added draft version of MultiGpuMgr (it isn't tested on multi GPU machine yet)

modules/gpu/include/opencv2/gpu/gpu.hpp

@@ -94,11 +94,15 @@ namespace cv
         class CV_EXPORTS DeviceInfo
         {
         public:
+            // Creates DeviceInfo object for the current GPU
             DeviceInfo() : device_id_(getDevice()) { query(); }
+
+            // Creates DeviceInfo object for the given GPU
             DeviceInfo(int device_id) : device_id_(device_id) { query(); }
 
             string name() const { return name_; }
 
+            // Return compute capability versions
             int majorVersion() const { return majorVersion_; }
             int minorVersion() const { return minorVersion_; }
 
@@ -107,7 +111,10 @@ namespace cv
             size_t freeMemory() const;
             size_t totalMemory() const;
 
+            // Checks whether device supports the given feature
             bool supports(GpuFeature feature) const;
+
+            // Checks whether the GPU module can be run on the given device
             bool isCompatible() const;
 
         private:
@@ -122,6 +129,33 @@ namespace cv
             int minorVersion_;
         };
 
+        /////////////////////////// Multi GPU Manager //////////////////////////////
+
+        // Provides functionality for working with many GPUs. Object of this
+        // class must be created before any OpenCV GPU call and no call must
+        // be done after its destruction.
+        class CV_EXPORTS MultiGpuMgr
+        {
+        public:
+            MultiGpuMgr();
+
+            // Returns the current GPU id (or BAD_GPU_ID if no GPU is active)
+            int currentGpuId() const;
+
+            // Makes the given GPU active
+            void gpuOn(int gpu_id);
+
+            // Finishes the piece of work with the current GPU
+            void gpuOff();
+
+            static const int BAD_GPU_ID = -1;
+
+        private:
+            class Impl;
+
+            Ptr<Impl> impl_;
+        };
+
         //////////////////////////////// Error handling ////////////////////////
 
         CV_EXPORTS void error(const char *error_string, const char *file, const int line, const char *func);
@@ -1129,7 +1163,7 @@ namespace cv
 
         /////////////////////////// StereoConstantSpaceBP ///////////////////////////
         // "A Constant-Space Belief Propagation Algorithm for Stereo Matching"
-        // Qingxiong Yang, Liang Wang†, Narendra Ahuja
+        // Qingxiong Yang, Liang Wang<EFBFBD>, Narendra Ahuja
         // http://vision.ai.uiuc.edu/~qyang6/
 
         class CV_EXPORTS StereoConstantSpaceBP
@@ -1195,7 +1229,7 @@ namespace cv
 
         /////////////////////////// DisparityBilateralFilter ///////////////////////////
         // Disparity map refinement using joint bilateral filtering given a single color image.
-        // Qingxiong Yang, Liang Wang†, Narendra Ahuja
+        // Qingxiong Yang, Liang Wang<EFBFBD>, Narendra Ahuja
         // http://vision.ai.uiuc.edu/~qyang6/
 
         class CV_EXPORTS DisparityBilateralFilter

modules/gpu/src/multi_gpu_mgr.cpp

@@ -0,0 +1,160 @@
+/*M///////////////////////////////////////////////////////////////////////////////////////
+//
+//  IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING.
+//
+//  By downloading, copying, installing or using the software you agree to this license.
+//  If you do not agree to this license, do not download, install,
+//  copy or use the software.
+//
+//
+//                           License Agreement
+//                For Open Source Computer Vision Library
+//
+// Copyright (C) 2000-2008, Intel Corporation, all rights reserved.
+// Copyright (C) 2009, Willow Garage Inc., all rights reserved.
+// Third party copyrights are property of their respective owners.
+//
+// Redistribution and use in source and binary forms, with or without modification,
+// are permitted provided that the following conditions are met:
+//
+//   * Redistribution's of source code must retain the above copyright notice,
+//     this list of conditions and the following disclaimer.
+//
+//   * Redistribution's 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.
+//
+//   * The name of the copyright holders may not be used to endorse or promote products
+//     derived from this software without specific prior written permission.
+//
+// This software is provided by the copyright holders 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 Intel Corporation or 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.
+//
+//M*/
+
+#include "precomp.hpp"
+
+#if !defined(HAVE_CUDA)
+
+namespace cv { namespace gpu {
+
+class MultiGpuMgr::Impl {};
+MultiGpuMgr::MultiGpuMgr() { throw_nogpu(); }
+int MultiGpuMgr::currentGpuId() const { throw_nogpu(); return 0; }
+void MultiGpuMgr::gpuOn(int) { throw_nogpu(); }
+void MultiGpuMgr::gpuOff() { throw_nogpu(); }
+
+}}
+
+#else
+
+#include <stack>
+#include <vector>
+#include <cuda.h>
+
+#define cuSafeCall(expr) safeCall(expr, #expr, __FILE__, __LINE__)
+
+using namespace std;
+
+namespace cv { namespace gpu {
+
+
+class MultiGpuMgr::Impl
+{
+public:
+    Impl();
+
+    ~Impl()
+    {
+        for (int i = 0; i < num_devices_; ++i)
+            cuSafeCall(cuCtxDestroy(contexts_[i]));
+    }
+
+    int currentGpuId() const
+    {
+        if (gpu_id_stack_.empty())
+            return MultiGpuMgr::BAD_GPU_ID;
+        return gpu_id_stack_.top();
+    }
+
+    void gpuOn(int gpu_id)
+    {
+        if (gpu_id != currentGpuId())
+        {
+            cuSafeCall(cuCtxPushCurrent(contexts_[gpu_id]));
+            gpu_id_stack_.push(gpu_id);
+        }
+    }
+
+    void gpuOff()
+    {
+        if (gpu_id_stack_.empty())
+            return;
+
+        CUcontext prev_context;
+        cuSafeCall(cuCtxPopCurrent(&prev_context));
+
+        gpu_id_stack_.pop();
+    }
+
+private:
+    void safeCall(CUresult code, const char* expr, const char* file, int line)
+    {
+        if (code != CUDA_SUCCESS)
+            error(expr, file, line, "");
+    }
+
+    int num_devices_;
+    stack<int> gpu_id_stack_;
+    vector<CUcontext> contexts_;
+};
+
+
+MultiGpuMgr::Impl::Impl(): num_devices_(0)
+{
+    num_devices_ = getCudaEnabledDeviceCount();
+    contexts_.resize(num_devices_);
+
+    CUdevice device;
+    CUcontext prev_context;
+    for (int i = 0; i < num_devices_; ++i)
+    {
+        cuSafeCall(cuDeviceGet(&device, i));
+        cuSafeCall(cuCtxCreate(&contexts_[i], 0, device));
+        cuSafeCall(cuCtxPopCurrent(&prev_context));
+    }
+}
+
+
+
+MultiGpuMgr::MultiGpuMgr(): impl_(new Impl()) {}
+
+
+int MultiGpuMgr::currentGpuId() const
+{
+    return impl_->currentGpuId();
+}
+
+
+void MultiGpuMgr::gpuOn(int gpu_id)
+{
+    impl_->gpuOn(gpu_id);
+}
+
+
+void MultiGpuMgr::gpuOff()
+{
+    impl_->gpuOff();
+}
+
+}}
+
+#endif