Fix API compatibility error

insoow · insoow · commit 2ef427db91b6 · 2017-01-31T11:26:26.000-08:00
This patch fixes a OCV API compatibility error. The error was reported
due to the interface changes of Kernel::run. To resolve the issue,
An overloaded function of Kernel::run is added. It take a flag indicating
whether there are more work to be done with the kernel object without
releasing resources related to it.

Signed-off-by: Woo, Insoo &lt;insoo.woo@intel.com&gt;
diff --git a/modules/core/include/opencv2/core/ocl.hpp b/modules/core/include/opencv2/core/ocl.hpp
@@ -580,6 +580,20 @@ class CV_EXPORTS Kernel
     @param localsize work-group size for each dimension.
     @param sync specify whether to wait for OpenCL computation to finish before return.
     @param q command queue
+    */
+    bool run(int dims, size_t globalsize[],
+             size_t localsize[], bool sync, const Queue& q=Queue());
+    /**
+    Run the OpenCL kernel.
+    @param dims the work problem dimensions. It is the length of globalsize and localsize. It can be either 1, 2 or 3.
+    @param globalsize work items for each dimension.
+    It is not the final globalsize passed to OpenCL.
+    Each dimension will be adjusted to the nearest integer divisible by the corresponding value in localsize.
+    If localsize is NULL, it will still be adjusted depending on dims.
+    The adjusted values are greater than or equal to the original values.
+    @param localsize work-group size for each dimension.
+    @param sync specify whether to wait for OpenCL computation to finish before return.
+    @param q command queue
     @param moreWorkDone specify whether there will the remaining work to be computed (more Kernel::run calls).
     When a computation requires multiple kernel execution by changing input and output buffer offset to get
     the final computation results.
@@ -590,7 +604,8 @@ class CV_EXPORTS Kernel
     kernel.run(..., q, false);
     */
     bool run(int dims, size_t globalsize[],
-             size_t localsize[], bool sync, const Queue& q=Queue(), bool moreWorkDone = false);
+             size_t localsize[], bool sync, bool moreWorkDone, const Queue& q);
+
     bool runTask(bool sync, const Queue& q=Queue());
 
     size_t workGroupSize() const;
diff --git a/modules/core/src/intel_gpu_gemm.cpp b/modules/core/src/intel_gpu_gemm.cpp
@@ -33,14 +33,16 @@
 namespace cv
 {
 
+bool intel_gpu_gemm( UMat A, Size sizeA, UMat B, Size sizeB, UMat D, Size sizeD, double alpha, double beta, bool atrans, bool btrans);
+
 bool intel_gpu_gemm(
     UMat A, Size sizeA,
     UMat B, Size sizeB,
     UMat D, Size sizeD,
     double alpha, double beta, 
     bool atrans, bool btrans)
 {
-    sizeA; sizeB;
+    CV_UNUSED(sizeA); CV_UNUSED(sizeB);
 
     int M = sizeD.height, N = sizeD.width, K = ((atrans)? sizeA.height : sizeA.width);
 
@@ -105,7 +107,7 @@ bool intel_gpu_gemm(
     ocl::Queue q;
     if(!atrans && btrans)
     {
-        ret = k.run(2, global, local, false, q, false);
+        ret = k.run(2, global, local, false, false, q);
     }
     else
     {
@@ -114,12 +116,12 @@ bool intel_gpu_gemm(
        	    k.set(14, &start_index, sizeof(start_index));
             if ((start_index + stride) < K)
     	    {
-    	        ret = k.run(2, global, local, false, q, true);
+    	        ret = k.run(2, global, local, false, true, q);
                 if (!ret) return ret;
     	    }
     	    else
             {
-                ret = k.run(2, global, local, false, q, false);
+                ret = k.run(2, global, local, false, false, q);
     	    }
         }
     }
diff --git a/modules/core/src/ocl.cpp b/modules/core/src/ocl.cpp
@@ -3465,7 +3465,7 @@ int Kernel::set(int i, const KernelArg& arg)
 }
 
 bool Kernel::run(int dims, size_t _globalsize[], size_t _localsize[],
-                 bool sync, const Queue& q, bool moreWorkDone)
+                 bool sync, bool moreWorkDone, const Queue& q)
 {
     CV_INSTRUMENT_REGION_OPENCL_RUN(p->name.c_str());
 
@@ -3511,6 +3511,12 @@ bool Kernel::run(int dims, size_t _globalsize[], size_t _localsize[],
     return retval == CL_SUCCESS;
 }
 
+bool Kernel::run(int dims, size_t _globalsize[], size_t _localsize[],
+                 bool sync, const Queue& q)
+{
+    return run(dims, _globalsize, _localsize, sync, false, q);
+}
+
 bool Kernel::runTask(bool sync, const Queue& q)
 {
     if(!p || !p->handle || p->e != 0)

Original file line number	Diff line number	Diff line change
`@@ -33,14 +33,16 @@`
`33`	`33`	`namespace cv`
`34`	`34`	`{`
`35`	`35`
	`36`	`+bool intel_gpu_gemm( UMat A, Size sizeA, UMat B, Size sizeB, UMat D, Size sizeD, double alpha, double beta, bool atrans, bool btrans);`
	`37`	`+`
`36`	`38`	`bool intel_gpu_gemm(`
`37`	`39`	`UMat A, Size sizeA,`
`38`	`40`	`UMat B, Size sizeB,`
`39`	`41`	`UMat D, Size sizeD,`
`40`	`42`	`double alpha, double beta,`
`41`	`43`	`bool atrans, bool btrans)`
`42`	`44`	`{`
`43`		`- sizeA; sizeB;`
	`45`	`+ CV_UNUSED(sizeA); CV_UNUSED(sizeB);`
`44`	`46`
`45`	`47`	`int M = sizeD.height, N = sizeD.width, K = ((atrans)? sizeA.height : sizeA.width);`
`46`	`48`
`@@ -105,7 +107,7 @@ bool intel_gpu_gemm(`
`105`	`107`	`ocl::Queue q;`
`106`	`108`	`if(!atrans && btrans)`
`107`	`109`	`{`
`108`		`- ret = k.run(2, global, local, false, q, false);`
	`110`	`+ ret = k.run(2, global, local, false, false, q);`
`109`	`111`	`}`
`110`	`112`	`else`
`111`	`113`	`{`
`@@ -114,12 +116,12 @@ bool intel_gpu_gemm(`
`114`	`116`	`k.set(14, &start_index, sizeof(start_index));`
`115`	`117`	`if ((start_index + stride) < K)`
`116`	`118`	`{`
`117`		`- ret = k.run(2, global, local, false, q, true);`
	`119`	`+ ret = k.run(2, global, local, false, true, q);`
`118`	`120`	`if (!ret) return ret;`
`119`	`121`	`}`
`120`	`122`	`else`
`121`	`123`	`{`
`122`		`- ret = k.run(2, global, local, false, q, false);`
	`124`	`+ ret = k.run(2, global, local, false, false, q);`
`123`	`125`	`}`
`124`	`126`	`}`
`125`	`127`	`}`
Original file line number	Diff line number	Diff line change
`@@ -3465,7 +3465,7 @@ int Kernel::set(int i, const KernelArg& arg)`
`3465`	`3465`	`}`
`3466`	`3466`
`3467`	`3467`	`bool Kernel::run(int dims, size_t _globalsize[], size_t _localsize[],`
`3468`		`- bool sync, const Queue& q, bool moreWorkDone)`
	`3468`	`+ bool sync, bool moreWorkDone, const Queue& q)`
`3469`	`3469`	`{`
`3470`	`3470`	`CV_INSTRUMENT_REGION_OPENCL_RUN(p->name.c_str());`
`3471`	`3471`
`@@ -3511,6 +3511,12 @@ bool Kernel::run(int dims, size_t _globalsize[], size_t _localsize[],`
`3511`	`3511`	`return retval == CL_SUCCESS;`
`3512`	`3512`	`}`
`3513`	`3513`
	`3514`	`+bool Kernel::run(int dims, size_t _globalsize[], size_t _localsize[],`
	`3515`	`+ bool sync, const Queue& q)`
	`3516`	`+{`
	`3517`	`+ return run(dims, _globalsize, _localsize, sync, false, q);`
	`3518`	`+}`
	`3519`	`+`
`3514`	`3520`	`bool Kernel::runTask(bool sync, const Queue& q)`
`3515`	`3521`	`{`
`3516`	`3522`	`if(!p \|\| !p->handle \|\| p->e != 0)`