/* BLIS An object-based framework for developing high-performance BLAS-like libraries. Copyright (C) 2014, The University of Texas at Austin Copyright (C) 2018 - 2019, Advanced Micro Devices, Inc. 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(s) of the copyright holder(s) 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" typedef void (*xpbys_mxn_l_vft) ( doff_t diagoff, dim_t m, dim_t n, void* x, inc_t rs_x, inc_t cs_x, void* b, void* y, inc_t rs_y, inc_t cs_y ); #undef GENTFUNC #define GENTFUNC(ctype,ch,op) \ \ void PASTEMAC(ch,op) \ ( \ doff_t diagoff, \ dim_t m, \ dim_t n, \ void* x, inc_t rs_x, inc_t cs_x, \ void* b, \ void* y, inc_t rs_y, inc_t cs_y \ ) \ { \ ctype* restrict x_cast = x; \ ctype* restrict b_cast = b; \ ctype* restrict y_cast = y; \ \ PASTEMAC3(ch,ch,ch,xpbys_mxn_l) \ ( \ diagoff, \ m, n, \ x_cast, rs_x, cs_x, \ b_cast, \ y_cast, rs_y, cs_y \ ); \ } INSERT_GENTFUNC_BASIC0(xpbys_mxn_l_fn); static xpbys_mxn_l_vft GENARRAY(xpbys_mxn_l, xpbys_mxn_l_fn); void bli_gemmt_l_ker_var2b ( const obj_t* a, const obj_t* b, const obj_t* c, const cntx_t* cntx, const cntl_t* cntl, thrinfo_t* thread_par ) { const num_t dt = bli_obj_exec_dt( c ); const dim_t dt_size = bli_dt_size( dt ); doff_t diagoffc = bli_obj_diag_offset( c ); const pack_t schema_a = bli_obj_pack_schema( a ); const pack_t schema_b = bli_obj_pack_schema( b ); dim_t m = bli_obj_length( c ); dim_t n = bli_obj_width( c ); dim_t k = bli_obj_width( a ); const void* buf_a = bli_obj_buffer_at_off( a ); const inc_t is_a = bli_obj_imag_stride( a ); const dim_t pd_a = bli_obj_panel_dim( a ); const inc_t ps_a = bli_obj_panel_stride( a ); const void* buf_b = bli_obj_buffer_at_off( b ); const inc_t is_b = bli_obj_imag_stride( b ); const dim_t pd_b = bli_obj_panel_dim( b ); const inc_t ps_b = bli_obj_panel_stride( b ); void* buf_c = bli_obj_buffer_at_off( c ); const inc_t rs_c = bli_obj_row_stride( c ); const inc_t cs_c = bli_obj_col_stride( c ); // Detach and multiply the scalars attached to A and B. obj_t scalar_a, scalar_b; bli_obj_scalar_detach( a, &scalar_a ); bli_obj_scalar_detach( b, &scalar_b ); bli_mulsc( &scalar_a, &scalar_b ); // Grab the addresses of the internal scalar buffers for the scalar // merged above and the scalar attached to C. const void* buf_alpha = bli_obj_internal_scalar_buffer( &scalar_b ); const void* buf_beta = bli_obj_internal_scalar_buffer( c ); // Index into the type combination array to extract the correct // function pointer. ftypes[dt_exec] ( diagoffc, schema_a, schema_b, m, n, k, ( void* )buf_alpha, ( void* )buf_a, cs_a, is_a, pd_a, ps_a, ( void* )buf_b, rs_b, is_b, pd_b, ps_b, ( void* )buf_beta, buf_c, rs_c, cs_c, ( cntx_t* )cntx, rntm, thread ); } #undef GENTFUNC #define GENTFUNC( ctype, ch, varname ) \ \ void PASTEMAC(ch,varname) \ ( \ doff_t diagoffc, \ pack_t schema_a, \ pack_t schema_b, \ dim_t m, \ dim_t n, \ dim_t k, \ void* alpha, \ void* a, inc_t cs_a, inc_t is_a, \ dim_t pd_a, inc_t ps_a, \ void* b, inc_t rs_b, inc_t is_b, \ dim_t pd_b, inc_t ps_b, \ void* beta, \ void* c, inc_t rs_c, inc_t cs_c, \ cntx_t* cntx, \ rntm_t* rntm, \ thrinfo_t* thread \ ) \ { \ const num_t dt = PASTEMAC(ch,type); \ \ /* Alias some constants to simpler names. */ \ const dim_t MR = pd_a; \ const dim_t NR = pd_b; \ /*const dim_t PACKMR = cs_a;*/ \ /*const dim_t PACKNR = rs_b;*/ \ \ /* Query the context for the micro-kernel address and cast it to its function pointer type. */ \ PASTECH(ch,gemm_ukr_ft) \ gemm_ukr = bli_cntx_get_l3_vir_ukr_dt( dt, BLIS_GEMM_UKR, cntx ); \ \ /* Temporary C buffer for edge cases. Note that the strides of this temporary buffer are set so that they match the storage of the original C matrix. For example, if C is column-stored, ct will be column-stored as well. */ \ ctype ct[ BLIS_STACK_BUF_MAX_SIZE \ / sizeof( ctype ) ] \ __attribute__((aligned(BLIS_STACK_BUF_ALIGN_SIZE))); \ const bool col_pref = bli_cntx_ukr_prefers_cols_dt( dt, BLIS_GEMM_VIR_UKR, cntx ); \ const inc_t rs_ct = ( col_pref ? 1 : NR ); \ const inc_t cs_ct = ( col_pref ? MR : 1 ); \ \ ctype* restrict zero = PASTEMAC(ch,0); \ ctype* restrict a_cast = a; \ ctype* restrict b_cast = b; \ ctype* restrict c_cast = c; \ ctype* restrict alpha_cast = alpha; \ ctype* restrict beta_cast = beta; \ \ auxinfo_t aux; \ \ /* Assumptions/assertions: rs_a == 1 cs_a == PACKMR pd_a == MR ps_a == stride to next micro-panel of A rs_b == PACKNR cs_b == 1 pd_b == NR ps_b == stride to next micro-panel of B rs_c == (no assumptions) cs_c == (no assumptions) */ \ \ /* If any dimension is zero, return immediately. */ \ if ( bli_zero_dim3( m, n, k ) ) return; \ \ /* Safeguard: If the current panel of C is entirely above the diagonal, it is not stored. So we do nothing. */ \ if ( bli_is_strictly_above_diag_n( diagoffc, m, n ) ) return; \ \ /* If there is a zero region above where the diagonal of C intersects the left edge of the panel, adjust the pointer to C and A and treat this case as if the diagonal offset were zero. NOTE: It's possible that after this pruning that the diagonal offset is still negative (though its absolute value is guaranteed to be less than MR). */ \ if ( diagoffc < 0 ) \ { \ const dim_t ip = -diagoffc / MR; \ const dim_t i = ip * MR; \ \ m = m - i; \ diagoffc = diagoffc % MR; \ c_cast = c_cast + (i )*rs_c; \ a_cast = a_cast + (ip )*ps_a; \ } \ \ /* If there is a zero region to the right of where the diagonal of C intersects the bottom of the panel, shrink it to prevent "no-op" iterations from executing. */ \ if ( diagoffc + m < n ) \ { \ n = diagoffc + m; \ } \ \ /* Clear the temporary C buffer in case it has any infs or NaNs. */ \ PASTEMAC(ch,set0s_mxn)( MR, NR, \ ct, rs_ct, cs_ct ); \ \ /* Compute number of primary and leftover components of the m and n dimensions. */ \ const dim_t n_iter = n / NR + ( n % NR ? 1 : 0 ); \ const dim_t n_left = n % NR; \ \ const dim_t m_iter = m / MR + ( m % MR ? 1 : 0 ); \ const dim_t m_left = m % MR; \ \ /* Determine some increments used to step through A, B, and C. */ \ const inc_t rstep_a = ps_a; \ \ const inc_t cstep_b = ps_b; \ \ const inc_t rstep_c = rs_c * MR; \ const inc_t cstep_c = cs_c * NR; \ \ /* Save the pack schemas of A and B to the auxinfo_t object. */ \ bli_auxinfo_set_schema_a( schema_a, &aux ); \ bli_auxinfo_set_schema_b( schema_b, &aux ); \ \ /* Save the imaginary stride of A and B to the auxinfo_t object. */ \ bli_auxinfo_set_is_a( is_a, &aux ); \ bli_auxinfo_set_is_b( is_b, &aux ); \ \ const dim_t jr_inc = 1; \ const dim_t ir_inc = 1; \ \ /* Determine the starting microtile offsets and number of microtiles to compute for each thread. Note that assignment of microtiles is done according to the tlb policy. */ \ dim_t jr_st, ir_st; \ const dim_t n_ut_for_me \ = \ bli_thread_range_tlb( thread, diagoffc, BLIS_LOWER, m, n, MR, NR, \ &jr_st, &ir_st ); \ \ /* It's possible that there are so few microtiles relative to the number of threads that one or more threads gets no work. If that happens, those threads can return early. */ \ if ( n_ut_for_me == 0 ) return; \ \ /* Start the jr/ir loops with the current thread's microtile offsets computed by bli_thread_range_tlb(). */ \ dim_t i = ir_st; \ dim_t j = jr_st; \ \ /* Initialize a counter to track the number of microtiles computed by the current thread. */ \ dim_t ut = 0; \ \ /* Loop over the n dimension (NR columns at a time). */ \ for ( ; true; ++j ) \ { \ ctype* restrict b1 = b_cast + j * cstep_b; \ ctype* restrict c1 = c_cast + j * cstep_c; \ \ /* Compute the diagonal offset for the column of microtiles at (0,j). */ \ const doff_t diagoffc_j = diagoffc - (doff_t)j*NR; \ const dim_t n_cur = ( bli_is_not_edge_f( j, n_iter, n_left ) \ ? NR : n_left ); \ \ /* Initialize our next panel of B to be the current panel of B. */ \ ctype* restrict b2 = b1; \ \ /* Interior loop over the m dimension (MR rows at a time). */ \ for ( ; i < m_iter; ++i ) \ { \ /* Compute the diagonal offset for the microtile at (i,j). */ \ const doff_t diagoffc_ij = diagoffc_j + (doff_t)i*MR; \ const dim_t m_cur = ( bli_is_not_edge_f( i, m_iter, m_left ) \ ? MR : m_left ); \ \ /* If the diagonal intersects the current MR x NR microtile, we compute it the temporary buffer and then add in the elements on or below the diagonal. Otherwise, if the microtile is strictly below the diagonal, we compute and store as we normally would. And if we're strictly above the diagonal, we simply advance to last microtile before the diagonal. */ \ if ( bli_intersects_diag_n( diagoffc_ij, m_cur, n_cur ) ) \ { \ ctype* restrict a1 = a_cast + i * rstep_a; \ ctype* restrict c11 = c1 + i * rstep_c; \ \ /* Compute the addresses of the next panels of A and B. */ \ ctype* restrict a2 \ = bli_gemmt_get_next_a_upanel( a1, rstep_a, ir_inc ); \ \ /* Save addresses of next panels of A and B to the auxinfo_t object. */ \ bli_auxinfo_set_next_a( a2, &aux ); \ bli_auxinfo_set_next_b( b2, &aux ); \ \ /* Invoke the gemm micro-kernel. */ \ gemm_ukr \ ( \ MR, \ NR, \ k, \ alpha_cast, \ a1, \ b1, \ zero, \ ct, rs_ct, cs_ct, \ &aux, \ cntx \ ); \ \ /* Scale C and add the result to only the stored part. */ \ PASTEMAC(ch,xpbys_mxn_l)( diagoffc_ij, \ m_cur, n_cur, \ ct, rs_ct, cs_ct, \ beta_cast, \ c11, rs_c, cs_c ); \ \ ut += 1; \ if ( ut == n_ut_for_me ) return; \ } \ else if ( bli_is_strictly_below_diag_n( diagoffc_ij, m_cur, n_cur ) ) \ { \ ctype* restrict a1 = a_cast + i * rstep_a; \ ctype* restrict c11 = c1 + i * rstep_c; \ \ /* Compute the addresses of the next panels of A and B. */ \ ctype* restrict a2 \ = bli_gemmt_get_next_a_upanel( a1, rstep_a, ir_inc ); \ if ( bli_is_last_iter_tlb_l( i, m_iter ) ) \ { \ a2 = bli_gemmt_l_wrap_a_upanel( a_cast, rstep_a, \ diagoffc_j, MR, NR ); \ b2 = bli_gemmt_get_next_b_upanel( b1, cstep_b, jr_inc ); \ /* We don't bother computing b2 for the last iteration of the jr loop since the current thread won't know its j_st until the next time it calls bli_thread_range_tlb(). */ \ } \ \ /* Save addresses of next panels of A and B to the auxinfo_t object. */ \ bli_auxinfo_set_next_a( a2, &aux ); \ bli_auxinfo_set_next_b( b2, &aux ); \ \ /* Invoke the gemm micro-kernel. */ \ gemm_ukr \ ( \ m_cur, \ n_cur, \ k, \ alpha_cast, \ a1, \ b1, \ beta_cast, \ c11, rs_c, cs_c, \ &aux, \ cntx \ ); \ \ ut += 1; \ if ( ut == n_ut_for_me ) return; \ } \ else /* if ( bli_is_strictly_above_diag_n( diagoffc_ij, m_cur, n_cur ) ) */ \ { \ /* Skip ahead to the last microtile strictly above the diagonal. */ \ i = -diagoffc_j / MR - 1; \ } \ } \ \ /* Upon reaching the end of the column of microtiles, get ready to begin at the beginning of the next column (i.e., the next jr loop iteration). */ \ i = 0; \ } \ } INSERT_GENTFUNC_BASIC0( gemmt_l_ker_var2b )