diff --git a/cuda_ffi/bindings.ml b/cuda_ffi/bindings.ml
index 428cfeb..ee484ed 100644
--- a/cuda_ffi/bindings.ml
+++ b/cuda_ffi/bindings.ml
@@ -102,6 +102,17 @@ module Functions (F : Ctypes.FOREIGN) = struct
   let cu_memset_d32 =
     F.foreign "cuMemsetD32" F.(cu_deviceptr @-> uint32_t @-> size_t @-> returning E.cu_result)
 
+  let cu_memset_d8_async =
+    F.foreign "cuMemsetD8Async" F.(cu_deviceptr @-> uchar @-> size_t @-> cu_stream @-> returning E.cu_result)
+
+  let cu_memset_d16_async =
+    F.foreign "cuMemsetD16Async"
+      F.(cu_deviceptr @-> ushort @-> size_t @-> cu_stream @-> returning E.cu_result)
+
+  let cu_memset_d32_async =
+    F.foreign "cuMemsetD32Async"
+      F.(cu_deviceptr @-> uint32_t @-> size_t @-> cu_stream @-> returning E.cu_result)
+
   let cu_module_get_global =
     F.foreign "cuModuleGetGlobal_v2"
       F.(ptr cu_deviceptr @-> ptr size_t @-> cu_module @-> string @-> returning E.cu_result)
diff --git a/cudajit.ml b/cudajit.ml
index 2164e71..221d552 100644
--- a/cudajit.ml
+++ b/cudajit.ml
@@ -499,6 +499,15 @@ let memset_d16 (Deviceptr dev) v ~length =
 let memset_d32 (Deviceptr dev) v ~length =
   check "cu_memset_d32" @@ Cuda.cu_memset_d32 dev v @@ Unsigned.Size_t.of_int length
 
+let memset_d8_async (Deviceptr dev) v ~length stream =
+  check "cu_memset_d8_async" @@ Cuda.cu_memset_d8_async dev v (Unsigned.Size_t.of_int length) stream
+
+let memset_d16_async (Deviceptr dev) v ~length stream =
+  check "cu_memset_d16_async" @@ Cuda.cu_memset_d16_async dev v (Unsigned.Size_t.of_int length) stream
+
+let memset_d32_async (Deviceptr dev) v ~length stream =
+  check "cu_memset_d32_async" @@ Cuda.cu_memset_d32_async dev v (Unsigned.Size_t.of_int length) stream
+
 let module_get_global module_ ~name =
   let open Ctypes in
   let device = allocate_n cu_deviceptr ~count:1 in