mirror of
https://github.com/amd/blis.git
synced 2026-05-11 01:30:00 +00:00
b444489f10
Details:
- Added infrastructure to support a new scalar representation, whereby
every object contains an internal scalar that defaults to 1.0. This
facilitates passing scalars around without having to house them in
separate objects. These "attached" scalars are stored in the internal
atom_t field of the obj_t struct, and are always stored to be the same
datatype as the object to which they are attached. Level-3 variants no
longer take scalar arguments, however, level-3 internal back-ends stll
do; this is so that the calling function can perform subproblems such
as C := C - alpha * A * B on-the-fly without needing to change either
of the scalars attached to A or B.
- Removed scalar argument from packm_int().
- Observe and apply attached scalars in scalm_int(), and removed scalar
from interface of scalm_unb_var1().
- Renamed the following functions (and corresponding invocations):
bli_obj_init_scalar_copy_of()
-> bli_obj_scalar_init_detached_copy_of()
bli_obj_init_scalar() -> bli_obj_scalar_init_detached()
bli_obj_create_scalar_with_attached_buffer()
-> bli_obj_create_1x1_with_attached_buffer()
bli_obj_scalar_equals() -> bli_obj_equals()
- Defined new functions:
bli_obj_scalar_detach()
bli_obj_scalar_attach()
bli_obj_scalar_apply_scalar()
bli_obj_scalar_reset()
bli_obj_scalar_has_nonzero_imag()
bli_obj_scalar_equals()
- Placed all bli_obj_scalar_* functions in a new file, bli_obj_scalar.c.
- Renamed the following macros:
bli_obj_scalar_buffer() -> bli_obj_buffer_for_1x1()
bli_obj_is_scalar() -> bli_obj_is_1x1()
- Defined new macros to set and copy internal scalars between objects:
bli_obj_set_internal_scalar()
bli_obj_copy_internal_scalar()
- In level-3 internal back-ends, added conditional blocks where alpha and
beta are checked for non-unit-ness. Those values for alpha and beta are
applied to the scalars attached to aliases of A/B/C, as appropriate,
before being passed into the variant specified by the control tree.
- In level-3 blocked variants, pass BLIS_ONE into subproblems instead of
alpha and/or beta.
- In level-3 macro-kernels, changed how scalars are obtained. Now, scalars
attached to A and B are multiplied together to obtain alpha, while beta
is obtained directly from C.
- In level-3 front-ends, removed old function calls meant to provide
future support for mixed domain/precision. These can be added back later
once that functionality is given proper treatment. Also, removed the
creating of copy-casts of alpha and beta since typecasting of scalars
is now implicitly handled in the internal back-ends when alpha and
beta are applied to the attached scalars.
222 lines
7.2 KiB
C
222 lines
7.2 KiB
C
/*
|
|
|
|
BLIS
|
|
An object-based framework for developing high-performance BLAS-like
|
|
libraries.
|
|
|
|
Copyright (C) 2013, The University of Texas
|
|
|
|
Redistribution and use in source and binary forms, with or without
|
|
modification, are permitted provided that the following conditions are
|
|
met:
|
|
- Redistributions of source code must retain the above copyright
|
|
notice, this list of conditions and the following disclaimer.
|
|
- Redistributions in binary form must reproduce the above copyright
|
|
notice, this list of conditions and the following disclaimer in the
|
|
documentation and/or other materials provided with the distribution.
|
|
- Neither the name of The University of Texas nor the names of its
|
|
contributors may be used to endorse or promote products derived
|
|
from this software without specific prior written permission.
|
|
|
|
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
|
|
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
|
|
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
|
|
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
|
|
HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
|
|
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
|
|
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
|
|
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
|
|
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
|
|
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
|
|
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
|
|
*/
|
|
|
|
#include "blis.h"
|
|
|
|
#define FUNCPTR_T trmv_fp
|
|
|
|
typedef void (*FUNCPTR_T)(
|
|
uplo_t uplo,
|
|
trans_t trans,
|
|
diag_t diag,
|
|
dim_t m,
|
|
void* alpha,
|
|
void* a, inc_t rs_a, inc_t cs_a,
|
|
void* x, inc_t incx
|
|
);
|
|
|
|
// If some mixed datatype functions will not be compiled, we initialize
|
|
// the corresponding elements of the function array to NULL.
|
|
#ifdef BLIS_ENABLE_MIXED_PRECISION_SUPPORT
|
|
static FUNCPTR_T GENARRAY2_ALL(ftypes,trmv_unb_var2);
|
|
#else
|
|
#ifdef BLIS_ENABLE_MIXED_DOMAIN_SUPPORT
|
|
static FUNCPTR_T GENARRAY2_EXT(ftypes,trmv_unb_var2);
|
|
#else
|
|
static FUNCPTR_T GENARRAY2_MIN(ftypes,trmv_unb_var2);
|
|
#endif
|
|
#endif
|
|
|
|
|
|
void bli_trmv_unb_var2( obj_t* alpha,
|
|
obj_t* a,
|
|
obj_t* x,
|
|
trmv_t* cntl )
|
|
{
|
|
num_t dt_a = bli_obj_datatype( *a );
|
|
num_t dt_x = bli_obj_datatype( *x );
|
|
|
|
uplo_t uplo = bli_obj_uplo( *a );
|
|
trans_t trans = bli_obj_conjtrans_status( *a );
|
|
diag_t diag = bli_obj_diag( *a );
|
|
|
|
dim_t m = bli_obj_length( *a );
|
|
|
|
void* buf_a = bli_obj_buffer_at_off( *a );
|
|
inc_t rs_a = bli_obj_row_stride( *a );
|
|
inc_t cs_a = bli_obj_col_stride( *a );
|
|
|
|
void* buf_x = bli_obj_buffer_at_off( *x );
|
|
inc_t incx = bli_obj_vector_inc( *x );
|
|
|
|
num_t dt_alpha;
|
|
void* buf_alpha;
|
|
|
|
FUNCPTR_T f;
|
|
|
|
// The datatype of alpha MUST be the type union of a and x. This is to
|
|
// prevent any unnecessary loss of information during computation.
|
|
dt_alpha = bli_datatype_union( dt_a, dt_x );
|
|
buf_alpha = bli_obj_buffer_for_1x1( dt_alpha, *alpha );
|
|
|
|
// Index into the type combination array to extract the correct
|
|
// function pointer.
|
|
f = ftypes[dt_a][dt_x];
|
|
|
|
// Invoke the function.
|
|
f( uplo,
|
|
trans,
|
|
diag,
|
|
m,
|
|
buf_alpha,
|
|
buf_a, rs_a, cs_a,
|
|
buf_x, incx );
|
|
}
|
|
|
|
|
|
#undef GENTFUNC2U
|
|
#define GENTFUNC2U( ctype_a, ctype_x, ctype_ax, cha, chx, chax, varname, kername ) \
|
|
\
|
|
void PASTEMAC2(cha,chx,varname)( \
|
|
uplo_t uplo, \
|
|
trans_t trans, \
|
|
diag_t diag, \
|
|
dim_t m, \
|
|
void* alpha, \
|
|
void* a, inc_t rs_a, inc_t cs_a, \
|
|
void* x, inc_t incx \
|
|
) \
|
|
{ \
|
|
ctype_ax* alpha_cast = alpha; \
|
|
ctype_a* a_cast = a; \
|
|
ctype_x* x_cast = x; \
|
|
ctype_a* a01; \
|
|
ctype_a* alpha11; \
|
|
ctype_a* a21; \
|
|
ctype_x* x0; \
|
|
ctype_x* chi1; \
|
|
ctype_x* x2; \
|
|
ctype_ax alpha_alpha11_conj; \
|
|
ctype_ax alpha_chi1; \
|
|
dim_t iter, i; \
|
|
dim_t n_behind; \
|
|
inc_t rs_at, cs_at; \
|
|
uplo_t uplo_trans; \
|
|
conj_t conja; \
|
|
\
|
|
if ( bli_zero_dim1( m ) ) return; \
|
|
\
|
|
if ( bli_does_notrans( trans ) ) \
|
|
{ \
|
|
rs_at = rs_a; \
|
|
cs_at = cs_a; \
|
|
uplo_trans = uplo; \
|
|
} \
|
|
else /* if ( bli_does_trans( trans ) ) */ \
|
|
{ \
|
|
rs_at = cs_a; \
|
|
cs_at = rs_a; \
|
|
uplo_trans = bli_uplo_toggled( uplo ); \
|
|
} \
|
|
\
|
|
conja = bli_extract_conj( trans ); \
|
|
\
|
|
/* We reduce all of the possible cases down to just lower/upper. */ \
|
|
if ( bli_is_upper( uplo_trans ) ) \
|
|
{ \
|
|
for ( iter = 0; iter < m; ++iter ) \
|
|
{ \
|
|
i = iter; \
|
|
n_behind = i; \
|
|
alpha11 = a_cast + (i )*rs_at + (i )*cs_at; \
|
|
a01 = a_cast + (0 )*rs_at + (i )*cs_at; \
|
|
chi1 = x_cast + (i )*incx; \
|
|
x0 = x_cast + (0 )*incx; \
|
|
\
|
|
/* x0 = x0 + alpha * chi1 * a01; */ \
|
|
PASTEMAC3(chax,chx,chax,scal2s)( *alpha_cast, *chi1, alpha_chi1 ); \
|
|
PASTEMAC3(chax,cha,chx,axpyv)( conja, \
|
|
n_behind, \
|
|
&alpha_chi1, \
|
|
a01, rs_at, \
|
|
x0, incx ); \
|
|
\
|
|
/* chi1 = alpha * alpha11 * chi1; */ \
|
|
PASTEMAC2(chax,chax,copys)( *alpha_cast, alpha_alpha11_conj ); \
|
|
if ( bli_is_nonunit_diag( diag ) ) \
|
|
PASTEMAC2(cha,chax,scalcjs)( conja, *alpha11, alpha_alpha11_conj ); \
|
|
PASTEMAC2(chax,chx,scals)( alpha_alpha11_conj, *chi1 ); \
|
|
} \
|
|
} \
|
|
else /* if ( bli_is_lower( uplo_trans ) ) */ \
|
|
{ \
|
|
for ( iter = 0; iter < m; ++iter ) \
|
|
{ \
|
|
i = m - iter - 1; \
|
|
n_behind = iter; \
|
|
alpha11 = a_cast + (i )*rs_at + (i )*cs_at; \
|
|
a21 = a_cast + (i+1)*rs_at + (i )*cs_at; \
|
|
chi1 = x_cast + (i )*incx; \
|
|
x2 = x_cast + (i+1)*incx; \
|
|
\
|
|
/* x2 = x2 + alpha * chi1 * a21; */ \
|
|
PASTEMAC3(chax,chx,chax,scal2s)( *alpha_cast, *chi1, alpha_chi1 ); \
|
|
PASTEMAC3(chax,cha,chx,kername)( conja, \
|
|
n_behind, \
|
|
&alpha_chi1, \
|
|
a21, rs_at, \
|
|
x2, incx ); \
|
|
\
|
|
/* chi1 = alpha * alpha11 * chi1; */ \
|
|
PASTEMAC2(chax,chax,copys)( *alpha_cast, alpha_alpha11_conj ); \
|
|
if ( bli_is_nonunit_diag( diag ) ) \
|
|
PASTEMAC2(cha,chax,scalcjs)( conja, *alpha11, alpha_alpha11_conj ); \
|
|
PASTEMAC2(chax,chx,scals)( alpha_alpha11_conj, *chi1 ); \
|
|
} \
|
|
} \
|
|
}
|
|
|
|
// Define the basic set of functions unconditionally, and then also some
|
|
// mixed datatype functions if requested.
|
|
INSERT_GENTFUNC2U_BASIC( trmv_unb_var2, AXPYV_KERNEL )
|
|
|
|
#ifdef BLIS_ENABLE_MIXED_DOMAIN_SUPPORT
|
|
INSERT_GENTFUNC2U_MIX_D( trmv_unb_var2, AXPYV_KERNEL )
|
|
#endif
|
|
|
|
#ifdef BLIS_ENABLE_MIXED_PRECISION_SUPPORT
|
|
INSERT_GENTFUNC2U_MIX_P( trmv_unb_var2, AXPYV_KERNEL )
|
|
#endif
|
|
|