PLASMA  2.4.5
PLASMA - Parallel Linear Algebra for Scalable Multi-core Architectures
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups
time_cgemm_tile.c
Go to the documentation of this file.
1 
6 #define _TYPE PLASMA_Complex32_t
7 #define _PREC float
8 #define _LAMCH LAPACKE_slamch_work
9 
10 #define _NAME "PLASMA_cgemm_Tile"
11 /* See Lawn 41 page 120 */
12 #define _FMULS FMULS_GEMM(M, N, K)
13 #define _FADDS FADDS_GEMM(M, N, K)
14 
15 #include "./timing.c"
16 
17 static int
18 RunTest(int *iparam, float *dparam, real_Double_t *t_)
19 {
20  PLASMA_Complex32_t alpha, beta;
21  PASTE_CODE_IPARAM_LOCALS( iparam );
22 
23  LDB = max(K, iparam[IPARAM_LDB]);
24  LDC = max(M, iparam[IPARAM_LDC]);
25 
26  /* Allocate Data */
30 
31  /* Initialiaze Data */
32  PLASMA_cplrnt_Tile( descA, 5373 );
33  PLASMA_cplrnt_Tile( descB, 7672 );
34  PLASMA_cplrnt_Tile( descC, 6387 );
35 
36  LAPACKE_clarnv_work(1, ISEED, 1, &alpha);
37  LAPACKE_clarnv_work(1, ISEED, 1, &beta);
38 
39  /* Save C for check */
40  PASTE_TILE_TO_LAPACK( descC, C2, check, PLASMA_Complex32_t, LDC, N );
41 
42  START_TIMING();
43  PLASMA_cgemm_Tile( PlasmaNoTrans, PlasmaNoTrans, alpha, descA, descB, beta, descC );
44  STOP_TIMING();
45 
46  /* Check the solution */
47  if (check)
48  {
49  PASTE_TILE_TO_LAPACK( descA, A, check, PLASMA_Complex32_t, LDA, K );
50  PASTE_TILE_TO_LAPACK( descB, B, check, PLASMA_Complex32_t, LDB, N );
51  PASTE_TILE_TO_LAPACK( descC, C, check, PLASMA_Complex32_t, LDC, N );
52 
54  alpha, A, LDA, B, LDB, beta, C, C2, LDC,
55  &(dparam[IPARAM_ANORM]),
56  &(dparam[IPARAM_BNORM]),
57  &(dparam[IPARAM_XNORM]));
58  free(C2);
59  }
60 
61  PASTE_CODE_FREE_MATRIX( descA );
62  PASTE_CODE_FREE_MATRIX( descB );
63  PASTE_CODE_FREE_MATRIX( descC );
64  return 0;
65 }