diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
index d354b3c..bb7a851 100644
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -141,6 +141,7 @@ CHECK_INCLUDE_FILES(stdarg.h HAVE_STDARG_H)
 CHECK_INCLUDE_FILES(signal.h HAVE_SIGNAL_H)
 CHECK_INCLUDE_FILES(execinfo.h HAVE_EXECINFO_H)
 CHECK_INCLUDE_FILES(malloc.h HAVE_MALLOC_H)
+CHECK_INCLUDE_FILES(dispatch/dispatch.h HAVE_DISPATCH_H)
 
 CONFIGURE_FILE(${CMAKE_CURRENT_SOURCE_DIR}/nvconfig.h.in ${CMAKE_CURRENT_BINARY_DIR}/nvconfig.h)
 
diff --git a/src/nvconfig.h.in b/src/nvconfig.h.in
index 9726768..0142d96 100644
--- a/src/nvconfig.h.in
+++ b/src/nvconfig.h.in
@@ -8,6 +8,7 @@
 #cmakedefine HAVE_MALLOC_H
 
 #cmakedefine HAVE_OPENMP
+#cmakedefine HAVE_DISPATCH_H
 
 #cmakedefine HAVE_PNG
 #cmakedefine HAVE_JPEG
diff --git a/src/nvmath/SimdVector_VE.h b/src/nvmath/SimdVector_VE.h
index dabb525..c850557 100644
--- a/src/nvmath/SimdVector_VE.h
+++ b/src/nvmath/SimdVector_VE.h
@@ -35,12 +35,13 @@ namespace nv {
 
     class SimdVector
     {
+	public:
         vector float vec;
 
         typedef SimdVector Arg;
 
         SimdVector() {}
-        explicit SimdVector(float v) : vec((vector float)(X)) {}	
+        explicit SimdVector(float v) : vec((vector float)(v)) {}	
         explicit SimdVector(vector float v) : vec(v) {}
         SimdVector(const SimdVector & arg) : vec(arg.vec) {}
 
@@ -115,34 +116,34 @@ namespace nv {
         }
     };
 
-    SimdVector operator+( SimdVector::Arg left, SimdVector::Arg right  )
+    inline SimdVector operator+( SimdVector::Arg left, SimdVector::Arg right  )
     {
         return SimdVector( vec_add( left.vec, right.vec ) );
     }
 
-    SimdVector operator-( SimdVector::Arg left, SimdVector::Arg right  )
+    inline SimdVector operator-( SimdVector::Arg left, SimdVector::Arg right  )
     {
         return SimdVector( vec_sub( left.vec, right.vec ) );
     }
 
-    SimdVector operator*( SimdVector::Arg left, SimdVector::Arg right  )
+    inline SimdVector operator*( SimdVector::Arg left, SimdVector::Arg right  )
     {
         return SimdVector( vec_madd( left.vec, right.vec, ( vector float )( -0.0f ) ) );
     }
 
     // Returns a*b + c
-    SimdVector multiplyAdd( SimdVector::Arg a, SimdVector::Arg b, SimdVector::Arg c )
+    inline SimdVector multiplyAdd( SimdVector::Arg a, SimdVector::Arg b, SimdVector::Arg c )
     {
         return SimdVector( vec_madd( a.vec, b.vec, c.vec ) );
     }
 
     // Returns -( a*b - c )
-    SimdVector negativeMultiplySubtract( SimdVector::Arg a, SimdVector::Arg b, SimdVector::Arg c )
+    inline SimdVector negativeMultiplySubtract( SimdVector::Arg a, SimdVector::Arg b, SimdVector::Arg c )
     {
         return SimdVector( vec_nmsub( a.vec, b.vec, c.vec ) );
     }
 
-    SimdVector reciprocal( SimdVector::Arg v )
+    inline SimdVector reciprocal( SimdVector::Arg v )
     {
         // get the reciprocal estimate
         vector float estimate = vec_re( v.vec );
@@ -152,32 +153,32 @@ namespace nv {
         return SimdVector( vec_madd( diff, estimate, estimate ) );
     }
 
-    SimdVector min( SimdVector::Arg left, SimdVector::Arg right )
+    inline SimdVector min( SimdVector::Arg left, SimdVector::Arg right )
     {
         return SimdVector( vec_min( left.vec, right.vec ) );
     }
 
-    SimdVector max( SimdVector::Arg left, SimdVector::Arg right )
+    inline SimdVector max( SimdVector::Arg left, SimdVector::Arg right )
     {
         return SimdVector( vec_max( left.vec, right.vec ) );
     }
 
-    SimdVector truncate( SimdVector::Arg v )
+    inline SimdVector truncate( SimdVector::Arg v )
     {
         return SimdVector( vec_trunc( v.vec ) );
     }
 
-    SimdVector compareEqual( SimdVector::Arg left, SimdVector::Arg right )
+    inline SimdVector compareEqual( SimdVector::Arg left, SimdVector::Arg right )
     {
         return SimdVector( ( vector float )vec_cmpeq( left.vec, right.vec ) );
     }
 
-    SimdVector select( SimdVector::Arg off, SimdVector::Arg on, SimdVector::Arg bits )
+    inline SimdVector select( SimdVector::Arg off, SimdVector::Arg on, SimdVector::Arg bits )
     {
         return SimdVector( vec_sel( off.vec, on.vec, ( vector unsigned int )bits.vec ) );
     }
 
-    bool compareAnyLessThan( SimdVector::Arg left, SimdVector::Arg right ) 
+    inline bool compareAnyLessThan( SimdVector::Arg left, SimdVector::Arg right ) 
     {
         return vec_any_lt( left.vec, right.vec ) != 0;
     }
diff --git a/src/nvtt/CompressorDXT.cpp b/src/nvtt/CompressorDXT.cpp
index 76f4ed6..a621feb 100644
--- a/src/nvtt/CompressorDXT.cpp
+++ b/src/nvtt/CompressorDXT.cpp
@@ -156,10 +156,10 @@ void FixedBlockCompressor::compress(nvtt::AlphaMode alphaMode, uint w, uint h, c
     context.compressor = this;
 
     static SequentialTaskDispatcher sequential;
-    static AppleTaskDispatcher concurrent;
+	//#static AppleTaskDispatcher concurrent;
 
-    //TaskDispatcher * dispatcher = &sequential;
-    TaskDispatcher * dispatcher = &concurrent;
+    TaskDispatcher * dispatcher = &sequential;
+    //TaskDispatcher * dispatcher = &concurrent;
 
     // Use a single thread to compress small textures.
     if (context.bh < 4) dispatcher = &sequential;
diff --git a/src/nvtt/Context.cpp b/src/nvtt/Context.cpp
index 27a7837..5e5a794 100644
--- a/src/nvtt/Context.cpp
+++ b/src/nvtt/Context.cpp
@@ -56,8 +56,6 @@
 using namespace nv;
 using namespace nvtt;
 
-#include <dispatch/dispatch.h>
-
 Compressor::Compressor() : m(*new Compressor::Private())
 {
     // CUDA initialization.
diff --git a/src/nvtt/TaskDispatcher.h b/src/nvtt/TaskDispatcher.h
index f11bd83..7dd81af 100644
--- a/src/nvtt/TaskDispatcher.h
+++ b/src/nvtt/TaskDispatcher.h
@@ -6,8 +6,7 @@
 #include <omp.h>
 #endif
 
-#if NV_OS_DARWIN
-//#if defined(HAVE_DISPATCH)
+#if NV_OS_DARWIN && defined(HAVE_DISPATCH_H)
 #include <dispatch/dispatch.h>
 #endif
 
@@ -22,7 +21,7 @@ namespace nvtt {
         }
     };
 
-#if NV_OS_DARWIN
+#if NV_OS_DARWIN && defined(HAVE_DISPATCH_H)
 
     // Task dispatcher using Apple's Grand Central Dispatch.
     struct AppleTaskDispatcher : public TaskDispatcher