diff --git a/tests/python/contrib/test_hexagon/hmx_qnn_conv2d_generator.py b/tests/python/contrib/test_hexagon/hmx_qnn_conv2d_generator.py
index 137c056f99305..d52e4d5bcaeb2 100644
--- a/tests/python/contrib/test_hexagon/hmx_qnn_conv2d_generator.py
+++ b/tests/python/contrib/test_hexagon/hmx_qnn_conv2d_generator.py
@@ -137,7 +137,9 @@ def make_hexagon_conv2d_quantized_no_layout_transform_nhwc(
 
     BH = AH + PH_low + PH_high - FH + 1
     BW = AW + PW_low + PW_high - FW + 1
-    output_shape = [1, K, BH // strides[0], BW // strides[1]]
+    SBH = (AH + PH_low + PH_high - FH) // strides[0] + 1
+    SBW = (AW + PW_low + PW_high - FW) // strides[1] + 1
+    output_shape = [1, K, SBH, SBW]
 
     BC_pad_w_low = (AC_pad_w_low - PW_low + (FW - 1)) % TW
     BC_pad_h_low = (AC_pad_h_low - PH_low) % TH
@@ -254,18 +256,15 @@ def output_tiled_to_NCHW(B_handle: T.handle, BC_handle: T.handle):
             B = T.match_buffer(B_handle, shape=output_shape, dtype=b_dtype)
             BC = T.match_buffer(BC_handle, shape=transformed_output_shape, dtype=B.dtype, scope=mem_scope)
             with T.block("output_tiled_to_NCHW"):
-                for n, k, bh, bw in T.grid(1, K, BH, BW):
+                for n, k, sbh, sbw in T.grid(1, K, SBH, SBW):
                     ko = k // 32
                     ki = k % 32
 
-                    sbh = T.floordiv(bh, strides[0])
-                    sbw = T.floordiv(bw, strides[1])
+                    bho = (sbh * strides[0] + BC_pad_h_low) // TH
+                    bhi = (sbh * strides[0] + BC_pad_h_low) % TH
 
-                    bho = (sbh + BC_pad_h_low) // TH
-                    bhi = (sbh + BC_pad_h_low) % TH
-
-                    bwo = (sbw + BC_pad_w_low) // TW
-                    bwi = (sbw + BC_pad_w_low) % TW
+                    bwo = (sbw * strides[1] + BC_pad_w_low) // TW
+                    bwi = (sbw * strides[1] + BC_pad_w_low) % TW
 
                     B[n, k, sbh, sbw] = BC[n, bho, bwo, ko, bhi, bwi, ki]