diff --git a/benchmarks/triton_kernels_benchmark/gemm_benchmark.py b/benchmarks/triton_kernels_benchmark/gemm_benchmark.py
index 4ad3d8d5e5..ac45d629d9 100644
--- a/benchmarks/triton_kernels_benchmark/gemm_benchmark.py
+++ b/benchmarks/triton_kernels_benchmark/gemm_benchmark.py
@@ -305,11 +305,9 @@ def benchmark(B, M, N, K, provider):
     elif provider == 'xetla':
         if B == 1:
             c = torch.zeros((M, N), device='xpu', dtype=torch.float32)
-            acc = torch.zeros((M, N), device='xpu', dtype=torch.float32)
             cnt = torch.zeros((M, N), device='xpu', dtype=torch.int32)
         else:
             c = torch.zeros((B, M, N), device='xpu', dtype=torch.float32)
-            acc = torch.zeros((B, M, N), device='xpu', dtype=torch.float32)
             cnt = torch.zeros((B, M, N), device='xpu', dtype=torch.int32)
         name = f'gemm_shape_{B}_{M}_{K}_{N}'
         # FIXME: Use gemm_streamk_benchmark.py when Triton streamk can get
@@ -317,7 +315,18 @@ def benchmark(B, M, N, K, provider):
         if (B, M, N, K) == (1, 3072, 3072, 4096):
             name = 'gemm_streamk_shape_3072_4096_3072'
         func = getattr(xetla_kernel, name)
-        xetla_fn = lambda: func(a, b, c, acc, cnt)
+
+
+        def xetla_func_with_acc_allocation():
+            # allocating `acc` matrix on every function call, to be as similar as
+            # possible to the triton kernel, which also does this on every call.
+            if B == 1:
+                acc = torch.zeros((M, N), device='xpu', dtype=torch.float32)
+            else:
+                acc = torch.zeros((B, M, N), device='xpu', dtype=torch.float32)
+            return func(a, b, c, acc, cnt)
+
+        xetla_fn = xetla_func_with_acc_allocation
         torch_fn = lambda: torch.matmul(a, b).to(torch.float32)
 
         # benchmark_suit.assert_close(xetla_fn(), torch_fn(), atol=1e-4, rtol=1.0, err_msg='xetla to torch')