Clean up canonical mt benchmarks and reorganize extra versions in /mt. All versions support support at least 1/2/4 threads.

1 files changed, 35 insertions, 0 deletions

diff --git a/mt/bb_matmul.c b/mt/bb_matmul.c
new file mode 100755
index 0000000..c335ed3
--- /dev/null
+++ b/mt/bb_matmul.c
@@ -0,0 +1,35 @@
+#include "dataset.h"
+void __attribute__((noinline)) matmul(const int coreid, const int ncores, const int lda,  const data_t A[], const data_t B[], data_t C[] )
+{
+  int i, j, k;
+
+  for (i = 0; i < lda; i += 2) {
+    for (j = coreid * (lda / ncores); j < (coreid + 1) * (lda / ncores); j += 4) {
+      register data_t c00 = 0, c01 = 0;
+      register data_t c10 = 0, c11 = 0;
+      register data_t c20 = 0, c21 = 0;
+      register data_t c30 = 0, c31 = 0;
+
+      register data_t a0, a1, a2, a3, b0, b1;
+      for (k = 0; k < lda; k++) {
+        a0 = A[j*lda + k + 0*lda];
+        a1 = A[j*lda + k + 1*lda];
+        a2 = A[j*lda + k + 2*lda];
+        a3 = A[j*lda + k + 3*lda];
+
+        b0 = B[k*lda + i + 0];
+        b1 = B[k*lda + i + 1];
+
+        c00 += a0 * b0; c01 += a0 * b1;
+        c10 += a1 * b0; c11 += a1 * b1;
+        c20 += a2 * b0; c21 += a2 * b1;
+        c30 += a3 * b0; c31 += a3 * b1;
+      }
+
+      C[i + j*lda + 0 + 0*lda] = c00; C[i + j*lda + 1 + 0*lda] = c01;
+      C[i + j*lda + 0 + 1*lda] = c10; C[i + j*lda + 1 + 1*lda] = c11;
+      C[i + j*lda + 0 + 2*lda] = c20; C[i + j*lda + 1 + 2*lda] = c21;
+      C[i + j*lda + 0 + 3*lda] = c30; C[i + j*lda + 1 + 3*lda] = c31;
+    }
+  }
+}