/*
 * JavaCL - Java API and utilities for OpenCL
 * http://javacl.googlecode.com/
 *
 * Copyright (c) 2009-2013, Olivier Chafik (http://ochafik.com/)
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are met:
 * 
 *     * Redistributions of source code must retain the above copyright
 *       notice, this list of conditions and the following disclaimer.
 *     * Redistributions in binary form must reproduce the above copyright
 *       notice, this list of conditions and the following disclaimer in the
 *       documentation and/or other materials provided with the distribution.
 *     * Neither the name of Olivier Chafik nor the
 *       names of its contributors may be used to endorse or promote products
 *       derived from this software without specific prior written permission.
 * 
 * THIS SOFTWARE IS PROVIDED BY OLIVIER CHAFIK AND CONTRIBUTORS ``AS IS'' AND ANY
 * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED. IN NO EVENT SHALL THE REGENTS AND CONTRIBUTORS BE LIABLE FOR ANY
 * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
 * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */
        




        

        
        



        

        
        
        

        
package com.nativelibs4java.opencl;
import com.nativelibs4java.util.Pair;

import static com.nativelibs4java.opencl.CLException.error;
import static com.nativelibs4java.opencl.CLException.failedForLackOfMemory;
import static com.nativelibs4java.opencl.JavaCL.CL;
import static com.nativelibs4java.opencl.library.OpenCLLibrary.*;
import static com.nativelibs4java.opencl.library.IOpenCLLibrary.*;
import com.nativelibs4java.opencl.library.IOpenCLLibrary.cl_event;
import static com.nativelibs4java.util.ImageUtils.getImageIntPixels;
import static com.nativelibs4java.util.NIOUtils.getSizeInBytes;

import java.awt.Image;
import java.io.IOException;
import java.nio.Buffer;
import java.nio.ByteBuffer;
import java.nio.ByteOrder;
import java.nio.CharBuffer;
import java.nio.DoubleBuffer;
import java.nio.FloatBuffer;
import java.nio.IntBuffer;
import java.nio.LongBuffer;
import java.nio.ShortBuffer;
import java.util.ArrayList;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.atomic.AtomicReference;
import java.util.List;
import java.util.logging.*;

import com.nativelibs4java.opencl.CLDevice.QueueProperties;
import com.nativelibs4java.opencl.CLSampler.AddressingMode;
import com.nativelibs4java.opencl.CLSampler.FilterMode;
import com.nativelibs4java.opencl.ImageIOUtils.ImageInfo;
import com.nativelibs4java.opencl.library.OpenGLContextUtils;
import com.nativelibs4java.opencl.library.cl_image_format;
import com.nativelibs4java.opencl.library.IOpenCLLibrary.cl_context;
import com.nativelibs4java.opencl.library.IOpenCLLibrary.cl_device_id;
import com.nativelibs4java.opencl.library.IOpenCLLibrary.cl_mem;
import com.nativelibs4java.opencl.library.IOpenCLLibrary.cl_sampler;
import com.nativelibs4java.util.EnumValue;
import com.nativelibs4java.util.EnumValues;
import com.nativelibs4java.util.NIOUtils;
import org.bridj.*;
import static org.bridj.Pointer.*;
import static com.nativelibs4java.opencl.proxy.PointerUtils.*;
import java.io.InputStream;
import java.util.Arrays;
import java.util.Map;


/**
 * OpenCL context.<br>
 * An OpenCL context is created with one or more devices.<br>
 * Contexts are used by the OpenCL runtime for managing objects such as command-queues, memory, program and kernel objects and for executing kernels on one or more devices specified in the context.
 * @author Olivier Chafik
 */
public class CLContext extends CLAbstractEntity {








        private final AtomicReference<ConcurrentHashMap<Object, Object>> propertiesMapRef =
                new AtomicReference<ConcurrentHashMap<Object, Object>>();
        
        public Object getClientProperty(Object key) {
                ConcurrentHashMap<Object, Object> propertiesMap = propertiesMapRef.get();
                return propertiesMap == null ? null : propertiesMap.get(key);
        }
        public Object putClientProperty(Object key, Object value) {
                ConcurrentHashMap<Object, Object> propertiesMap = propertiesMapRef.get();
                if (propertiesMap == null) {
                        propertiesMap = new ConcurrentHashMap<Object, Object>();
                        if (!propertiesMapRef.compareAndSet(null, propertiesMap))
                                propertiesMap = propertiesMapRef.get();
                }
                return propertiesMap.put(key, value);
        }
        
        private volatile long maxMemAllocSize = -1;
        
        /**
     * Max size of memory object allocation in bytes. The minimum value is max (1/4th of CL_DEVICE_GLOBAL_MEM_SIZE , 128*1024*1024)
     */
    public long getMaxMemAllocSize() {
        if (maxMemAllocSize < 0) {
                        long min = Long.MAX_VALUE;
                        for (CLDevice device : getDevices()) {
                                long m = device.getMaxMemAllocSize();
                                if (m < min)
                                        min = m;
                        }
                        maxMemAllocSize = min;
                }
        return maxMemAllocSize;
    }
    
    volatile Boolean cacheBinaries;
        
        /**
         * Change whether program binaries are automatically cached or not.<br>
         * By default it is true, it can be set to false with the "javacl.cacheBinaries" Java property or the "JAVACL_CACHE_BINARIES" environment variable (when set to "0").<br>
         * Each program can be set to be cached or not using {@link CLProgram#setCached(boolean) }.<br>
         * Caching of binaries might be disabled by default on some platforms (ATI Stream, for instance).
         */ 
        public synchronized void setCacheBinaries(boolean cacheBinaries) {
                this.cacheBinaries = cacheBinaries;
        }
        /**
         * Says whether program binaries are automatically cached or not.<br>
         * By default it is true, it can be set to false with the "javacl.cacheBinaries" Java property, the "JAVACL_CACHE_BINARIES" environment variable (when set to "0") or the {@link CLContext#setCacheBinaries(boolean) } method.<br>
         * Each program can be set to be cached or not using {@link CLProgram#setCached(boolean) }.<br>
         * Caching of binaries might be disabled by default on some platforms (ATI Stream, for instance).
         */ 
        public synchronized boolean getCacheBinaries() {
                if (cacheBinaries == null) {
                        String prop = System.getProperty("javacl.cacheBinaries"), env = System.getenv("JAVACL_CACHE_BINARIES");
                        if ("true".equals(prop) || "1".equals(env))
                                cacheBinaries = true;
                        else if ("false".equals(prop) || "0".equals(env))
                                cacheBinaries = false;
                        else {
                                cacheBinaries = !PlatformUtils.PlatformKind.AMDApp.equals(PlatformUtils.guessPlatformKind(getPlatform()));
                        }
                        //System.out.println("CACHE BINARIES = " + cacheBinaries);
                }
                return cacheBinaries;
        }
        
                protected static CLInfoGetter infos = new CLInfoGetter() {
                @Override
                protected int getInfo(long entity, int infoTypeEnum, long size, Pointer out, Pointer<SizeT> sizeOut) {
                        return CL.clGetContextInfo(entity, infoTypeEnum, size, getPeer(out), getPeer(sizeOut));
                }
        };
        
        CLPlatform platform;
        protected Pointer<SizeT> deviceIds;

        CLContext(CLPlatform platform, Pointer<SizeT> deviceIds, long context) {
                super(context);
                this.platform = platform;
                this.deviceIds = deviceIds;
                
                if (getByteOrder() == null) {
                        JavaCL.log(Level.WARNING, "The devices in this context have mismatching byte orders. This mandates the use of __attribute__((endian(host))) in kernel sources or *very* careful use of buffers to avoid facing endianness issues");   
                }
        }
    
        /**
         * Creates a user event object. <br>
         * User events allow applications to enqueue commands that wait on a user event to finish before the command is executed by the device.
         * @since OpenCL 1.1
         */
        public CLUserEvent createUserEvent() {
                platform.requireMinVersionValue("clCreateUserEvent", 1.1);
                                ReusablePointers ptrs = ReusablePointers.get();
                Pointer<Integer> pErr = ptrs.pErr;
                long evt = CL.clCreateUserEvent(getEntity(), getPeer(pErr));
                        error(pErr.getInt());
                return (CLUserEvent)CLEvent.createEvent(null, evt, true);
        }

        /**
         * Create an OpenCL queue on the first device of this context.<br>
         * Equivalent to calling <code>getDevices()[0].createQueue(context)</code>
         * @return new OpenCL queue
         */
        public CLQueue createDefaultQueue(QueueProperties... queueProperties) {
                return new CLDevice(platform, deviceIds.getSizeT()).createQueue(this, queueProperties);
        }

        /**
         * Create an out-of-order OpenCL queue on the first device of this context.<br>
         * Equivalent to calling <code>getDevices()[0].createOutOfOrderQueue(context)</code>
         * @return new out-of-order OpenCL queue
         */
        public CLQueue createDefaultOutOfOrderQueue() {
                return new CLDevice(platform, deviceIds.getSizeT()).createOutOfOrderQueue(this);
        }

        
    public String toString() {
        StringBuilder b = new StringBuilder("CLContext(platform = ").append(getPlatform().getName()).append("; devices = ");
        boolean first = true;
        for (CLDevice d : getDevices()) {
            if (first)
                first = false;
            else
                b.append(", ");
            b.append(d.getName());
        }
        b.append(")");
        return b.toString();
    }
        public CLQueue createDefaultOutOfOrderQueueIfPossible() {
                try {
                        return createDefaultOutOfOrderQueue();
                } catch (Throwable th) {//CLException.InvalidQueueProperties ex) {
                        return createDefaultQueue();
                }
    }

        /**
         * Create an profiling-enabled OpenCL queue on the first device of this context.<br>
         * Equivalent to calling <code>getDevices()[0].createProfilingQueue(context)</code>
         * @return new profiling-enabled OpenCL queue
         */
        public CLQueue createDefaultProfilingQueue() {
                return new CLDevice(platform, deviceIds.getSizeT()).createProfilingQueue(this);
        }

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clGetSupportedImageFormats.html">clGetSupportedImageFormats</a>.<br>
        */
        @SuppressWarnings("deprecation")
        public CLImageFormat[] getSupportedImageFormats(CLBuffer.Flags flags, CLBuffer.ObjectType imageType) {
                Pointer<Integer> pCount = allocateInt();
                int memFlags = (int) flags.value();
                int imTyp = (int) imageType.value();
                CL.clGetSupportedImageFormats(getEntity(), memFlags, imTyp, 0, 0, getPeer(pCount));
                //cl_image_format ft = new cl_image_format();
                //int sz = ft.size();
                int n = pCount.getInt();
                if (n == 0) {
                        n = 30; // There HAS to be at least one format. the spec even says even more, but in fact on Mac OS X / CPU there's only one...
                }
        Pointer<cl_image_format> formats = allocateArray(cl_image_format.class, n);
                CL.clGetSupportedImageFormats(getEntity(), memFlags, imTyp, n, getPeer(formats), 0);
                List<CLImageFormat> ret = new ArrayList<CLImageFormat>(n);
        for (cl_image_format ft : formats) {
            if (ft.image_channel_data_type() == 0 && ft.image_channel_order() == 0)
                                break;

                        ret.add(new CLImageFormat(ft));
                }
                return ret.toArray(new CLImageFormat[ret.size()]);
        }

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateSampler.html">clCreateSampler</a>.<br>
        */
        @SuppressWarnings("deprecation")
        public CLSampler createSampler(boolean normalized_coords, AddressingMode addressing_mode, FilterMode filter_mode) {
                                ReusablePointers ptrs = ReusablePointers.get();
                Pointer<Integer> pErr = ptrs.pErr;
                long sampler = CL.clCreateSampler(
                        getEntity(), 
                        normalized_coords ? CL_TRUE : CL_FALSE, 
                        (int) addressing_mode.value(), 
                        (int) filter_mode.value(), 
                        getPeer(pErr)
                );
                        error(pErr.getInt());
                return new CLSampler(sampler);
        }

        public int getDeviceCount() {
                return infos.getOptionalFeatureInt(getEntity(), CL.CL_CONTEXT_NUM_DEVICES);
        }
        
        /**
         * Lists the devices of this context
         * @return array of the devices that form this context
         */
        public synchronized CLDevice[] getDevices() {
                if (deviceIds == null) {
                        deviceIds = infos.getMemory(getEntity(), CL_CONTEXT_DEVICES).as(SizeT.class);
                }
        int n = (int)deviceIds.getValidElements();

                CLDevice[] devices = new CLDevice[n];
                for (int i = n; i-- != 0;) {
                        devices[i] = new CLDevice(platform, deviceIds.getSizeTAtIndex(i));
                }
                return devices;
        }

        /**
         * Create a program with all the C source code content provided as argument.
         * @param srcs list of the content of source code for the program
         * @return a program that needs to be built
         */
        public CLProgram createProgram(String... srcs) {
        return createProgram(null, srcs);
        }

    public CLProgram createProgram(CLDevice[] devices, String... srcs) {
        CLProgram program = new CLProgram(this, devices);
        for (String src : srcs)
            program.addSource(src);
        return program;
    }

    /**
     * Restore a program previously saved with {@link CLProgram#store(java.io.OutputStream) }
     * @param in will be closed
     * @return a CLProgram object representing the previously saved program
     * @throws IOException
     */
    public CLProgram loadProgram(InputStream in) throws IOException {
        Pair<Map<CLDevice, byte[]>, String> binaries = CLProgram.readBinaries(Arrays.asList(getDevices()), null, in);
        return createProgram(binaries.getFirst(), binaries.getSecond());
    }

        public CLProgram createProgram(Map<CLDevice, byte[]> binaries, String source) {
                return new CLProgram(this, binaries, source);
        }

        //cl_queue queue;
        @Override
        protected void clear() {
                error(CL.clReleaseContext(getEntity()));
        }



    @Deprecated
    public CLDevice guessCurrentGLDevice() {
        long[] props = platform.getContextProps(CLPlatform.getGLContextProperties(getPlatform()));
        Pointer<SizeT> propsRef = pointerToSizeTs(props);
        
        Pointer<SizeT> pCount = allocateSizeT();
        Pointer<Pointer<?>> mem = allocatePointer();
        if (Platform.isMacOSX())
            error(CL.clGetGLContextInfoAPPLE(
                getEntity(), 
                getPeer(OpenGLContextUtils.CGLGetCurrentContext()),
                CL_CURRENT_DEVICE_FOR_GL_CONTEXT_KHR,
                Pointer.SIZE, 
                getPeer(mem), 
                getPeer(pCount)));
        else
            error(CL.clGetGLContextInfoKHR(
                getPeer(propsRef), 
                CL_CURRENT_DEVICE_FOR_GL_CONTEXT_KHR, 
                Pointer.SIZE, 
                getPeer(mem), 
                getPeer(pCount)
                        ));

        if (pCount.getSizeT() != Pointer.SIZE)
            throw new RuntimeException("Not a device : len = " + pCount.get().intValue());

        Pointer p = mem.getPointer();
        if (p.equals(Pointer.NULL))
            return null;
        return new CLDevice(null, getPeer(p));
    }

    private static <T extends CLMem> T markAsGL(T mem) {
        mem.isGL = true;
        return mem;
    }

    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateFromGLBuffer.html">clCreateFromGLBuffer</a>.<br>
     * Makes an OpenGL Vertex Buffer Object (VBO) visible to OpenCL as a buffer object.<br>
     * Note that memory objects shared with OpenGL must be acquired / released before / after use from OpenCL.
     * see {@link CLMem#acquireGLObject(com.nativelibs4java.opencl.CLQueue, com.nativelibs4java.opencl.CLEvent[]) }
     * see {@link CLMem#releaseGLObject(com.nativelibs4java.opencl.CLQueue, com.nativelibs4java.opencl.CLEvent[]) } 
         * @param usage flags
     * @param openGLBufferObject Identifier of a VBO, as generated by glGenBuffers
     */
        @SuppressWarnings("deprecation")
        public CLBuffer<Byte> createBufferFromGLBuffer(CLMem.Usage usage, int openGLBufferObject) {
                                ReusablePointers ptrs = ReusablePointers.get();
                Pointer<Integer> pErr = ptrs.pErr;
                long mem;
                int previousAttempts = 0;
                do {
                        mem = CL.clCreateFromGLBuffer(
                                getEntity(), 
                                usage.getIntFlags(), 
                                openGLBufferObject, 
                                getPeer(pErr)
                        );
                } while (failedForLackOfMemory(pErr.getInt(), previousAttempts++));
        return markAsGL(new CLBuffer(this, -1, mem, null, PointerIO.getByteInstance()));
        }

    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateFromGLRenderbuffer.html">clCreateFromGLRenderbuffer</a>.<br>
     * Makes an OpenGL Render Buffer visible to OpenCL as a 2D image.<br>
     * Note that memory objects shared with OpenGL must be acquired / released before / after use from OpenCL.
     * see {@link CLMem#acquireGLObject(com.nativelibs4java.opencl.CLQueue, com.nativelibs4java.opencl.CLEvent[]) }
     * see {@link CLMem#releaseGLObject(com.nativelibs4java.opencl.CLQueue, com.nativelibs4java.opencl.CLEvent[]) }
     * @param openGLRenderBuffer Identifier of an OpenGL render buffer
     */
        @SuppressWarnings("deprecation")
        public CLImage2D createImage2DFromGLRenderBuffer(CLMem.Usage usage, int openGLRenderBuffer) {
                                ReusablePointers ptrs = ReusablePointers.get();
                Pointer<Integer> pErr = ptrs.pErr;
                long mem;
                int previousAttempts = 0;
                do {
                        mem = CL.clCreateFromGLRenderbuffer(
                                getEntity(), 
                                usage.getIntFlags(), 
                                openGLRenderBuffer, 
                                getPeer(pErr)
                        );
                } while (failedForLackOfMemory(pErr.getInt(), previousAttempts++));
                return markAsGL(new CLImage2D(this, mem, null));
        }
        
        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateFromGLTexture2D.html">clCreateFromGLTexture2D</a>.<br>
         * Creates an OpenCL 2D image object from an OpenGL 2D texture object, or a single face of an OpenGL cubemap texture object.<br>
         * Note that memory objects shared with OpenGL must be acquired / released before / after use from OpenCL.
     * @param usage
         * @param textureTarget Must be one of GL_TEXTURE_2D, GL_TEXTURE_CUBE_MAP_POSITIVE_X, GL_TEXTURE_CUBE_MAP_POSITIVE_Y, GL_TEXTURE_CUBE_MAP_POSITIVE_Z, GL_TEXTURE_CUBE_MAP_NEGATIVE_X, GL_TEXTURE_CUBE_MAP_NEGATIVE_Y, GL_TEXTURE_CUBE_MAP_NEGATIVE_Z, or GL_TEXTURE_RECTANGLE47. texture_target is used only to define the image type of texture. No reference to a bound GL texture object is made or implied by this parameter.
         * @param mipLevel Mipmap level to be used (Implementations may return CL_INVALID_OPERATION for miplevel values &gt; 0)
         * @param texture Name of a GL 2D, cubemap or rectangle texture object. The texture object must be a complete texture as per OpenGL rules on texture completeness. The texture format and dimensions defined by OpenGL for the specified miplevel of the texture will be used to create the 2D image object. Only GL texture objects with an internal format that maps to appropriate image channel order and data type specified in tables 5.4 and 5.5 may be used to create a 2D image object.
         * @return valid OpenCL image object if the image object is created successfully
         * @throws CLException.InvalidMipLevel if miplevel is less than the value of levelbase (for OpenGL implementations) or zero (for OpenGL ES implementations); or greater than the value of q (for both OpenGL and OpenGL ES). levelbase and q are defined for the texture in section 3.8.10 (Texture Completeness) of the OpenGL 2.1 specification and section 3.7.10 of the OpenGL ES 2.0, or if miplevel is greather than zero and the OpenGL implementation does not support creating from non-zero mipmap levels.
         * @throws CLException.InvalidGLObject if texture is not a GL texture object whose type matches texture_target, if the specified miplevel of texture is not defined, or if the width or height of the specified miplevel is zero.
         * @throws CLException.InvalidImageFormatDescriptor if the OpenGL texture internal format does not map to a supported OpenCL image format.
         */
        @SuppressWarnings("deprecation")
        public CLImage2D createImage2DFromGLTexture2D(CLMem.Usage usage, GLTextureTarget textureTarget, int texture, int mipLevel) {
                platform.requireMinVersionValue("clCreateFromGLTexture2D", 1.1, 1.2);
                                ReusablePointers ptrs = ReusablePointers.get();
                Pointer<Integer> pErr = ptrs.pErr;
                long mem;
                int previousAttempts = 0;
                do {
                        mem = CL.clCreateFromGLTexture2D(
                                getEntity(), 
                                usage.getIntFlags(), 
                                (int)textureTarget.value(), 
                                mipLevel, 
                                texture, 
                                getPeer(pErr)
                        );
                } while (failedForLackOfMemory(pErr.getInt(), previousAttempts++));
                return markAsGL(new CLImage2D(this, mem, null));
        }

        private static final int 
                GL_TEXTURE_2D = 0x0DE1,
                GL_TEXTURE_3D = 0x806F,
                GL_TEXTURE_CUBE_MAP_POSITIVE_X = 0x8515,
                GL_TEXTURE_CUBE_MAP_NEGATIVE_X = 0x8516,
                GL_TEXTURE_CUBE_MAP_POSITIVE_Y = 0x8517,
                GL_TEXTURE_CUBE_MAP_NEGATIVE_Y = 0x8518,
                GL_TEXTURE_CUBE_MAP_POSITIVE_Z = 0x8519,
                GL_TEXTURE_CUBE_MAP_NEGATIVE_Z = 0x851A,
                GL_TEXTURE_RECTANGLE = 0x84F5;

        public CLPlatform getPlatform() {
                return platform;
        }
                

        public enum GLTextureTarget implements com.nativelibs4java.util.ValuedEnum {
                
                Texture2D(GL_TEXTURE_2D),
                //Texture3D(GL_TEXTURE_3D),
                CubeMapPositiveX(GL_TEXTURE_CUBE_MAP_POSITIVE_X),
                CubeMapNegativeX(GL_TEXTURE_CUBE_MAP_NEGATIVE_X),
                CubeMapPositiveY(GL_TEXTURE_CUBE_MAP_POSITIVE_Y),
                CubeMapNegativeY(GL_TEXTURE_CUBE_MAP_NEGATIVE_Y),
                CubeMapPositiveZ(GL_TEXTURE_CUBE_MAP_POSITIVE_Z),
                CubeMapNegativeZ(GL_TEXTURE_CUBE_MAP_NEGATIVE_Z),
                Rectangle(GL_TEXTURE_RECTANGLE);

                GLTextureTarget(long value) { this.value = value; }
                long value;
                @Override
                public long value() { return value; }
                public static GLTextureTarget getEnum(int v) { return EnumValues.getEnum(v, GLTextureTarget.class); }
        }
        
        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateFromGLTexture3D.html">clCreateFromGLTexture3D</a>.<br>
         * Creates an OpenCL 3D image object from an OpenGL 3D texture object<br>
         * Note that memory objects shared with OpenGL must be acquired / released before / after use from OpenCL.
         * @param usage
         * @param mipLevel Mipmap level to be used (Implementations may return CL_INVALID_OPERATION for miplevel values &gt; 0)
         * @param texture Name of a GL 3D texture object. The texture object must be a complete texture as per OpenGL rules on texture completeness. The texture format and dimensions defined by OpenGL for the specified miplevel of the texture will be used to create the 3D image object. Only GL texture objects with an internal format that maps to appropriate image channel order and data type specified in tables 5.4 and 5.5 can be used to create the 3D image object.
         * @return valid OpenCL image object if the image object is created successfully
         * @throws CLException.InvalidMipLevel if miplevel is less than the value of levelbase (for OpenGL implementations) or zero (for OpenGL ES implementations); or greater than the value of q (for both OpenGL and OpenGL ES). levelbase and q are defined for the texture in section 3.8.10 (Texture Completeness) of the OpenGL 2.1 specification and section 3.7.10 of the OpenGL ES 2.0, or if miplevel is greather than zero and the OpenGL implementation does not support creating from non-zero mipmap levels.
         * @throws CLException.InvalidGLObject if texture is not a GL texture object whose type matches texture_target, if the specified miplevel of texture is not defined, or if the width or height of the specified miplevel is zero.
         * @throws CLException.InvalidImageFormatDescriptor if the OpenGL texture internal format does not map to a supported OpenCL image format.
         */
        @SuppressWarnings("deprecation")
        public CLImage3D createImage3DFromGLTexture3D(CLMem.Usage usage, int texture, int mipLevel) {
                platform.requireMinVersionValue("clCreateFromGLTexture3D", 1.1, 1.2);
                                ReusablePointers ptrs = ReusablePointers.get();
                Pointer<Integer> pErr = ptrs.pErr;
                long mem;
                int previousAttempts = 0;
                do {
                        mem = CL.clCreateFromGLTexture3D(
                                getEntity(), 
                                usage.getIntFlags(), 
                                GL_TEXTURE_3D, 
                                mipLevel, 
                                texture, 
                                getPeer(pErr)
                        );
                } while (failedForLackOfMemory(pErr.getInt(), previousAttempts++));
                return markAsGL(new CLImage3D(this, mem, null));
        }
        
        /**
         * Create an ARGB input 2D image with the content provided
         * @param allowUnoptimizingDirectRead Some images expose their internal data for direct read, leading to performance increase during the creation of the OpenCL image. However, direct access to the image data disables some Java2D optimizations for this image, leading to degraded performance in subsequent uses with AWT/Swing.
         */
        public CLImage2D createImage2D(CLMem.Usage usage, Image image, boolean allowUnoptimizingDirectRead) {
                int width = image.getWidth(null), height = image.getHeight(null);
                ImageInfo info = ImageIOUtils.getImageInfo(image);
        return createImage2D(
                                usage,
                                info.clImageFormat,
                                width, height,
                                0,
                                info.dataGetter.getData(image, null, true, allowUnoptimizingDirectRead, getByteOrder()),
                                true);
        }

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateImage2D.html">clCreateImage2D</a>.<br>
     * @param usage Usage intended for the image.
     * @param format Format of the image.
     * @param width Image width
     * @param height Image height
     * @return A new image allocated on this context.
   * @param rowPitch Row pitch.
   * @param buffer Image data buffer.
   * @param copy Whether to copy the input buffer.
        */
        @SuppressWarnings("deprecation")
        public CLImage2D createImage2D(CLMem.Usage usage, CLImageFormat format, long width, long height, long rowPitch, Buffer buffer, boolean copy) {
                platform.requireMinVersionValue("clCreateImage2D", 1.1, 1.2);
                
                long memFlags = usage.getIntFlags();
                if (buffer != null) {
                        memFlags |= copy ? CL_MEM_COPY_HOST_PTR : CL_MEM_USE_HOST_PTR;
                }

                                ReusablePointers ptrs = ReusablePointers.get();
                Pointer<Integer> pErr = ptrs.pErr;
                Pointer<cl_image_format> pImageFormat = getPointer(format.to_cl_image_format());
                Pointer<?> pBuffer = buffer == null ? null : pointerToBuffer(buffer);
                long mem;
                int previousAttempts = 0;
                do {
                        mem = CL.clCreateImage2D(
                                getEntity(),
                                memFlags,
                                getPeer(pImageFormat),
                                width,
                                height,
                                rowPitch,
                                getPeer(pBuffer),
                                getPeer(pErr));
                } while (failedForLackOfMemory(pErr.getInt(), previousAttempts++));
                return new CLImage2D(this, mem, format);
        }

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateImage2D.html">clCreateImage2D</a>.<br>
     * @param usage Usage intended for the image.
     * @param format Format of the image.
     * @param width Image width
     * @param height Image height
     * @return A new image allocated on this context.
   * @param rowPitch Row pitch.
  */
  public CLImage2D createImage2D(CLMem.Usage usage, CLImageFormat format, long width, long height, long rowPitch) {
                return createImage2D(usage, format, width, height, rowPitch, null, false);
        }

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateImage2D.html">clCreateImage2D</a>.<br>
     * @param usage Usage intended for the image.
     * @param format Format of the image.
     * @param width Image width
     * @param height Image height
     * @return A new image allocated on this context.
  */
  public CLImage2D createImage2D(CLMem.Usage usage, CLImageFormat format, long width, long height) {
                return createImage2D(usage, format, width, height, 0, null, false);
        }

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateImage3D.html">clCreateImage3D</a>.<br>
     * @param usage Usage intended for the image.
     * @param format Format of the image.
     * @param width Image width
     * @param height Image height
     * @return A new image allocated on this context.
     * @param depth Image depth
   * @param rowPitch Row pitch.
   * @param slicePitch Slice pitch.
   * @param buffer Image data buffer.
   * @param copy Whether to copy the input buffer.
  */
  @SuppressWarnings("deprecation")
        public CLImage3D createImage3D(CLMem.Usage usage, CLImageFormat format, long width, long height, long depth, long rowPitch, long slicePitch, Buffer buffer, boolean copy) {
                platform.requireMinVersionValue("clCreateImage3D", 1.1, 1.2);
                long memFlags = usage.getIntFlags();
                if (buffer != null) {
                        memFlags |= copy ? CL_MEM_COPY_HOST_PTR : CL_MEM_USE_HOST_PTR;
                }

                                ReusablePointers ptrs = ReusablePointers.get();
                Pointer<Integer> pErr = ptrs.pErr;
                Pointer<cl_image_format> pImageFormat = getPointer(format.to_cl_image_format());
                Pointer<?> pBuffer = buffer == null ? null : pointerToBuffer(buffer);
                long mem;
                int previousAttempts = 0;
                do {
                        mem = CL.clCreateImage3D(
                                getEntity(),
                                memFlags,
                                getPeer(pImageFormat),
                                width,
                                height,
                                depth,
                                rowPitch,
                                slicePitch,
                                getPeer(pBuffer),
                                getPeer(pErr));
                } while (failedForLackOfMemory(pErr.getInt(), previousAttempts++));
                return new CLImage3D(this, mem, format);
        }

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateImage3D.html">clCreateImage3D</a>.<br>
     * @param usage Usage intended for the image.
     * @param format Format of the image.
     * @param width Image width
     * @param height Image height
     * @return A new image allocated on this context.
     * @param depth Image depth
   * @param rowPitch Row pitch.
   * @param slicePitch Slice pitch.
  */
  public CLImage3D createImage3D(CLMem.Usage usage, CLImageFormat format, long width, long height, long depth, long rowPitch, long slicePitch) {
                return createImage3D(usage, format, width, height, depth, rowPitch, slicePitch, null, false);
        }

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateImage3D.html">clCreateImage3D</a>.<br>
     * @param usage Usage intended for the image.
     * @param format Format of the image.
     * @param width Image width
     * @param height Image height
     * @return A new image allocated on this context.
     * @param depth Image depth
  */
  public CLImage3D createImage3D(CLMem.Usage usage, CLImageFormat format, long width, long height, long depth) {
                return createImage3D(usage, format, width, height, depth, 0, 0, null, false);
        }


        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Integer&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Integer> createIntBuffer(CLMem.Usage usage,  IntBuffer  data, boolean copy) {
                return createBuffer(usage, Pointer.pointerToInts(data), copy);
        }

    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Integer&gt;</code> OpenCL buffer big enough to hold the valid  values of the provided data pointer.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
     * @param data Host data to copy to the buffer
     * @return A new buffer allocated on this context.
   */
        public CLBuffer<Integer> createIntBuffer(CLMem.Usage usage, Pointer<Integer> data) {
                return createIntBuffer(usage, data, true);
        }
        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Integer&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Integer> createIntBuffer(CLMem.Usage usage, Pointer<Integer> data, boolean copy) {
                return createBuffer(usage, data, copy);
        }
        
    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Integer&gt;</code> OpenCL buffer big enough to hold 'length' values of type int.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
    
   * @param elementCount Length of the buffer expressed in elements (for instance, a <code>CLBuffer&lt;Integer&gt;</code> of length 10 will actually contain 10 * 4 bytes, as ints are 4-bytes-long)
   * @return A new buffer allocated on this context.
   */
        public CLBuffer<Integer> createIntBuffer(CLMem.Usage usage, long elementCount) {
                return createBuffer(usage, Integer.class, elementCount);
        }
        

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Long&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Long> createLongBuffer(CLMem.Usage usage,  LongBuffer  data, boolean copy) {
                return createBuffer(usage, Pointer.pointerToLongs(data), copy);
        }

    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Long&gt;</code> OpenCL buffer big enough to hold the valid  values of the provided data pointer.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
     * @param data Host data to copy to the buffer
     * @return A new buffer allocated on this context.
   */
        public CLBuffer<Long> createLongBuffer(CLMem.Usage usage, Pointer<Long> data) {
                return createLongBuffer(usage, data, true);
        }
        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Long&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Long> createLongBuffer(CLMem.Usage usage, Pointer<Long> data, boolean copy) {
                return createBuffer(usage, data, copy);
        }
        
    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Long&gt;</code> OpenCL buffer big enough to hold 'length' values of type long.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
    
   * @param elementCount Length of the buffer expressed in elements (for instance, a <code>CLBuffer&lt;Long&gt;</code> of length 10 will actually contain 10 * 8 bytes, as longs are 8-bytes-long)
   * @return A new buffer allocated on this context.
   */
        public CLBuffer<Long> createLongBuffer(CLMem.Usage usage, long elementCount) {
                return createBuffer(usage, Long.class, elementCount);
        }
        

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Short&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Short> createShortBuffer(CLMem.Usage usage,  ShortBuffer  data, boolean copy) {
                return createBuffer(usage, Pointer.pointerToShorts(data), copy);
        }

    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Short&gt;</code> OpenCL buffer big enough to hold the valid  values of the provided data pointer.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
     * @param data Host data to copy to the buffer
     * @return A new buffer allocated on this context.
   */
        public CLBuffer<Short> createShortBuffer(CLMem.Usage usage, Pointer<Short> data) {
                return createShortBuffer(usage, data, true);
        }
        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Short&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Short> createShortBuffer(CLMem.Usage usage, Pointer<Short> data, boolean copy) {
                return createBuffer(usage, data, copy);
        }
        
    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Short&gt;</code> OpenCL buffer big enough to hold 'length' values of type short.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
    
   * @param elementCount Length of the buffer expressed in elements (for instance, a <code>CLBuffer&lt;Short&gt;</code> of length 10 will actually contain 10 * 2 bytes, as shorts are 2-bytes-long)
   * @return A new buffer allocated on this context.
   */
        public CLBuffer<Short> createShortBuffer(CLMem.Usage usage, long elementCount) {
                return createBuffer(usage, Short.class, elementCount);
        }
        

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Byte&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Byte> createByteBuffer(CLMem.Usage usage,  Buffer  data, boolean copy) {
                return createBuffer(usage, Pointer.pointerToBuffer(data).as(Byte.class), copy);
        }

    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Byte&gt;</code> OpenCL buffer big enough to hold the valid  values of the provided data pointer.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
     * @param data Host data to copy to the buffer
     * @return A new buffer allocated on this context.
   */
        public CLBuffer<Byte> createByteBuffer(CLMem.Usage usage, Pointer<Byte> data) {
                return createByteBuffer(usage, data, true);
        }
        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Byte&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Byte> createByteBuffer(CLMem.Usage usage, Pointer<Byte> data, boolean copy) {
                return createBuffer(usage, data, copy);
        }
        
    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Byte&gt;</code> OpenCL buffer big enough to hold 'length' values of type byte.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
    
   * @param elementCount Length of the buffer expressed in elements (for instance, a <code>CLBuffer&lt;Byte&gt;</code> of length 10 will actually contain 10 * 1 bytes, as bytes are 1-bytes-long)
   * @return A new buffer allocated on this context.
   */
        public CLBuffer<Byte> createByteBuffer(CLMem.Usage usage, long elementCount) {
                return createBuffer(usage, Byte.class, elementCount);
        }
        

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Character&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Character> createCharBuffer(CLMem.Usage usage,  CharBuffer  data, boolean copy) {
                return createBuffer(usage, Pointer.pointerToChars(data), copy);
        }

    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Character&gt;</code> OpenCL buffer big enough to hold the valid  values of the provided data pointer.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
     * @param data Host data to copy to the buffer
     * @return A new buffer allocated on this context.
   */
        public CLBuffer<Character> createCharBuffer(CLMem.Usage usage, Pointer<Character> data) {
                return createCharBuffer(usage, data, true);
        }
        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Character&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Character> createCharBuffer(CLMem.Usage usage, Pointer<Character> data, boolean copy) {
                return createBuffer(usage, data, copy);
        }
        
    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Character&gt;</code> OpenCL buffer big enough to hold 'length' values of type char.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
    
   * @param elementCount Length of the buffer expressed in elements (for instance, a <code>CLBuffer&lt;Character&gt;</code> of length 10 will actually contain 10 * 2 bytes, as chars are 2-bytes-long)
   * @return A new buffer allocated on this context.
   */
        public CLBuffer<Character> createCharBuffer(CLMem.Usage usage, long elementCount) {
                return createBuffer(usage, Character.class, elementCount);
        }
        

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Float&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Float> createFloatBuffer(CLMem.Usage usage,  FloatBuffer  data, boolean copy) {
                return createBuffer(usage, Pointer.pointerToFloats(data), copy);
        }

    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Float&gt;</code> OpenCL buffer big enough to hold the valid  values of the provided data pointer.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
     * @param data Host data to copy to the buffer
     * @return A new buffer allocated on this context.
   */
        public CLBuffer<Float> createFloatBuffer(CLMem.Usage usage, Pointer<Float> data) {
                return createFloatBuffer(usage, data, true);
        }
        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Float&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Float> createFloatBuffer(CLMem.Usage usage, Pointer<Float> data, boolean copy) {
                return createBuffer(usage, data, copy);
        }
        
    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Float&gt;</code> OpenCL buffer big enough to hold 'length' values of type float.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
    
   * @param elementCount Length of the buffer expressed in elements (for instance, a <code>CLBuffer&lt;Float&gt;</code> of length 10 will actually contain 10 * 4 bytes, as floats are 4-bytes-long)
   * @return A new buffer allocated on this context.
   */
        public CLBuffer<Float> createFloatBuffer(CLMem.Usage usage, long elementCount) {
                return createBuffer(usage, Float.class, elementCount);
        }
        

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Double&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Double> createDoubleBuffer(CLMem.Usage usage,  DoubleBuffer  data, boolean copy) {
                return createBuffer(usage, Pointer.pointerToDoubles(data), copy);
        }

    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Double&gt;</code> OpenCL buffer big enough to hold the valid  values of the provided data pointer.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
     * @param data Host data to copy to the buffer
     * @return A new buffer allocated on this context.
   */
        public CLBuffer<Double> createDoubleBuffer(CLMem.Usage usage, Pointer<Double> data) {
                return createDoubleBuffer(usage, data, true);
        }
        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;Double&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
        public CLBuffer<Double> createDoubleBuffer(CLMem.Usage usage, Pointer<Double> data, boolean copy) {
                return createBuffer(usage, data, copy);
        }
        
    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;Double&gt;</code> OpenCL buffer big enough to hold 'length' values of type double.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
    
   * @param elementCount Length of the buffer expressed in elements (for instance, a <code>CLBuffer&lt;Double&gt;</code> of length 10 will actually contain 10 * 8 bytes, as doubles are 8-bytes-long)
   * @return A new buffer allocated on this context.
   */
        public CLBuffer<Double> createDoubleBuffer(CLMem.Usage usage, long elementCount) {
                return createBuffer(usage, Double.class, elementCount);
        }
        

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create an OpenCL buffer with the provided initial values, in copy mode (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>).
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()})
         */
    public <T> CLBuffer<T> createBuffer(CLMem.Usage usage, Pointer<T> data) {
                return createBuffer(usage, data, true);
        }
        
        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create a <code>CLBuffer&lt;N&gt;</code> OpenCL buffer  with the provided initial values.<br>
         * If copy is true (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_COPY_HOST_PTR</a>), then the buffer will be hosted in OpenCL and will have the best performance, but any change done to the OpenCL buffer won't be propagated to the original data pointer.<br>
         * If copy is false (see <a href="http://www.khronos.org/registry/cl/sdk/1.0/docs/man/xhtml/clCreateBuffer.html">CL_MEM_USE_HOST_PTR</a>), then the provided data pointer will be used for storage of the OpenCL buffer. OpenCL might still cache the data in the OpenCL land, so careful use of {@link CLBuffer#map(CLQueue, CLMem.MapFlags, CLEvent...) CLBuffer#map(CLQueue, MapFlags, CLEvent...)} is then necessary to ensure the data is properly synchronized with the buffer. 
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param data Pointer to the initial values, must have known bounds (see {@link Pointer#getValidElements()}).
   * @param data Data buffer.
   * @param copy Whether to copy the input data.
   * @return A new buffer allocated on this context.
         */
    public <T> CLBuffer<T> createBuffer(CLMem.Usage usage, Pointer<T> data, boolean copy) {
        return createBuffer(data.getIO(), data, data.getValidBytes(), usage.getIntFlags() | (copy ? CL_MEM_COPY_HOST_PTR : CL_MEM_USE_HOST_PTR), copy);
        }

    /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
     * Create a <code>CLBuffer&lt;N&gt;</code> OpenCL buffer big enough to hold 'length' values of type T.
     * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
   * @param elementClass Primitive type of the buffer. For instance a buffer of 'int' values can be created with elementClass being Integer.class or int.class indifferently. 
   * @param elementCount Length of the buffer expressed in elements 
   * @return A new buffer allocated on this context.
   */
    public <T> CLBuffer<T> createBuffer(CLMem.Usage usage, Class<T> elementClass, long elementCount) {
        PointerIO<T> io = PointerIO.getInstance(elementClass);
        if (io == null)
                throw new IllegalArgumentException("Unknown target type : " + elementClass.getName());
        return createBuffer(usage, io, elementCount);
        }

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
         * Create an OpenCL buffer big enough to hold the provided amount of values of the specified type.
         * @param usage Usage intended for the pointer in OpenCL kernels : a pointer created with {@link CLMem.Usage#Input} cannot be written to in a kernel.
         * @param io Delegate responsible for reading and writing values.
         * @param elementCount Length of the buffer expressed in elements (for instance, a CLBuffer&lt;Integer&gt; of length 4 will actually contain 4 * 4 bytes, as ints are 4-bytes-long)
         * @deprecated Intended for advanced uses in conjunction with BridJ.
         */
    @Deprecated
    public <T> CLBuffer<T> createBuffer(CLMem.Usage usage, PointerIO<T> io, long elementCount) {
        return createBuffer(io, null, io.getTargetSize() * elementCount, usage.getIntFlags(), false);
        }

        /**
  * Calls <a href="http://www.khronos.org/registry/cl/sdk/1.2/docs/man/xhtml/clCreateBuffer.html">clCreateBuffer</a>.<br>
        */
        @SuppressWarnings("deprecation")
        private <T> CLBuffer<T> createBuffer(PointerIO<T> io, Pointer<T> data, long byteCount, final int CLBufferFlags, final boolean retainBufferReference) {
        if (byteCount <= 0)
                        throw new IllegalArgumentException("Buffer size must be greater than zero (asked for size " + byteCount + ")");
                
                if (byteCount > getMaxMemAllocSize())
            throw new OutOfMemoryError("Requested size for buffer allocation is more than the maximum for this context : " + byteCount + " > " + getMaxMemAllocSize());

        if (data != null) {
                        ByteOrder contextOrder = getByteOrder();
                        ByteOrder dataOrder = data.order();
                        if (contextOrder != null && !dataOrder.equals(contextOrder) && data.getTargetSize() > 1)
                                throw new IllegalArgumentException("Byte order of this context is " + contextOrder + ", but was given pointer to data with order " + dataOrder + ". Please create a pointer with correct byte order (Pointer.order(CLContext.getByteOrder())).");
                }
        
                                ReusablePointers ptrs = ReusablePointers.get();
                Pointer<Integer> pErr = ptrs.pErr;
                long mem;
                int previousAttempts = 0;
                do {
                        mem = CL.clCreateBuffer(
                                getEntity(),
                                CLBufferFlags,
                                byteCount,
                                getPeer(data),
                                getPeer(pErr));
                } while (failedForLackOfMemory(pErr.getInt(), previousAttempts++));
                return new CLBuffer<T>(this, byteCount, mem, retainBufferReference ? data : null, io);
        }

    public ByteOrder getKernelsDefaultByteOrder() {
        if (kernelsDefaultByteOrder == null) {
                        ByteOrder order = null;
                        for (CLDevice device : getDevices()) {
                                ByteOrder devOrder = device.getKernelsDefaultByteOrder();
                                if (order != null && devOrder != order)
                                        return null;
                                order = devOrder;
                        }
                        kernelsDefaultByteOrder = order;
                }
        return kernelsDefaultByteOrder;
    }

    private volatile ByteOrder byteOrder, kernelsDefaultByteOrder;
    
    /**
     * Get the endianness common to all devices of this context, or null if the devices have mismatching endiannesses.
     */
    public ByteOrder getByteOrder() {
        if (byteOrder == null) {
                        ByteOrder order = null;
                        for (CLDevice device : getDevices()) {
                                ByteOrder devOrder = device.getByteOrder();
                                if (order != null && devOrder != order)
                                        return null;
                                order = devOrder;
                        }
                        byteOrder = order;
                }
        return byteOrder;
    }

    private volatile int addressBits = -2;
    
    /**
     * Return the number of bits used to represent a pointer on all of the context's devices, or -1 if not all devices use the same number of bits.<br>
     * Size of size_t type in OpenCL kernels can be obtained with getAddressBits() / 8.
     * @return -1 if the address bits of the context's devices do not match, common address bits otherwise
     */
    public int getAddressBits() {
        if (addressBits == -2) {
            synchronized (this) {
                if (addressBits == -2) {
                    for (CLDevice device : getDevices()) {
                        int bits = device.getAddressBits();
                        if (addressBits != -2 && bits != addressBits) {
                            addressBits = -1;
                            break;
                        }
                        addressBits = bits;
                    }
                }
            }
        }
        return addressBits;
    }

    private volatile Boolean doubleSupported;
    
    /**
     * Whether all the devices in this context support any double-precision numbers (see {@link CLDevice#isDoubleSupported()}).
     */
    public boolean isDoubleSupported() {
        if (doubleSupported == null) {
                boolean supported = true;
                        for (CLDevice device : getDevices()) {
                                if (!device.isDoubleSupported()) {
                                        supported = false;
                                        break;
                                }
                        }
                        doubleSupported = supported;
                }
                return doubleSupported;
        }
        
        private volatile Boolean halfSupported;
    
    /**
     * Whether all the devices in this context support half-precision numbers (see {@link CLDevice#isHalfSupported()}).
     */
    public boolean isHalfSupported() {
                if (halfSupported == null) {
                boolean supported = true;
                        for (CLDevice device : getDevices()) {
                                if (!device.isHalfSupported()) {
                                        supported = false;
                                        break;
                                }
                        }
                        halfSupported = supported;
                }
                return halfSupported;
        }
        
        private volatile Boolean byteAddressableStoreSupported;
    
    public boolean isByteAddressableStoreSupported() {
        if (byteAddressableStoreSupported == null) {
                boolean supported = true;
                        for (CLDevice device : getDevices()) {
                                if (!device.isByteAddressableStoreSupported()) {
                                        supported = false;
                                        break;
                                }
                        }
                        byteAddressableStoreSupported = supported;
                }
                return byteAddressableStoreSupported;
        }
}