From 72361b3afebe0fae5abcb8b294b3d480d6a54e66 Mon Sep 17 00:00:00 2001
From: Euna Kim <eunakim@gatech.edu>
Date: Mon, 25 Nov 2019 20:53:41 -0500
Subject: [PATCH] merging error fixed

---
 benchmarks/vector/sgemm_nn/vx_vec_sgemm_nn.c | 10 ++++------
 benchmarks/vector/sgemm_nn/vx_vec_sgemm_nn.s | 12 +++++++-----
 2 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/benchmarks/vector/sgemm_nn/vx_vec_sgemm_nn.c b/benchmarks/vector/sgemm_nn/vx_vec_sgemm_nn.c
index 467b29a6..e12e244e 100644
--- a/benchmarks/vector/sgemm_nn/vx_vec_sgemm_nn.c
+++ b/benchmarks/vector/sgemm_nn/vx_vec_sgemm_nn.c
@@ -53,8 +53,6 @@ int main()
        for (int i = 0; i < w; i=+4) {
            for (int m = 0; m < d; m++) {
                  vx_vec_sgemm_nn(i, m, n, a1, b1, c1, ldc, vsize);
-                //d1[i+n*ldc] += a1[m+n*ldc]*b1[m*ldc+i];
-                 vx_vec_sgemm_nn(i, r, c, a1, b1, c1, ldc, vsize);
                  i = i + vsize;
            }
        }
@@ -77,10 +75,10 @@ int main()
     }
 
 #if 1
-   printf("\n\nc[%d]:\n", m*n);
-   for(int i = 0; i < m; ++i) {
-      for(int j = 0; j < n; ++j) {
-          printf("%d ", d1[i*m+j]);
+   printf("\n\nc[%d]:\n", w*d);
+   for(int i = 0; i < w; ++i) {
+      for(int j = 0; j < d; ++j) {
+          printf("%d ", d1[i*w+j]);
       }
       printf("\n");
     }
diff --git a/benchmarks/vector/sgemm_nn/vx_vec_sgemm_nn.s b/benchmarks/vector/sgemm_nn/vx_vec_sgemm_nn.s
index 639676ba..7d88bc01 100644
--- a/benchmarks/vector/sgemm_nn/vx_vec_sgemm_nn.s
+++ b/benchmarks/vector/sgemm_nn/vx_vec_sgemm_nn.s
@@ -16,17 +16,19 @@ vx_vec_sgemm_nn:
     vsetvli t0, a7, e32 # <--- vsize 
     mul x11, a6, a2  # n*ldc
     add x12, x11, a1  # i + (n*ldc)
+    slli x12, x12, 2 
     add a3, x12, a3  # a[i+ n*ldc]
     lw x13, (a3)
 
     mul x14, a1, a6  # m*ldc
     add x15, a0, x14  # i + m*ldc
+    slli x15, x15, 2
     add a4, x15, a4  # b[i + m*ldc]
     vlw.v v0, (a4)
     vmul.vx v2, v1, x13
 ##   lw x6, (a4)
-#     lw x10, (a4)  # b
-#     mul x11, x3, x10
+##   lw x10, (a4)  # b
+##   mul x11, x3, x10
 
     mul x6, a2, a6  # n*ldc
     add x7, a0, x6  # i + n*ldc
@@ -34,7 +36,7 @@ vx_vec_sgemm_nn:
     vlw.v v3, (a5) # c
     vadd.vv v3, v3, v2
     vsw.v v3, (a5)
-#    lw x12, (a5)
-#    add x12, x12, x11
-#    sw x12, (a5)    
+##   lw x12, (a5)
+##   add x12, x12, x11
+##   sw x12, (a5)    
     ret