diff --git a/dsv4/kernels/attention/fmha.py b/dsv4/kernels/attention/fmha.py
index 0585724a..86829eed 100644
--- a/dsv4/kernels/attention/fmha.py
+++ b/dsv4/kernels/attention/fmha.py
@@ -288,6 +288,8 @@ class FmhaKernel:
             #   sP_addr = 64*m + (k%16) + 16*((k//16)%4) + 8192*(k//64)
             # This is representable as a CuTe layout: (128, (16, 4, 2)) -> (64, (1, 16, 8192))
             _sP_nostage = sP[(None, None, None, 0)]  # remove stage dim
+            print(f"SMEM-P DEBUG: sP_nostage shape={cute.shape(_sP_nostage)}, sP outer={cute.shape(sP)}")
+            print(f"SMEM-P DEBUG: pv_n_tile={self.pv_n_tile}, p_cols_fp32={p_cols_fp32}")
 
             row_max = -Float32.inf
             row_sum = Float32(0.0)