1 /*********************************************************************/ 2 /* Copyright 2009, 2010 The University of Texas at Austin. */ 3 /* All rights reserved. */ 4 /* */ 5 /* Redistribution and use in source and binary forms, with or */ 6 /* without modification, are permitted provided that the following */ 7 /* conditions are met: */ 8 /* */ 9 /* 1. Redistributions of source code must retain the above */ 10 /* copyright notice, this list of conditions and the following */ 11 /* disclaimer. */ 12 /* */ 13 /* 2. Redistributions in binary form must reproduce the above */ 14 /* copyright notice, this list of conditions and the following */ 15 /* disclaimer in the documentation and/or other materials */ 16 /* provided with the distribution. */ 17 /* */ 18 /* THIS SOFTWARE IS PROVIDED BY THE UNIVERSITY OF TEXAS AT */ 19 /* AUSTIN ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, */ 20 /* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF */ 21 /* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE */ 22 /* DISCLAIMED. IN NO EVENT SHALL THE UNIVERSITY OF TEXAS AT */ 23 /* AUSTIN OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, */ 24 /* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES */ 25 /* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE */ 26 /* GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR */ 27 /* BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF */ 28 /* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT */ 29 /* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT */ 30 /* OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE */ 31 /* POSSIBILITY OF SUCH DAMAGE. */ 32 /* */ 33 /* The views and conclusions contained in the software and */ 34 /* documentation are those of the authors and should not be */ 35 /* interpreted as representing official policies, either expressed */ 36 /* or implied, of The University of Texas at Austin. */ 37 /*********************************************************************/ 38 39 #include <stdio.h> 40 #include "common.h" 41 42 #undef TIMING 43 44 #define BETA_OPERATION(M_FROM, M_TO, N_FROM, N_TO, BETA, C, LDC) \ 45 GEMM_BETA((M_TO) - (M_FROM), (N_TO - N_FROM), 0, \ 46 BETA[0], BETA[1], NULL, 0, NULL, 0, \ 47 (FLOAT *)(C) + (M_FROM) + (N_FROM) * (LDC) * COMPSIZE, LDC) 48 49 #ifndef RSIDE 50 #ifndef LOWER 51 #define ICOPYB_OPERATION(M, N, A, LDA, X, Y, BUFFER) SYMM3M_IUCOPYB(M, N, A, LDA, Y, X, BUFFER) 52 #define ICOPYR_OPERATION(M, N, A, LDA, X, Y, BUFFER) SYMM3M_IUCOPYR(M, N, A, LDA, Y, X, BUFFER) 53 #define ICOPYI_OPERATION(M, N, A, LDA, X, Y, BUFFER) SYMM3M_IUCOPYI(M, N, A, LDA, Y, X, BUFFER) 54 #else 55 #define ICOPYB_OPERATION(M, N, A, LDA, X, Y, BUFFER) SYMM3M_ILCOPYB(M, N, A, LDA, Y, X, BUFFER) 56 #define ICOPYR_OPERATION(M, N, A, LDA, X, Y, BUFFER) SYMM3M_ILCOPYR(M, N, A, LDA, Y, X, BUFFER) 57 #define ICOPYI_OPERATION(M, N, A, LDA, X, Y, BUFFER) SYMM3M_ILCOPYI(M, N, A, LDA, Y, X, BUFFER) 58 #endif 59 #endif 60 61 #ifdef RSIDE 62 #ifndef LOWER 63 #define OCOPYB_OPERATION(M, N, A, LDA, ALPHA_R, ALPHA_I, X, Y, BUFFER) \ 64 SYMM3M_OUCOPYB(M, N, A, LDA, Y, X, ALPHA_R, ALPHA_I, BUFFER) 65 #define OCOPYR_OPERATION(M, N, A, LDA, ALPHA_R, ALPHA_I, X, Y, BUFFER) \ 66 SYMM3M_OUCOPYR(M, N, A, LDA, Y, X, ALPHA_R, ALPHA_I, BUFFER) 67 #define OCOPYI_OPERATION(M, N, A, LDA, ALPHA_R, ALPHA_I, X, Y, BUFFER) \ 68 SYMM3M_OUCOPYI(M, N, A, LDA, Y, X, ALPHA_R, ALPHA_I, BUFFER) 69 #else 70 #define OCOPYB_OPERATION(M, N, A, LDA, ALPHA_R, ALPHA_I, X, Y, BUFFER) \ 71 SYMM3M_OLCOPYB(M, N, A, LDA, Y, X, ALPHA_R, ALPHA_I, BUFFER) 72 #define OCOPYR_OPERATION(M, N, A, LDA, ALPHA_R, ALPHA_I, X, Y, BUFFER) \ 73 SYMM3M_OLCOPYR(M, N, A, LDA, Y, X, ALPHA_R, ALPHA_I, BUFFER) 74 #define OCOPYI_OPERATION(M, N, A, LDA, ALPHA_R, ALPHA_I, X, Y, BUFFER) \ 75 SYMM3M_OLCOPYI(M, N, A, LDA, Y, X, ALPHA_R, ALPHA_I, BUFFER) 76 #endif 77 #endif 78 79 #ifndef RSIDE 80 #define K args -> m 81 #ifndef LOWER 82 #define GEMM3M_LOCAL SYMM3M_LU 83 #else 84 #define GEMM3M_LOCAL SYMM3M_LL 85 #endif 86 #else 87 #define K args -> n 88 #ifndef LOWER 89 #define GEMM3M_LOCAL SYMM3M_RU 90 #else 91 #define GEMM3M_LOCAL SYMM3M_RL 92 #endif 93 #endif 94 95 #ifdef THREADED_LEVEL3 96 #include "level3_gemm3m_thread.c" 97 #else 98 #include "gemm3m_level3.c" 99 #endif 100 101