6 void __attribute__((noinline
)) matmul(const int coreid
, const int ncores
, const int lda
, const data_t A
[], const data_t B
[], data_t C
[] )
8 int i
,j
,k
,a
,b
,b1
,a1
,a2
,a3
,c
,c1
,c2
,c3
,b2
,b3
;
9 for (j
=coreid
*4; j
<lda
; j
+=4*ncores
){
26 C
[i
+a1
]+=A
[a1
+k
+1]*c1
;
27 C
[i
+a2
]+=A
[a2
+k
+1]*c1
;
28 C
[i
+a3
]+=A
[a3
+k
+1]*c1
;