make host>device only take one parameter, add 2<<< for calling cuda functions where shared-memory is 0, simplify hello-world example

extra/cuda/demos/hello-world/hello-world.factor

@@ -10,10 +10,9 @@ CUDA-FUNCTION: helloWorld ( char* string-ptr ) ;
 
 :: cuda-hello-world ( -- )
     T{ launcher { device 0 } } [
-        "Hello World!" [ - ] map-index malloc-device-string
+        "Hello World!" [ - ] map-index host>device &dispose :> str
-        &dispose dup :> str
 
-        { 6 1 1 } { 2 1 } 1 3<<< helloWorld
+        str { 6 1 1 } { 2 1 } 2<<< helloWorld
 
         str device>host utf8 alien>string print
     ] with-cuda ;

extra/cuda/memory/memory.factor

@@ -2,7 +2,7 @@
 ! See http://factorcode.org/license.txt for BSD license.
 USING: accessors alien alien.data assocs byte-arrays cuda.ffi
 cuda.utils destructors io.encodings.string io.encodings.utf8
-kernel locals namespaces sequences ;
+kernel locals namespaces sequences strings ;
 QUALIFIED-WITH: alien.c-types a
 IN: cuda.memory
 
@@ -61,14 +61,15 @@ M: cuda-memory dispose ( ptr -- )
 : memcpy-array>array ( dest-array dest-index src-array src-ptr count -- )
     cuMemcpyAtoA cuda-error ;
 
-: host>device ( dest-ptr src-ptr -- )
+GENERIC: host>device ( obj -- ptr )
-    [ ptr>> ] dip dup length cuMemcpyHtoD cuda-error ;
+
+M: string host>device utf8 encode host>device ;
+
+M: byte-array host>device ( byte-array -- ptr )
+    [ length cuda-malloc ] keep
+    [ [ ptr>> ] dip dup length cuMemcpyHtoD cuda-error ]
+    [ drop ] 2bi ;
 
 :: device>host ( ptr -- seq )
     ptr byte-length <byte-array>
     [ ptr [ ptr>> ] [ byte-length ] bi cuMemcpyDtoH cuda-error ] keep ;
-
-: malloc-device-string ( string -- n )
-    utf8 encode
-    [ length cuda-malloc ] keep
-    [ host>device ] [ drop ] 2bi ;

extra/cuda/syntax/syntax.factor

@@ -13,6 +13,9 @@ SYNTAX: CUDA-FUNCTION:
     scan [ create-in current-cuda-library get ] [ ] bi
     ";" scan-c-args drop define-cuda-word ;
 
+: 2<<< ( dim-block dim-grid -- function-launcher )
+    0 f function-launcher boa ;
+
 : 3<<< ( dim-block dim-grid shared-size -- function-launcher )
     f function-launcher boa ;