add groupsize to the exported device API

Author: vchuravy

docs/src/api.md

@@ -9,6 +9,7 @@
 @localmem
 @private
 @synchronize
+groupsize
 ```
 
 ## Host interface

src/KernelAbstractions.jl

@@ -1,7 +1,7 @@
 module KernelAbstractions
 
 export @kernel
-export @Const, @localmem, @private, @synchronize, @index
+export @Const, @localmem, @private, @synchronize, @index, groupsize
 export Device, GPU, CPU, CUDA 
 
 using StaticArrays
@@ -69,8 +69,16 @@ function async_copy! end
 # - @private
 # - @synchronize
 # - @index
+# - groupsize
 ###
 
+"""
+    groupsize()
+
+Query the workgroupsize on the device.
+"""
+function groupsize end
+
 const shmem_id = Ref(0)
 
 """
@@ -256,8 +264,6 @@ include("compiler.jl")
 # Compiler/Frontend
 ###
 
-function groupsize end
-
 @inline function __workitems_iterspace()
     return 1:groupsize()
 end