libflame revision_anchor
Functions
FLA_Trmm_luh.h File Reference

(r)

Go to the source code of this file.

Functions

FLA_Error FLA_Trmm_luh_blk_var1 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, fla_trmm_t *cntl)
 
FLA_Error FLA_Trmm_luh_blk_var2 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, fla_trmm_t *cntl)
 
FLA_Error FLA_Trmm_luh_blk_var3 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, fla_trmm_t *cntl)
 
FLA_Error FLA_Trmm_luh_blk_var4 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B, fla_trmm_t *cntl)
 
FLA_Error FLA_Trmm_luh_unb_var1 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B)
 
FLA_Error FLA_Trmm_luh_unb_var2 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B)
 
FLA_Error FLA_Trmm_luh_unb_var3 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B)
 
FLA_Error FLA_Trmm_luh_unb_var4 (FLA_Diag diagA, FLA_Obj alpha, FLA_Obj A, FLA_Obj B)
 

Function Documentation

◆ FLA_Trmm_luh_blk_var1()

FLA_Error FLA_Trmm_luh_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 BT, B0,
22 BB, B1,
23 B2;
24
25 dim_t b;
26
27 FLA_Part_2x2( A, &ATL, &ATR,
28 &ABL, &ABR, 0, 0, FLA_BR );
29
30 FLA_Part_2x1( B, &BT,
31 &BB, 0, FLA_BOTTOM );
32
33 while ( FLA_Obj_length( ABR ) < FLA_Obj_length( A ) ){
34
36
37 FLA_Repart_2x2_to_3x3( ATL, /**/ ATR, &A00, &A01, /**/ &A02,
38 &A10, &A11, /**/ &A12,
39 /* ************* */ /* ******************** */
40 ABL, /**/ ABR, &A20, &A21, /**/ &A22,
41 b, b, FLA_TL );
42
44 &B1,
45 /* ** */ /* ** */
46 BB, &B2, b, FLA_TOP );
47
48 /*------------------------------------------------------------*/
49
50 /* B1 = triu( A11' ) * B1; */
52 alpha, A11, B1,
53 FLA_Cntl_sub_trmm( cntl ) );
54
55 /* B1 = B1 + A01' * B0; */
57 alpha, A01, B0, FLA_ONE, B1,
58 FLA_Cntl_sub_gemm( cntl ) );
59
60 /*------------------------------------------------------------*/
61
62 FLA_Cont_with_3x3_to_2x2( &ATL, /**/ &ATR, A00, /**/ A01, A02,
63 /* ************** */ /* ****************** */
64 A10, /**/ A11, A12,
65 &ABL, /**/ &ABR, A20, /**/ A21, A22,
66 FLA_BR );
67
69 /* ** */ /* ** */
70 B1,
71 &BB, B2, FLA_BOTTOM );
72
73 }
74
75 return FLA_SUCCESS;
76}
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_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_Obj_length(), FLA_ONE, FLA_Part_2x1(), FLA_Part_2x2(), FLA_Repart_2x1_to_3x1(), FLA_Repart_2x2_to_3x3(), FLA_Trmm_internal(), and i.

Referenced by FLA_Trmm_luh().

◆ FLA_Trmm_luh_blk_var2()

FLA_Error FLA_Trmm_luh_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 BT, B0,
20 BB, B1,
21 B2;
22
23 dim_t b;
24
26 FLA_Cntl_sub_scal( cntl ) );
27
28 FLA_Part_2x2( A, &ATL, &ATR,
29 &ABL, &ABR, 0, 0, FLA_BR );
30
31 FLA_Part_2x1( B, &BT,
32 &BB, 0, FLA_BOTTOM );
33
34 while ( FLA_Obj_length( ABR ) < FLA_Obj_length( A ) ){
35
37
38 FLA_Repart_2x2_to_3x3( ATL, /**/ ATR, &A00, &A01, /**/ &A02,
39 &A10, &A11, /**/ &A12,
40 /* ************* */ /* ******************** */
41 ABL, /**/ ABR, &A20, &A21, /**/ &A22,
42 b, b, FLA_TL );
43
45 &B1,
46 /* ** */ /* ** */
47 BB, &B2, b, FLA_TOP );
48
49 /*------------------------------------------------------------*/
50
51 /* B2 = B2 + A12' * B1; */
54 FLA_Cntl_sub_gemm( cntl ) );
55
56 /* B1 = triu( A11' ) * B1; */
58 FLA_ONE, A11, B1,
59 FLA_Cntl_sub_trmm( cntl ) );
60
61 /*------------------------------------------------------------*/
62
63 FLA_Cont_with_3x3_to_2x2( &ATL, /**/ &ATR, A00, /**/ A01, A02,
64 /* ************** */ /* ****************** */
65 A10, /**/ A11, A12,
66 &ABL, /**/ &ABR, A20, /**/ A21, A22,
67 FLA_BR );
68
70 /* ** */ /* ** */
71 B1,
72 &BB, B2, FLA_BOTTOM );
73
74 }
75
76 return FLA_SUCCESS;
77}
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_3x1_to_2x1(), FLA_Cont_with_3x3_to_2x2(), FLA_Determine_blocksize(), FLA_Gemm_internal(), FLA_Obj_length(), FLA_ONE, FLA_Part_2x1(), FLA_Part_2x2(), FLA_Repart_2x1_to_3x1(), FLA_Repart_2x2_to_3x3(), FLA_Scal_internal(), FLA_Trmm_internal(), and i.

Referenced by FLA_Trmm_luh().

◆ FLA_Trmm_luh_blk_var3()

FLA_Error FLA_Trmm_luh_blk_var3 ( FLA_Diag  diagA,
FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B,
fla_trmm_t cntl 
)
14{
15 FLA_Obj BL, BR, B0, B1, B2;
16
17 dim_t b;
18
19 FLA_Part_1x2( B, &BL, &BR, 0, FLA_LEFT );
20
21 while ( FLA_Obj_width( BL ) < FLA_Obj_width( B ) ){
22
24
25 FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, /**/ &B1, &B2,
26 b, FLA_RIGHT );
27
28 /*------------------------------------------------------------*/
29
30 /* B1 = triu( A )' * B1; */
32 alpha, A, B1,
33 FLA_Cntl_sub_trmm( cntl ) );
34
35 /*------------------------------------------------------------*/
36
37 FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, B1, /**/ B2,
38 FLA_LEFT );
39
40 }
41
42 return FLA_SUCCESS;
43}
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
dim_t FLA_Obj_width(FLA_Obj obj)
Definition FLA_Query.c:123
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
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

References FLA_Cont_with_1x3_to_1x2(), FLA_Determine_blocksize(), FLA_Obj_width(), FLA_Part_1x2(), FLA_Repart_1x2_to_1x3(), FLA_Trmm_internal(), and i.

Referenced by FLA_Trmm_luh().

◆ FLA_Trmm_luh_blk_var4()

FLA_Error FLA_Trmm_luh_blk_var4 ( FLA_Diag  diagA,
FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B,
fla_trmm_t cntl 
)
16{
17 FLA_Obj BL, BR, B0, B1, B2;
18
19 dim_t b;
20
21 FLA_Part_1x2( B, &BL, &BR, 0, FLA_RIGHT );
22
23 while ( FLA_Obj_width( BR ) < FLA_Obj_width( B ) ){
24
26
27 FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, &B1, /**/ &B2,
28 b, FLA_LEFT );
29
30 /*------------------------------------------------------------*/
31
32 /* B1 = triu( A )' * B1; */
34 alpha, A, B1,
35 FLA_Cntl_sub_trmm( cntl ) );
36
37 /*------------------------------------------------------------*/
38
39 FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, /**/ B1, B2,
40 FLA_RIGHT );
41
42 }
43
44 return FLA_SUCCESS;
45}

References FLA_Cont_with_1x3_to_1x2(), FLA_Determine_blocksize(), FLA_Obj_width(), FLA_Part_1x2(), FLA_Repart_1x2_to_1x3(), FLA_Trmm_internal(), and i.

Referenced by FLA_Trmm_luh().

◆ FLA_Trmm_luh_unb_var1()

FLA_Error FLA_Trmm_luh_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 BT, B0,
22 BB, b1t,
23 B2;
24
26
27 FLA_Part_2x2( A, &ATL, &ATR,
28 &ABL, &ABR, 0, 0, FLA_BR );
29
30 FLA_Part_2x1( B, &BT,
31 &BB, 0, FLA_BOTTOM );
32
33 while ( FLA_Obj_length( ABR ) < FLA_Obj_length( A ) ){
34
35 FLA_Repart_2x2_to_3x3( ATL, /**/ ATR, &A00, &a01, /**/ &A02,
36 &a10t, &alpha11, /**/ &a12t,
37 /* ************* */ /* ************************** */
38 ABL, /**/ ABR, &A20, &a21, /**/ &A22,
39 1, 1, FLA_TL );
40
42 &b1t,
43 /* ** */ /* *** */
44 BB, &B2, 1, FLA_TOP );
45
46 /*------------------------------------------------------------*/
47
48 /* b1t = alpha11 * b1t */
49 if ( diagA != FLA_UNIT_DIAG )
51
52 /* b1t = b1t + a01' * B0; */
54
55 /*------------------------------------------------------------*/
56
57 FLA_Cont_with_3x3_to_2x2( &ATL, /**/ &ATR, A00, /**/ a01, A02,
58 /* ************** */ /* ************************ */
59 a10t, /**/ alpha11, a12t,
60 &ABL, /**/ &ABR, A20, /**/ a21, A22,
61 FLA_BR );
62
64 /* ** */ /* *** */
65 b1t,
66 &BB, B2, FLA_BOTTOM );
67
68 }
69
70 return FLA_SUCCESS;
71}
FLA_Error FLA_Scalc_external(FLA_Conj conjalpha, FLA_Obj alpha, FLA_Obj A)
Definition FLA_Scalc_external.c:13
FLA_Error FLA_Scal_external(FLA_Obj alpha, FLA_Obj A)
Definition FLA_Scal_external.c:13
FLA_Error FLA_Gemvc_external(FLA_Trans transa, FLA_Conj conjx, FLA_Obj alpha, FLA_Obj A, FLA_Obj x, FLA_Obj beta, FLA_Obj y)
Definition FLA_Gemvc_external.c:13

References FLA_Cont_with_3x1_to_2x1(), FLA_Cont_with_3x3_to_2x2(), FLA_Gemvc_external(), FLA_Obj_length(), FLA_ONE, FLA_Part_2x1(), FLA_Part_2x2(), FLA_Repart_2x1_to_3x1(), FLA_Repart_2x2_to_3x3(), FLA_Scal_external(), FLA_Scalc_external(), and i.

Referenced by FLA_Trmm_luh().

◆ FLA_Trmm_luh_unb_var2()

FLA_Error FLA_Trmm_luh_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 BT, B0,
22 BB, b1t,
23 B2;
24
26
27 FLA_Part_2x2( A, &ATL, &ATR,
28 &ABL, &ABR, 0, 0, FLA_BR );
29
30 FLA_Part_2x1( B, &BT,
31 &BB, 0, FLA_BOTTOM );
32
33 while ( FLA_Obj_length( ABR ) < FLA_Obj_length( A ) ){
34
35 FLA_Repart_2x2_to_3x3( ATL, /**/ ATR, &A00, &a01, /**/ &A02,
36 &a10t, &alpha11, /**/ &a12t,
37 /* ************* */ /* ************************** */
38 ABL, /**/ ABR, &A20, &a21, /**/ &A22,
39 1, 1, FLA_TL );
40
42 &b1t,
43 /* ** */ /* *** */
44 BB, &B2, 1, FLA_TOP );
45
46 /*------------------------------------------------------------*/
47
48 /* B2 = B2 + a12t' * b1t; */
50
51 /* b1t = alpha11 * b1t; */
52 if ( diagA != FLA_UNIT_DIAG )
54
55 /*------------------------------------------------------------*/
56
57 FLA_Cont_with_3x3_to_2x2( &ATL, /**/ &ATR, A00, /**/ a01, A02,
58 /* ************** */ /* ************************ */
59 a10t, /**/ alpha11, a12t,
60 &ABL, /**/ &ABR, A20, /**/ a21, A22,
61 FLA_BR );
62
64 /* ** */ /* *** */
65 b1t,
66 &BB, B2, FLA_BOTTOM );
67
68 }
69
70 return FLA_SUCCESS;
71}
FLA_Error FLA_Gerc_external(FLA_Conj conjx, FLA_Conj conjy, FLA_Obj alpha, FLA_Obj x, FLA_Obj y, FLA_Obj A)
Definition FLA_Gerc_external.c:13

References FLA_Cont_with_3x1_to_2x1(), FLA_Cont_with_3x3_to_2x2(), FLA_Gerc_external(), FLA_Obj_length(), FLA_ONE, FLA_Part_2x1(), FLA_Part_2x2(), FLA_Repart_2x1_to_3x1(), FLA_Repart_2x2_to_3x3(), FLA_Scal_external(), FLA_Scalc_external(), and i.

Referenced by FLA_Trmm_luh().

◆ FLA_Trmm_luh_unb_var3()

FLA_Error FLA_Trmm_luh_unb_var3 ( FLA_Diag  diagA,
FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B 
)
16{
17 FLA_Obj BL, BR, B0, b1, B2;
18
20
21 FLA_Part_1x2( B, &BL, &BR, 0, FLA_LEFT );
22
23 while ( FLA_Obj_width( BL ) < FLA_Obj_width( B ) ){
24
25 FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, /**/ &b1, &B2,
26 1, FLA_RIGHT );
27
28 /*------------------------------------------------------------*/
29
30 /* b1 = triu( A )' * b1; */
32
33 /*------------------------------------------------------------*/
34
35 FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, b1, /**/ B2,
36 FLA_LEFT );
37
38 }
39
40 return FLA_SUCCESS;
41}
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_1x3_to_1x2(), FLA_Obj_width(), FLA_Part_1x2(), FLA_Repart_1x2_to_1x3(), FLA_Scal_external(), FLA_Trmv_external(), and i.

Referenced by FLA_Trmm_luh().

◆ FLA_Trmm_luh_unb_var4()

FLA_Error FLA_Trmm_luh_unb_var4 ( FLA_Diag  diagA,
FLA_Obj  alpha,
FLA_Obj  A,
FLA_Obj  B 
)
16{
17 FLA_Obj BL, BR, B0, b1, B2;
18
20
21 FLA_Part_1x2( B, &BL, &BR, 0, FLA_RIGHT );
22
23 while ( FLA_Obj_width( BR ) < FLA_Obj_width( B ) ){
24
25 FLA_Repart_1x2_to_1x3( BL, /**/ BR, &B0, &b1, /**/ &B2,
26 1, FLA_LEFT );
27
28 /*------------------------------------------------------------*/
29
30 /* b1 = triu( A )' * b1; */
32
33 /*------------------------------------------------------------*/
34
35 FLA_Cont_with_1x3_to_1x2( &BL, /**/ &BR, B0, /**/ b1, B2,
36 FLA_RIGHT );
37
38 }
39
40 return FLA_SUCCESS;
41}

References FLA_Cont_with_1x3_to_1x2(), FLA_Obj_width(), FLA_Part_1x2(), FLA_Repart_1x2_to_1x3(), FLA_Scal_external(), FLA_Trmv_external(), and i.

Referenced by FLA_Trmm_luh().