diff --git a/csrc/quantization/cutlass_w8a8/broadcast_load_epilogue_c3x.hpp b/csrc/quantization/cutlass_w8a8/broadcast_load_epilogue_c3x.hpp
index 8f38bbf50790..877a9f5b9e5d 100644
--- a/csrc/quantization/cutlass_w8a8/broadcast_load_epilogue_c3x.hpp
+++ b/csrc/quantization/cutlass_w8a8/broadcast_load_epilogue_c3x.hpp
@@ -153,7 +153,7 @@ struct Sm90RowOrScalarBroadcast {
 
     CUTLASS_DEVICE void
     begin(uint64_t* full_mbarrier_ptr, int load_iteration, bool issue_tma_load) {
-      if (params.ptr_row == nullptr) {
+      if (!params.row_broadcast) {
         return;
       }
 
diff --git a/vllm/model_executor/layers/quantization/fp8.py b/vllm/model_executor/layers/quantization/fp8.py
index bc08bfcc32b3..e89fd65813c0 100644
--- a/vllm/model_executor/layers/quantization/fp8.py
+++ b/vllm/model_executor/layers/quantization/fp8.py
@@ -257,9 +257,7 @@ def apply(self,
         #   If dynamic, layer.input_scale is None and x_scale computed from x.
         #   If static, layer.input_scale is scalar and x_scale is input_scale.
 
-        # Temporarily disable CUTLASS kernels due to an illegal memory access
-        #if  bias is None and self.cutlass_fp8_supported:
-        if False:
+        if bias is None and self.cutlass_fp8_supported:
             qinput, x_scale = ops.scaled_fp8_quant(x, layer.input_scale)
 
             # Fused GEMM_DQ