)
16{
17 FLA_Obj BT, B0,
18 BB, B1,
19 B2;
20
21 dim_t b;
22
23 FLA_Part_2x1( B, &BT,
24 &BB, 0, FLA_BOTTOM );
25
26 while ( FLA_Obj_length( BB ) < FLA_Obj_length( B ) ){
27
29
31 &B1,
32 /* ** */ /* ** */
33 BB, &B2, b, FLA_TOP );
34
35 /*------------------------------------------------------------*/
36
37 /* B1 = B1 * tril( A ); */
39 alpha, A, B1,
40 FLA_Cntl_sub_trmm( cntl ) );
41
42 /*------------------------------------------------------------*/
43
45 /* ** */ /* ** */
46 B1,
47 &BB, B2, FLA_BOTTOM );
48
49 }
50
51 return FLA_SUCCESS;
52}

References FLA_Cont_with_3x1_to_2x1(), FLA_Determine_blocksize(), FLA_Obj_length(), FLA_Part_2x1(), FLA_Repart_2x1_to_3x1(), FLA_Trmm_internal(), and i.

Referenced by FLA_Trmm_rln().

◆ FLA_Trmm_rln_unb_var1()

FLA_Error FLA_Trmm_rln_unb_var1 ( FLA_Diag  diagA,
FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B 
)
16{
17 FLA_Obj ATL, ATR, A00, a01, A02,
19 A20, a21, A22;
20
21 FLA_Obj BL, BR, B0, b1, B2;
22
24
25 FLA_Part_2x2( A, &ATL, &ATR,
26 &ABL, &ABR, 0, 0, FLA_TL );
27
28 FLA_Part_1x2( B, &BL, &BR, 0, FLA_LEFT );
29
30 while ( FLA_Obj_length( ATL ) < FLA_Obj_length( A ) ){
31
32 FLA_Repart_2x2_to_3x3( ATL, /**/ ATR, &A00, /**/ &a01, &A02,
33 /* ************* */ /* ************************** */
34 &a10t, /**/ &alpha11, &a12t,
35 ABL, /**/ ABR, &A20, /**/ &a21, &A22,
36 1, 1, FLA_BR );
37
38 FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, /**/ &b1, &B2,
39 1, FLA_RIGHT );
40
41 /*------------------------------------------------------------*/
42
43 /* b1 = b1 * alpha11; */
44 if ( diagA != FLA_UNIT_DIAG )
46
47 /* b1 = b1 + B2 * a21; */
49
50 /*------------------------------------------------------------*/
51
52 FLA_Cont_with_3x3_to_2x2( &ATL, /**/ &ATR, A00, a01, /**/ A02,
53 a10t, alpha11, /**/ a12t,
54 /* ************** */ /* ************************ */
55 &ABL, /**/ &ABR, A20, a21, /**/ A22,
56 FLA_TL );
57
58 FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, b1, /**/ B2,
59 FLA_LEFT );
60
61 }
62
63 return FLA_SUCCESS;
64}
FLA_Error FLA_Scal_external(FLA_Obj alpha, FLA_Obj A)
Definition FLA_Scal_external.c:13
FLA_Error FLA_Gemv_external(FLA_Trans transa, FLA_Obj alpha, FLA_Obj A, FLA_Obj x, FLA_Obj beta, FLA_Obj y)
Definition FLA_Gemv_external.c:13

References FLA_Cont_with_1x3_to_1x2(), FLA_Cont_with_3x3_to_2x2(), FLA_Gemv_external(), FLA_Obj_length(), FLA_ONE, FLA_Part_1x2(), FLA_Part_2x2(), FLA_Repart_1x2_to_1x3(), FLA_Repart_2x2_to_3x3(), FLA_Scal_external(), and i.

Referenced by FLA_Trmm_rln().

◆ FLA_Trmm_rln_unb_var2()

FLA_Error FLA_Trmm_rln_unb_var2 ( FLA_Diag  diagA,
FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B 
)
16{
17 FLA_Obj ATL, ATR, A00, a01, A02,
19 A20, a21, A22;
20
21 FLA_Obj BL, BR, B0, b1, B2;
22
24
25 FLA_Part_2x2( A, &ATL, &ATR,
26 &ABL, &ABR, 0, 0, FLA_TL );
27
28 FLA_Part_1x2( B, &BL, &BR, 0, FLA_LEFT );
29
30 while ( FLA_Obj_length( ATL ) < FLA_Obj_length( A ) ){
31
32 FLA_Repart_2x2_to_3x3( ATL, /**/ ATR, &A00, /**/ &a01, &A02,
33 /* ************* */ /* ************************** */
34 &a10t, /**/ &alpha11, &a12t,
35 ABL, /**/ ABR, &A20, /**/ &a21, &A22,
36 1, 1, FLA_BR );
37
38 FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, /**/ &b1, &B2,
39 1, FLA_RIGHT );
40
41 /*------------------------------------------------------------*/
42
43 /* B0 = B0 + b1 * a10t; */
45
46 /* b1 = b1 * alpha11; */
47 if ( diagA != FLA_UNIT_DIAG )
49
50 /*------------------------------------------------------------*/
51
52 FLA_Cont_with_3x3_to_2x2( &ATL, /**/ &ATR, A00, a01, /**/ A02,
53 a10t, alpha11, /**/ a12t,
54 /* ************** */ /* ************************ */
55 &ABL, /**/ &ABR, A20, a21, /**/ A22,
56 FLA_TL );
57
58 FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, b1, /**/ B2,
59 FLA_LEFT );
60
61 }
62
63 return FLA_SUCCESS;
64}
FLA_Error FLA_Ger_external(FLA_Obj alpha, FLA_Obj x, FLA_Obj y, FLA_Obj A)
Definition FLA_Ger_external.c:13

References FLA_Cont_with_1x3_to_1x2(), FLA_Cont_with_3x3_to_2x2(), FLA_Ger_external(), FLA_Obj_length(), FLA_ONE, FLA_Part_1x2(), FLA_Part_2x2(), FLA_Repart_1x2_to_1x3(), FLA_Repart_2x2_to_3x3(), FLA_Scal_external(), and i.

Referenced by FLA_Trmm_rln().

◆ FLA_Trmm_rln_unb_var3()

FLA_Error FLA_Trmm_rln_unb_var3 ( FLA_Diag  diagA,
FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B 
)
16{
17 FLA_Obj BT, B0,
18 BB, b1t,
19 B2;
20
22
23 FLA_Part_2x1( B, &BT,
24 &BB, 0, FLA_TOP );
25
26 while ( FLA_Obj_length( BT ) < FLA_Obj_length( B ) ){
27
29 /* ** */ /* *** */
30 &b1t,
31 BB, &B2, 1, FLA_BOTTOM );
32
33 /*------------------------------------------------------------*/
34
35 /* b1t = b1t * tril( A ); */
37
38 /*------------------------------------------------------------*/
39
41 b1t,
42 /* ** */ /* *** */
43 &BB, B2, FLA_TOP );
44
45 }
46
47 return FLA_SUCCESS;
48}
FLA_Error FLA_Trmv_external(FLA_Uplo uplo, FLA_Trans transa, FLA_Diag diag, FLA_Obj A, FLA_Obj x)
Definition FLA_Trmv_external.c:13

References FLA_Cont_with_3x1_to_2x1(), FLA_Obj_length(), FLA_Part_2x1(), FLA_Repart_2x1_to_3x1(), FLA_Scal_external(), FLA_Trmv_external(), and i.

Referenced by FLA_Trmm_rln().

◆ FLA_Trmm_rln_unb_var4()

FLA_Error FLA_Trmm_rln_unb_var4 ( FLA_Diag  diagA,
FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B 
)
16{
17 FLA_Obj BT, B0,
18 BB, b1t,
19 B2;
20
22
23 FLA_Part_2x1( B, &BT,
24 &BB, 0, FLA_BOTTOM );
25
26 while ( FLA_Obj_length( BB ) < FLA_Obj_length( B ) ){
27
29 &b1t,
30 /* ** */ /* *** */
31 BB, &B2, 1, FLA_TOP );
32
33 /*------------------------------------------------------------*/
34
35 /* b1t = b1t * tril( A ); */
37
38 /*------------------------------------------------------------*/
39
41 /* ** */ /* *** */
42 b1t,
43 &BB, B2, FLA_BOTTOM );
44
45 }
46
47 return FLA_SUCCESS;
48}

References FLA_Cont_with_3x1_to_2x1(), FLA_Obj_length(), FLA_Part_2x1(), FLA_Repart_2x1_to_3x1(), FLA_Scal_external(), FLA_Trmv_external(), and i.

Referenced by FLA_Trmm_rln().