refactor: namespace csrc kernels and extract common helpers

- attention family -> astrai::attention; fp8 family -> astrai::fp8
- new common/reduce.cuh (warp/group reductions, atomic_max_float)
- new common/cp_async.cuh (predicated cp_async_16, commit/wait group)
- move MAX_SPLITS into attention/common.h; delete warp_utils.cuh
- .cu bindings and pure C tests open family namespaces via using

csrc/kernels/fp8/common.h

@@ -5,6 +5,12 @@
 #include <cuda_runtime.h>
 #include <cstdint>
 
+// Pure POD/traits header — no .cuh/CUDA-kernel includes; raw __nv_* type
+// spellings only.
+
+namespace astrai {
+namespace fp8 {
+
 // Compile-time FP8 format: E4M3 (forward / high precision, max 448) or
 // E5M2 (gradient / large dynamic range, max 57344).
 enum class FP8Format : int {
@@ -64,3 +70,6 @@ struct FP8Params {
 
     int total;
 };
+
+}  // namespace fp8
+}  // namespace astrai