trivial bugfixes, typo and javadoc warning fixes.

began to switch to gluegen's libloading infrastructure.
added CL extensions accessors to CLPlatform.
optimized isFooEnabled() methods for CLCommandQueue.

14 files changed, 119 insertions, 16 deletions

diff --git a/nbproject/project.properties b/nbproject/project.properties
index eab4a1e4..8622d80a 100644
--- a/nbproject/project.properties
+++ b/nbproject/project.properties
@@ -40,7 +40,7 @@ javac.classpath=\
     ${file.reference.newt.all.jar}:\
     ${file.reference.nativewindow.all.jar}
 # Space-separated list of extra javac options
-javac.compilerargs=
+javac.compilerargs=-Xlint
 javac.deprecation=true
 javac.source=1.5
 javac.target=1.5
diff --git a/resources/cl-common.cfg b/resources/cl-common.cfg
index 6a42f300..79c2845c 100644
--- a/resources/cl-common.cfg
+++ b/resources/cl-common.cfg
@@ -26,6 +26,10 @@ Ignore CL_LONG_MAX
 Ignore CL_LONG_MIN
 Ignore CL_ULONG_MAX
 
+
+#use CLException instead of RunTimeException
+RuntimeExceptionType CLException
+
 #enforce client side "good behavior" by generating direct-memory-only bindings for
 #performance critical functions.
 #NioDirectOnly __ALL__
@@ -78,6 +82,7 @@ NioDirectOnly clWaitForEvents
 #extensions
 NioDirectOnly clGetGLContextInfoKHR
 
+
 #common rename emitted struct accessors
 #struct cl_image_format
 RenameJavaType cl_image_format CLImageFormatImpl
@@ -87,10 +92,12 @@ StructPackage cl_image_format com.mbien.opencl.impl
 RenameJavaMethod image_channel_order imageChannelOrder
 RenameJavaMethod image_channel_data_type imageChannelDataType
 
+
 ClassJavadoc CLImageFormatImpl /**
 ClassJavadoc CLImageFormatImpl  * Struct accessor for cl_image_format.
 ClassJavadoc CLImageFormatImpl  * @author Michael Bien
 ClassJavadoc CLImageFormatImpl  */
 
+
 # Pick up on-line OpenCL doc and link it with the javadoc
 TagNativeBinding true
\ No newline at end of file
diff --git a/resources/cl-impl.cfg b/resources/cl-impl.cfg
index ec2718df..ac67e482 100644
--- a/resources/cl-impl.cfg
+++ b/resources/cl-impl.cfg
@@ -22,6 +22,10 @@ GetProcAddressTableExpr addressTable
 ProcAddressNameExpr $UpperCase(arg)
 ForceProcAddressGen clGetGLContextInfoKHR
 
+Unignore clGetExtensionFunctionAddress
+#AccessControl clGetExtensionFunctionAddress PACKAGE_PRIVATE
+ArgumentIsString clGetExtensionFunctionAddress 0
+
 #append to generated c files
 CustomCCode #include <CL/cl.h>
 CustomCCode #include <GL3/gl3.h>
diff --git a/resources/clImplCustomCode.java b/resources/clImplCustomCode.java
index 6ee9193b..814bb219 100644
--- a/resources/clImplCustomCode.java
+++ b/resources/clImplCustomCode.java
@@ -1,5 +1,9 @@
 
-    CLProcAddressTable addressTable = new CLProcAddressTable();
+    final static CLProcAddressTable addressTable = new CLProcAddressTable();
+
+//    static{
+//        ProcAddressHelper.resetProcAddressTable(addressTable, );
+//    }
 
     public long clCreateContext(PointerBuffer properties, PointerBuffer devices, CreateContextCallback pfn_notify, Object userData, IntBuffer errcode_ret) {
 
diff --git a/src/com/mbien/opencl/CLCommandQueue.java b/src/com/mbien/opencl/CLCommandQueue.java
index d7586a93..0967a27b 100644
--- a/src/com/mbien/opencl/CLCommandQueue.java
+++ b/src/com/mbien/opencl/CLCommandQueue.java
@@ -20,9 +20,11 @@ import static com.mbien.opencl.CL.*;
 public class CLCommandQueue implements CLResource {
 
     public final long ID;
+
     private final CLContext context;
     private final CLDevice device;
     private final CL cl;
+    private long properties;
 
     /*
      * Those direct memory buffers are used to pass data between the JVM and OpenCL.
@@ -35,6 +37,7 @@ public class CLCommandQueue implements CLResource {
         this.context = context;
         this.cl = context.cl;
         this.device = device;
+        this.properties = properties;
 
         this.bufferA = PointerBuffer.allocateDirect(3);
         this.bufferB = PointerBuffer.allocateDirect(3);
@@ -450,7 +453,7 @@ public class CLCommandQueue implements CLResource {
         if(localWorkSizeX != 0 && localWorkSizeY !=0) {
             localWorkSize = copy2NIO(bufferC, localWorkSizeX, localWorkSizeY);
         }
-        this.putNDRangeKernel(kernel, 2, globalWorkOffset, globalWorkSize, localWorkSize);
+        this.putNDRangeKernel(kernel, 2, globalWorkOffset, globalWorkSize, localWorkSize, events);
         return this;
     }
 
@@ -525,13 +528,26 @@ public class CLCommandQueue implements CLResource {
         return this;
     }
 
-
     public CLCommandQueue finish() {
         int ret = cl.clFinish(ID);
         checkForError(ret, "can not finish command queue");
         return this;
     }
 
+    /**
+     * Returns true only when {@link Mode#PROFILING_MODE} has been enabled.
+     */
+    public boolean isProfilingEnabled() {
+        return (Mode.PROFILING_MODE.QUEUE_MODE & properties) != 0;
+    }
+
+    /**
+     * Returns true only when {@link Mode#OUT_OF_ORDER_EXEC_MODE} mode has been enabled.
+     */
+    public boolean isOutOfOrderModeEnabled() {
+        return (Mode.OUT_OF_ORDER_EXEC_MODE.QUEUE_MODE & properties) != 0;
+    }
+
     public void release() {
         int ret = cl.clReleaseCommandQueue(ID);
         context.onCommandQueueReleased(device, this);
diff --git a/src/com/mbien/opencl/CLDevice.java b/src/com/mbien/opencl/CLDevice.java
index e754a4bc..c6134248 100644
--- a/src/com/mbien/opencl/CLDevice.java
+++ b/src/com/mbien/opencl/CLDevice.java
@@ -263,14 +263,14 @@ public final class CLDevice {
     /**
      * Returns the size of global memory cache line in bytes.
      */
-    public long getGlobalMemCachlineSize() {
+    public long getGlobalMemCachelineSize() {
         return deviceInfo.getLong(CL_DEVICE_GLOBAL_MEM_CACHELINE_SIZE);
     }
 
     /**
      * Returns the size of global memory cache in bytes.
      */
-    public long getGlobalMemCachSize() {
+    public long getGlobalMemCacheSize() {
         return deviceInfo.getLong(CL_DEVICE_GLOBAL_MEM_CACHE_SIZE);
     }
 
@@ -466,6 +466,13 @@ public final class CLDevice {
     }
 
     /**
+     * Returns true if the extension is supported on this device.
+     */
+    public boolean isExtensionAvailable(String extension) {
+        return getExtensions().contains(extension);
+    }
+
+    /**
      * Returns all device extension names as unmodifiable Set.
      */
     public Set<String> getExtensions() {
diff --git a/src/com/mbien/opencl/CLException.java b/src/com/mbien/opencl/CLException.java
index 3703f235..4ca216c4 100644
--- a/src/com/mbien/opencl/CLException.java
+++ b/src/com/mbien/opencl/CLException.java
@@ -12,6 +12,11 @@ public class CLException extends RuntimeException {
 
     private final static String ERROR_CODE_DOC = "http://www.khronos.org/opencl/sdk/1.0/docs/man/xhtml/errors.html";
 
+    public CLException(String message) {
+        super(message);
+        errorcode = 0;
+    }
+
     public CLException(int error, String message) {
         super(message + "\nerror: " + identifyError(error) + " (man page: "+ERROR_CODE_DOC+")");
         errorcode = error;
diff --git a/src/com/mbien/opencl/CLImageFormat.java b/src/com/mbien/opencl/CLImageFormat.java
index b5551798..d2625f7f 100644
--- a/src/com/mbien/opencl/CLImageFormat.java
+++ b/src/com/mbien/opencl/CLImageFormat.java
@@ -172,7 +172,7 @@ public final class CLImageFormat {
 
         /**
          * Represents a normalized 5-6-5 3-channel RGB image. The channel order must
-         * be {@link ChannelOrder#CL_RGB}.
+         * be {@link ChannelOrder#RGB}.
          */
         UNORM_SHORT_565(CL_UNORM_SHORT_565),
 
diff --git a/src/com/mbien/opencl/CLKernel.java b/src/com/mbien/opencl/CLKernel.java
index cb03cf8e..0dc9db19 100644
--- a/src/com/mbien/opencl/CLKernel.java
+++ b/src/com/mbien/opencl/CLKernel.java
@@ -12,11 +12,11 @@ import static com.mbien.opencl.CL.*;
 
 /**
  * High level abstraction for an OpenCL Kernel.
- * CLKernel is not threadsafe.
- * "A kernel is a function declared in a program. A kernel is identified by the <code>kernel</code> qualifier
+ * A kernel is a function declared in a program. A kernel is identified by the <code>kernel</code> qualifier
  * applied to any function in a program. A kernel object encapsulates the specific <code>kernel</code>
  * function declared in a program and the argument values to be used when executing this
- * <code>kernel</code> function."
+ * <code>kernel</code> function.
+ * CLKernel is not threadsafe.
  * @author Michael Bien
  */
 public class CLKernel implements CLResource/*, Cloneable*/ {
diff --git a/src/com/mbien/opencl/CLPlatform.java b/src/com/mbien/opencl/CLPlatform.java
index fc1b78f3..2f3cbf54 100644
--- a/src/com/mbien/opencl/CLPlatform.java
+++ b/src/com/mbien/opencl/CLPlatform.java
@@ -2,9 +2,14 @@ package com.mbien.opencl;
 
 import com.mbien.opencl.impl.CLImpl;
 import com.sun.gluegen.runtime.PointerBuffer;
+import com.sun.gluegen.runtime.ProcAddressHelper;
 import java.nio.ByteBuffer;
 import java.nio.ByteOrder;
 import java.nio.IntBuffer;
+import java.util.Collections;
+import java.util.HashSet;
+import java.util.Scanner;
+import java.util.Set;
 
 import static com.mbien.opencl.CLException.*;
 import static com.mbien.opencl.CL.*;
@@ -22,9 +27,12 @@ public final class CLPlatform {
 
     private static final CL cl;
 
+    private Set<String> extensions;
+
     static{
-        System.loadLibrary("gluegen-rt");
-        System.loadLibrary("jocl");
+        NativeLibLoader.loadJOCL();
+//        System.loadLibrary("gluegen-rt");
+//        ProcAddressHelper.resetProcAddressTable(table, null);
         cl = new CLImpl();
     }
 
@@ -186,6 +194,32 @@ public final class CLPlatform {
     }
 
     /**
+     * Returns true if the extension is supported on this platform.
+     */
+    public boolean isExtensionAvailable(String extension) {
+        return getExtensions().contains(extension);
+    }
+
+    /**
+     * Returns all platform extension names as unmodifiable Set.
+     */
+    public Set<String> getExtensions() {
+
+        if(extensions == null) {
+            extensions = new HashSet<String>();
+            String ext = getInfoString(CL_PLATFORM_EXTENSIONS);
+            Scanner scanner = new Scanner(ext);
+
+            while(scanner.hasNext())
+                extensions.add(scanner.next());
+
+            extensions = Collections.unmodifiableSet(extensions);
+        }
+
+        return extensions;
+    }
+
+    /**
      * Returns a info string in exchange for a key (CL_PLATFORM_*).
      */
     public String getInfoString(int key) {
diff --git a/src/com/mbien/opencl/CLUtils.java b/src/com/mbien/opencl/CLUtils.java
index eb66a7a3..03fb559c 100644
--- a/src/com/mbien/opencl/CLUtils.java
+++ b/src/com/mbien/opencl/CLUtils.java
@@ -32,6 +32,7 @@ class CLUtils {
         map.put("CL_PLATFORM_PROFILE",  platform.getProfile());
         map.put("CL_PLATFORM_VERSION",  platform.getVersion());
         map.put("CL_PLATFORM_VENDOR",   platform.getVendor());
+        map.put("CL_PLATFORM_EXTENSIONS",   platform.getExtensions().toString());
 //        map.put("fastest device (estimated)", platform.getMaxFlopsDevice().toString());
 
         return map;
@@ -53,7 +54,8 @@ class CLUtils {
         map.put("CL_DEVICE_MAX_PARAMETER_SIZE", dev.getMaxParameterSize()+" Byte");
         map.put("CL_DEVICE_LOCAL_MEM_SIZE",     dev.getLocalMemSize()/1024+" KB");
         map.put("CL_DEVICE_LOCAL_MEM_TYPE",     dev.getLocalMemType()+"");
-        map.put("CL_DEVICE_GLOBAL_MEM_CACHE_SIZE", dev.getGlobalMemCachSize()+"");
+        map.put("CL_DEVICE_GLOBAL_MEM_CACHE_SIZE", dev.getGlobalMemCacheSize()+"");
+        map.put("CL_DEVICE_GLOBAL_MEM_CACHELINE_SIZE", dev.getGlobalMemCachelineSize()+"");
         map.put("CL_DEVICE_GLOBAL_MEM_CACHE_TYPE", dev.getGlobalMemCacheType()+"");
         map.put("CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE", dev.getMaxConstantBufferSize()+"");
         map.put("CL_DEVICE_MAX_CONSTANT_ARGS",        dev.getMaxConstantArgs()+"");
diff --git a/src/com/mbien/opencl/NativeLibLoader.java b/src/com/mbien/opencl/NativeLibLoader.java
new file mode 100644
index 00000000..b8148432
--- /dev/null
+++ b/src/com/mbien/opencl/NativeLibLoader.java
@@ -0,0 +1,22 @@
+package com.mbien.opencl;
+
+import java.security.AccessController;
+import java.security.PrivilegedAction;
+import com.sun.nativewindow.impl.NativeLibLoaderBase;
+
+/**
+ *
+ * @author Michael Bien
+ */
+class NativeLibLoader extends NativeLibLoaderBase {
+
+    @SuppressWarnings("unchecked")
+    public static void loadJOCL() {
+        AccessController.doPrivileged(new PrivilegedAction() {
+            public Object run() {
+                loadLibrary("jocl", null, true);
+                return null;
+            }
+        });
+    }
+}
diff --git a/test/com/mbien/opencl/CLConcurrencyTest.java b/test/com/mbien/opencl/CLConcurrencyTest.java
index 94e8c4ee..e7a244c3 100644
--- a/test/com/mbien/opencl/CLConcurrencyTest.java
+++ b/test/com/mbien/opencl/CLConcurrencyTest.java
@@ -19,7 +19,7 @@ public class CLConcurrencyTest {
      @Test
      public void testEvents() throws IOException {
 
-        out.println(" - - - event synchronisation test - - - ");
+        out.println(" - - - event synchronization test - - - ");
 
         final int groupSize = 256;
         final int elements = roundUp(groupSize, ONE_MB/SIZEOF_INT * 5); // 5MB per buffer
@@ -82,7 +82,7 @@ public class CLConcurrencyTest {
      @Test
      public void concurrencyTest() throws IOException, InterruptedException {
 
-        out.println(" - - - queue synchronisation test - - - ");
+        out.println(" - - - QueueBarrier test - - - ");
 
         final int elements = ONE_MB/SIZEOF_INT * 10; // 20MB per buffer
 
diff --git a/test/com/mbien/opencl/HighLevelBindingTest.java b/test/com/mbien/opencl/HighLevelBindingTest.java
index 6ae9ac7e..9c0e2343 100644
--- a/test/com/mbien/opencl/HighLevelBindingTest.java
+++ b/test/com/mbien/opencl/HighLevelBindingTest.java
@@ -58,6 +58,7 @@ public class HighLevelBindingTest {
             out.println("    version: "+platform.getVersion());
             out.println("    vendor: "+platform.getVendor());
             out.println("    max FLOPS device: "+platform.getMaxFlopsDevice());
+            out.println("    extensions: "+platform.getExtensions());
 
             CLDevice[] clDevices = platform.listCLDevices();
             for (CLDevice device : clDevices) {
@@ -74,7 +75,8 @@ public class HighLevelBindingTest {
                 out.println("    max param size: "+device.getMaxParameterSize()+" byte");
                 out.println("    local mem: "+device.getLocalMemSize()/1024+" KB");
                 out.println("    local mem type: "+device.getLocalMemType());
-                out.println("    global mem cache size: "+device.getGlobalMemCachSize());
+                out.println("    global mem cache size: "+device.getGlobalMemCacheSize());
+                out.println("    global mem cacheline size: "+device.getGlobalMemCachelineSize());
                 out.println("    global mem cache type: "+device.getGlobalMemCacheType());
                 out.println("    constant buffer size: "+device.getMaxConstantBufferSize());
                 out.println("    error correction support: "+device.isErrorCorrectionSupported());