prefix many objects with starpu_

examples/axpy/axpy.c

@@ -66,9 +66,9 @@ void axpy_gpu(void *descr[], __attribute__((unused)) void *arg)
 static starpu_codelet axpy_cl = {
         .where =
 #ifdef USE_CUDA
-                CUDA|
+                STARPU_CUDA|
 #endif
-                CORE,
+                STARPU_CORE,
 
 	.core_func = axpy_cpu,
 #ifdef USE_CUDA
@@ -133,10 +133,10 @@ int main(int argc, char **argv)
 
 		task->cl_arg = α
 
-		task->buffers[0].handle = get_sub_data(handle_x, 1, b);
+		task->buffers[0].handle = starpu_get_sub_data(handle_x, 1, b);
 		task->buffers[0].mode = STARPU_R;
 		
-		task->buffers[1].handle = get_sub_data(handle_y, 1, b);
+		task->buffers[1].handle = starpu_get_sub_data(handle_y, 1, b);
 		task->buffers[1].mode = STARPU_RW;
 		
 		starpu_submit_task(task);

examples/basic-examples/hello-world.c

@@ -54,7 +54,7 @@ starpu_codelet cl =
 {
 	/* this codelet may only be executed on a CPU, and its cpu
  	 * implementation is function "cpu_func" */
-	.where = CORE,
+	.where = STARPU_CORE,
 	.core_func = cpu_func,
 	/* the codelet does not manipulate any data that is managed
 	 * by our DSM */

examples/basic-examples/mult.c

@@ -116,9 +116,9 @@ static void cpu_mult(void *descr[], __attribute__((unused))  void *arg)
 	uint32_t ldA, ldB, ldC;
 
 	/* .blas.ptr gives a pointer to the first element of the local copy */
-	subA = (float *)GET_BLAS_PTR(descr[0]);
-	subB = (float *)GET_BLAS_PTR(descr[1]);
-	subC = (float *)GET_BLAS_PTR(descr[2]);
+	subA = (float *)STARPU_GET_BLAS_PTR(descr[0]);
+	subB = (float *)STARPU_GET_BLAS_PTR(descr[1]);
+	subC = (float *)STARPU_GET_BLAS_PTR(descr[2]);
 
 	/* .blas.nx is the number of rows (consecutive elements) and .blas.ny
 	 * is the number of lines that are separated by .blas.ld elements (ld
@@ -126,13 +126,13 @@ static void cpu_mult(void *descr[], __attribute__((unused))  void *arg)
 	 * NB: in case some filters were used, the leading dimension is not
 	 * guaranteed to be the same in main memory (on the original matrix)
 	 * and on the accelerator! */
-	nxC = GET_BLAS_NX(descr[2]);
-	nyC = GET_BLAS_NY(descr[2]);
-	nyA = GET_BLAS_NY(descr[0]);
+	nxC = STARPU_GET_BLAS_NX(descr[2]);
+	nyC = STARPU_GET_BLAS_NY(descr[2]);
+	nyA = STARPU_GET_BLAS_NY(descr[0]);
 
-	ldA = GET_BLAS_LD(descr[0]);
-	ldB = GET_BLAS_LD(descr[1]);
-	ldC = GET_BLAS_LD(descr[2]);
+	ldA = STARPU_GET_BLAS_LD(descr[0]);
+	ldB = STARPU_GET_BLAS_LD(descr[1]);
+	ldC = STARPU_GET_BLAS_LD(descr[2]);
 
 	/* we assume a FORTRAN-ordering! */
 	unsigned i,j,k;
@@ -277,7 +277,7 @@ static void partition_mult_data(void)
 }
 
 static struct starpu_perfmodel_t mult_perf_model = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = "mult_perf_model"
 };
 
@@ -293,7 +293,7 @@ static void launch_tasks(void)
 
 	starpu_codelet cl = {
 		/* we can only execute that kernel on a CPU yet */
-		.where = CORE,
+		.where = STARPU_CORE,
 		/* CPU implementation of the codelet */
 		.core_func = cpu_mult,
 		/* the codelet manipulates 3 buffers that are managed by the
@@ -339,9 +339,9 @@ static void launch_tasks(void)
 			 * identified by "tasky" (respectively "taskx). The "1"
 			 * tells StarPU that there is a single argument to the
 			 * variable-arity function get_sub_data */
-			task->buffers[0].handle = get_sub_data(A_handle, 1, tasky);
+			task->buffers[0].handle = starpu_get_sub_data(A_handle, 1, tasky);
 			task->buffers[0].mode = STARPU_R;
-			task->buffers[1].handle = get_sub_data(B_handle, 1, taskx);
+			task->buffers[1].handle = starpu_get_sub_data(B_handle, 1, taskx);
 			task->buffers[1].mode = STARPU_R;
 
 			/* 2 filters were applied on matrix C, so we give
@@ -353,8 +353,7 @@ static void launch_tasks(void)
 			 * NB2: get_sub_data(C_handle, 2, taskx, tasky) is
 			 * equivalent to
 			 * get_sub_data(get_sub_data(C_handle, 1, taskx), 1, tasky)*/
-			task->buffers[2].handle = 
-				get_sub_data(C_handle, 2, taskx, tasky);
+			task->buffers[2].handle = starpu_get_sub_data(C_handle, 2, taskx, tasky);
 			task->buffers[2].mode = STARPU_W;
 
 			/* this is not a blocking call since task->synchronous = 0 */

examples/basic-examples/vector-scal.c

@@ -100,7 +100,7 @@ int main(int argc, char **argv)
 	task->synchronous = 1;
 
 	starpu_codelet cl = {
-		.where = CORE,
+		.where = STARPU_CORE,
 		/* CPU implementation of the codelet */
 		.core_func = scal_func,
 		.nbuffers = 1

examples/cholesky/dw_cholesky.c

@@ -37,7 +37,7 @@ static struct starpu_task *create_task(starpu_tag_t id)
 
 static starpu_codelet cl11 =
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = chol_core_codelet_update_u11,
 #ifdef USE_CUDA
 	.cuda_func = chol_cublas_codelet_update_u11,
@@ -55,12 +55,12 @@ static struct starpu_task * create_task_11(starpu_data_handle dataA, unsigned k)
 	task->cl = &cl11;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k);
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k);
 	task->buffers[0].mode = STARPU_RW;
 
 	/* this is an important task */
 	if (!noprio)
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 
 	/* enforce dependencies ... */
 	if (k > 0) {
@@ -72,7 +72,7 @@ static struct starpu_task * create_task_11(starpu_data_handle dataA, unsigned k)
 
 static starpu_codelet cl21 =
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = chol_core_codelet_update_u21,
 #ifdef USE_CUDA
 	.cuda_func = chol_cublas_codelet_update_u21,
@@ -88,13 +88,13 @@ static void create_task_21(starpu_data_handle dataA, unsigned k, unsigned j)
 	task->cl = &cl21;	
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k); 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k); 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, k, j); 
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, k, j); 
 	task->buffers[1].mode = STARPU_RW;
 
 	if (!noprio && (j == k+1)) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */
@@ -110,7 +110,7 @@ static void create_task_21(starpu_data_handle dataA, unsigned k, unsigned j)
 
 static starpu_codelet cl22 =
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = chol_core_codelet_update_u22,
 #ifdef USE_CUDA
 	.cuda_func = chol_cublas_codelet_update_u22,
@@ -128,15 +128,15 @@ static void create_task_22(starpu_data_handle dataA, unsigned k, unsigned i, uns
 	task->cl = &cl22;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, i); 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, i); 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, k, j); 
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, k, j); 
 	task->buffers[1].mode = STARPU_R;
-	task->buffers[2].handle = get_sub_data(dataA, 2, i, j); 
+	task->buffers[2].handle = starpu_get_sub_data(dataA, 2, i, j); 
 	task->buffers[2].mode = STARPU_RW;
 
 	if (!noprio && (i == k + 1) && (j == k +1) ) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */

examples/cholesky/dw_cholesky_grain.c

@@ -37,7 +37,7 @@ static struct starpu_task *create_task(starpu_tag_t id)
 
 static starpu_codelet cl11 =
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = chol_core_codelet_update_u11,
 #ifdef USE_CUDA
 	.cuda_func = chol_cublas_codelet_update_u11,
@@ -55,11 +55,11 @@ static struct starpu_task * create_task_11(starpu_data_handle dataA, unsigned k,
 	task->cl = &cl11;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k);
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k);
 	task->buffers[0].mode = STARPU_RW;
 
 	/* this is an important task */
-	task->priority = MAX_PRIO;
+	task->priority = STARPU_MAX_PRIO;
 
 	/* enforce dependencies ... */
 	if (k > 0) {
@@ -71,7 +71,7 @@ static struct starpu_task * create_task_11(starpu_data_handle dataA, unsigned k,
 
 static starpu_codelet cl21 =
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = chol_core_codelet_update_u21,
 #ifdef USE_CUDA
 	.cuda_func = chol_cublas_codelet_update_u21,
@@ -87,13 +87,13 @@ static void create_task_21(starpu_data_handle dataA, unsigned k, unsigned j, uns
 	task->cl = &cl21;	
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k); 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k); 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, k, j); 
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, k, j); 
 	task->buffers[1].mode = STARPU_RW;
 
 	if (j == k+1) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */
@@ -109,7 +109,7 @@ static void create_task_21(starpu_data_handle dataA, unsigned k, unsigned j, uns
 
 static starpu_codelet cl22 =
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = chol_core_codelet_update_u22,
 #ifdef USE_CUDA
 	.cuda_func = chol_cublas_codelet_update_u22,
@@ -127,15 +127,15 @@ static void create_task_22(starpu_data_handle dataA, unsigned k, unsigned i, uns
 	task->cl = &cl22;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, i); 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, i); 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, k, j); 
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, k, j); 
 	task->buffers[1].mode = STARPU_R;
-	task->buffers[2].handle = get_sub_data(dataA, 2, i, j); 
+	task->buffers[2].handle = starpu_get_sub_data(dataA, 2, i, j); 
 	task->buffers[2].mode = STARPU_RW;
 
 	if ( (i == k + 1) && (j == k +1) ) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */

examples/cholesky/dw_cholesky_kernels.c

@@ -30,17 +30,17 @@
 static inline void chol_common_core_codelet_update_u22(void *descr[], int s, __attribute__((unused)) void *_args)
 {
 	//printf("22\n");
-	float *left 	= (float *)GET_BLAS_PTR(descr[0]);
-	float *right 	= (float *)GET_BLAS_PTR(descr[1]);
-	float *center 	= (float *)GET_BLAS_PTR(descr[2]);
+	float *left 	= (float *)STARPU_GET_BLAS_PTR(descr[0]);
+	float *right 	= (float *)STARPU_GET_BLAS_PTR(descr[1]);
+	float *center 	= (float *)STARPU_GET_BLAS_PTR(descr[2]);
 
-	unsigned dx = GET_BLAS_NY(descr[2]);
-	unsigned dy = GET_BLAS_NX(descr[2]);
-	unsigned dz = GET_BLAS_NY(descr[0]);
+	unsigned dx = STARPU_GET_BLAS_NY(descr[2]);
+	unsigned dy = STARPU_GET_BLAS_NX(descr[2]);
+	unsigned dz = STARPU_GET_BLAS_NY(descr[0]);
 
-	unsigned ld21 = GET_BLAS_LD(descr[0]);
-	unsigned ld12 = GET_BLAS_LD(descr[1]);
-	unsigned ld22 = GET_BLAS_LD(descr[2]);
+	unsigned ld21 = STARPU_GET_BLAS_LD(descr[0]);
+	unsigned ld12 = STARPU_GET_BLAS_LD(descr[1]);
+	unsigned ld22 = STARPU_GET_BLAS_LD(descr[2]);
 
 #ifdef USE_CUDA
 	cublasStatus st;
@@ -91,14 +91,14 @@ static inline void chol_common_codelet_update_u21(void *descr[], int s, __attrib
 	float *sub11;
 	float *sub21;
 
-	sub11 = (float *)GET_BLAS_PTR(descr[0]);
-	sub21 = (float *)GET_BLAS_PTR(descr[1]);
+	sub11 = (float *)STARPU_GET_BLAS_PTR(descr[0]);
+	sub21 = (float *)STARPU_GET_BLAS_PTR(descr[1]);
 
-	unsigned ld11 = GET_BLAS_LD(descr[0]);
-	unsigned ld21 = GET_BLAS_LD(descr[1]);
+	unsigned ld11 = STARPU_GET_BLAS_LD(descr[0]);
+	unsigned ld21 = STARPU_GET_BLAS_LD(descr[1]);
 
-	unsigned nx21 = GET_BLAS_NY(descr[1]);
-	unsigned ny21 = GET_BLAS_NX(descr[1]);
+	unsigned nx21 = STARPU_GET_BLAS_NY(descr[1]);
+	unsigned ny21 = STARPU_GET_BLAS_NX(descr[1]);
 
 	switch (s) {
 		case 0:
@@ -137,10 +137,10 @@ static inline void chol_common_codelet_update_u11(void *descr[], int s, __attrib
 //	printf("11\n");
 	float *sub11;
 
-	sub11 = (float *)GET_BLAS_PTR(descr[0]); 
+	sub11 = (float *)STARPU_GET_BLAS_PTR(descr[0]); 
 
-	unsigned nx = GET_BLAS_NY(descr[0]);
-	unsigned ld = GET_BLAS_LD(descr[0]);
+	unsigned nx = STARPU_GET_BLAS_NY(descr[0]);
+	unsigned ld = STARPU_GET_BLAS_LD(descr[0]);
 
 	unsigned z;
 

examples/cholesky/dw_cholesky_models.c

@@ -129,7 +129,7 @@ struct starpu_perfmodel_t chol_model_11 = {
 		[STARPU_CORE_DEFAULT] = { .cost_model = core_chol_task_11_cost },
 		[STARPU_CUDA_DEFAULT] = { .cost_model = cuda_chol_task_11_cost }
 	},
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = "chol_model_11"
 };
 
@@ -138,7 +138,7 @@ struct starpu_perfmodel_t chol_model_21 = {
 		[STARPU_CORE_DEFAULT] = { .cost_model = core_chol_task_21_cost },
 		[STARPU_CUDA_DEFAULT] = { .cost_model = cuda_chol_task_21_cost }
 	},
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = "chol_model_21"
 };
 
@@ -147,6 +147,6 @@ struct starpu_perfmodel_t chol_model_22 = {
 		[STARPU_CORE_DEFAULT] = { .cost_model = core_chol_task_22_cost },
 		[STARPU_CUDA_DEFAULT] = { .cost_model = cuda_chol_task_22_cost }
 	},
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = "chol_model_22"
 };

examples/cholesky/dw_cholesky_no_stride.c

@@ -47,7 +47,7 @@ static void terminal_callback(void *argcb)
 
 static starpu_codelet cl11 =
 {
-	.where = CORE|CUDA|GORDON,
+	.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON,
 	.core_func = chol_core_codelet_update_u11,
 #ifdef USE_CUDA
 	.cuda_func = chol_cublas_codelet_update_u11,
@@ -76,7 +76,7 @@ static struct starpu_task * create_task_11(unsigned k, unsigned nblocks, sem_t *
 	task->buffers[0].mode = STARPU_RW;
 
 	/* this is an important task */
-	task->priority = MAX_PRIO;
+	task->priority = STARPU_MAX_PRIO;
 
 	/* enforce dependencies ... */
 	if (k > 0) {
@@ -94,7 +94,7 @@ static struct starpu_task * create_task_11(unsigned k, unsigned nblocks, sem_t *
 
 static starpu_codelet cl21 =
 {
-	.where = CORE|CUDA|GORDON,
+	.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON,
 	.core_func = chol_core_codelet_update_u21,
 #ifdef USE_CUDA
 	.cuda_func = chol_cublas_codelet_update_u21,
@@ -123,7 +123,7 @@ static void create_task_21(unsigned k, unsigned j)
 	task->buffers[1].mode = STARPU_RW;
 
 	if (j == k+1) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */
@@ -139,7 +139,7 @@ static void create_task_21(unsigned k, unsigned j)
 
 static starpu_codelet cl22 =
 {
-	.where = CORE|CUDA|GORDON,
+	.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON,
 	.core_func = chol_core_codelet_update_u22,
 #ifdef USE_CUDA
 	.cuda_func = chol_cublas_codelet_update_u22,
@@ -172,7 +172,7 @@ static void create_task_22(unsigned k, unsigned i, unsigned j)
 	task->buffers[2].mode = STARPU_RW;
 
 	if ( (i == k + 1) && (j == k +1) ) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */

examples/common/blas_model.h

@@ -23,7 +23,7 @@ double gemm_cost(starpu_buffer_descr *descr);
 
 static struct starpu_perfmodel_t sgemm_model = {
 	.cost_model = gemm_cost,
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = "sgemm_atlas"
 #elif defined(GOTO)
@@ -35,7 +35,7 @@ static struct starpu_perfmodel_t sgemm_model = {
 
 static struct starpu_perfmodel_t sgemm_model_common = {
 	.cost_model = gemm_cost,
-	.type = COMMON,
+	.type = STARPU_COMMON,
 };
 
 #endif // __BLAS_MODEL_H__

examples/heat/dw_factolu.c

@@ -32,7 +32,7 @@ static unsigned no_prio = 0;
 
 static starpu_codelet cl11 =
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dw_core_codelet_update_u11,
 #ifdef USE_CUDA
 	.cuda_func = dw_cublas_codelet_update_u11,
@@ -43,7 +43,7 @@ static starpu_codelet cl11 =
 
 static starpu_codelet cl12 =
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dw_core_codelet_update_u12,
 #ifdef USE_CUDA
 	.cuda_func = dw_cublas_codelet_update_u12,
@@ -54,7 +54,7 @@ static starpu_codelet cl12 =
 
 static starpu_codelet cl21 =
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dw_core_codelet_update_u21,
 #ifdef USE_CUDA
 	.cuda_func = dw_cublas_codelet_update_u21,
@@ -65,7 +65,7 @@ static starpu_codelet cl21 =
 
 static starpu_codelet cl22 =
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dw_core_codelet_update_u22,
 #ifdef USE_CUDA
 	.cuda_func = dw_cublas_codelet_update_u22,
@@ -106,7 +106,7 @@ void dw_callback_v2_codelet_update_u22(void *argcb)
 			task->cl_arg = u11arg;
 
 			task->buffers[0].handle =
-				get_sub_data(args->dataA, 2, k+1, k+1);
+				starpu_get_sub_data(args->dataA, 2, k+1, k+1);
 			task->buffers[0].mode = STARPU_RW;
 	
 		u11arg->dataA = args->dataA;
@@ -115,7 +115,7 @@ void dw_callback_v2_codelet_update_u22(void *argcb)
 
 		/* schedule the codelet */
 		if (!no_prio)
-			task->priority = MAX_PRIO;
+			task->priority = STARPU_MAX_PRIO;
 
 		starpu_submit_task(task);
 	}
@@ -145,10 +145,10 @@ void dw_callback_v2_codelet_update_u22(void *argcb)
 					u21a->dataA = args->dataA;
 
 					task21->buffers[0].handle = 
-						get_sub_data(args->dataA, 2, u21a->i, u21a->i);
+						starpu_get_sub_data(args->dataA, 2, u21a->i, u21a->i);
 					task21->buffers[0].mode = STARPU_R;
 					task21->buffers[1].handle =
-						get_sub_data(args->dataA, 2, u21a->i, u21a->k);
+						starpu_get_sub_data(args->dataA, 2, u21a->i, u21a->k);
 					task21->buffers[1].mode = STARPU_RW;
 		
 					starpu_submit_task(task21);
@@ -179,9 +179,9 @@ void dw_callback_v2_codelet_update_u22(void *argcb)
 					u12a->nblocks = args->nblocks;
 					u12a->dataA = args->dataA;
 
-					task12->buffers[0].handle = get_sub_data(args->dataA, 2, u12a->i, u12a->i); 
+					task12->buffers[0].handle = starpu_get_sub_data(args->dataA, 2, u12a->i, u12a->i); 
 					task12->buffers[0].mode = STARPU_R;
-					task12->buffers[1].handle = get_sub_data(args->dataA, 2, u12a->k, u12a->i); 
+					task12->buffers[1].handle = starpu_get_sub_data(args->dataA, 2, u12a->k, u12a->i); 
 					task12->buffers[1].mode = STARPU_RW;
 					
 					starpu_submit_task(task12);
@@ -231,18 +231,18 @@ void dw_callback_v2_codelet_update_u12(void *argcb)
 				u22a->dataA = args->dataA;
 				u22a->nblocks = nblocks;
 
-				task22->buffers[0].handle = get_sub_data(args->dataA, 2, u22a->i, u22a->k);
+				task22->buffers[0].handle = starpu_get_sub_data(args->dataA, 2, u22a->i, u22a->k);
 				task22->buffers[0].mode = STARPU_R;
 
-				task22->buffers[1].handle = get_sub_data(args->dataA, 2, u22a->k, u22a->j);
+				task22->buffers[1].handle = starpu_get_sub_data(args->dataA, 2, u22a->k, u22a->j);
 				task22->buffers[1].mode = STARPU_R;
 
-				task22->buffers[2].handle = get_sub_data(args->dataA, 2, u22a->i, u22a->j);
+				task22->buffers[2].handle = starpu_get_sub_data(args->dataA, 2, u22a->i, u22a->j);
 				task22->buffers[2].mode = STARPU_RW;
 				
 				/* schedule that codelet */
 				if (!no_prio && (slicey == i+1))
-					task22->priority = MAX_PRIO;
+					task22->priority = STARPU_MAX_PRIO;
 
 				starpu_submit_task(task22);
 			}
@@ -290,18 +290,18 @@ void dw_callback_v2_codelet_update_u21(void *argcb)
 				u22a->dataA = args->dataA;
 				u22a->nblocks = nblocks;
 
-				task22->buffers[0].handle = get_sub_data(args->dataA, 2, u22a->i, u22a->k);
+				task22->buffers[0].handle = starpu_get_sub_data(args->dataA, 2, u22a->i, u22a->k);
 				task22->buffers[0].mode = STARPU_R;
 
-				task22->buffers[1].handle = get_sub_data(args->dataA, 2, u22a->k, u22a->j);
+				task22->buffers[1].handle = starpu_get_sub_data(args->dataA, 2, u22a->k, u22a->j);
 				task22->buffers[1].mode = STARPU_R;
 
-				task22->buffers[2].handle = get_sub_data(args->dataA, 2, u22a->i, u22a->j);
+				task22->buffers[2].handle = starpu_get_sub_data(args->dataA, 2, u22a->i, u22a->j);
 				task22->buffers[2].mode = STARPU_RW;
 				
 				/* schedule that codelet */
 				if (!no_prio && (slicex == i+1))
-					task22->priority = MAX_PRIO;
+					task22->priority = STARPU_MAX_PRIO;
 
 				starpu_submit_task(task22);
 			}
@@ -362,13 +362,13 @@ void dw_callback_v2_codelet_update_u11(void *argcb)
 					u12a->nblocks = args->nblocks;
 					u12a->dataA = args->dataA;
 
-					task12->buffers[0].handle = get_sub_data(args->dataA, 2, u12a->i, u12a->i); 
+					task12->buffers[0].handle = starpu_get_sub_data(args->dataA, 2, u12a->i, u12a->i); 
 					task12->buffers[0].mode = STARPU_R;
-					task12->buffers[1].handle = get_sub_data(args->dataA, 2, u12a->k, u12a->i); 
+					task12->buffers[1].handle = starpu_get_sub_data(args->dataA, 2, u12a->k, u12a->i); 
 					task12->buffers[1].mode = STARPU_RW;
 
 					if (!no_prio && (slice == i +1))
-						task12->priority = MAX_PRIO;
+						task12->priority = STARPU_MAX_PRIO;
 
 					starpu_submit_task(task12);
 				}
@@ -399,13 +399,13 @@ void dw_callback_v2_codelet_update_u11(void *argcb)
 					u21a->nblocks = args->nblocks;
 					u21a->dataA = args->dataA;
 
-					task21->buffers[0].handle = get_sub_data(args->dataA, 2, u21a->i, u21a->i);
+					task21->buffers[0].handle = starpu_get_sub_data(args->dataA, 2, u21a->i, u21a->i);
 					task21->buffers[0].mode = STARPU_R;
-					task21->buffers[1].handle = get_sub_data(args->dataA, 2, u21a->i, u21a->k);
+					task21->buffers[1].handle = starpu_get_sub_data(args->dataA, 2, u21a->i, u21a->k);
 					task21->buffers[1].mode = STARPU_RW;
 		
 					if (!no_prio && (slice == i +1))
-						task21->priority = MAX_PRIO;
+						task21->priority = STARPU_MAX_PRIO;
 
 					starpu_submit_task(task21);
 				}
@@ -479,17 +479,17 @@ void dw_callback_codelet_update_u11(void *argcb)
 			u21a->remaining = remaining;
 
 			task12->buffers[0].handle = 
-				get_sub_data(args->dataA, 2, u12a->i, u12a->i); 
+				starpu_get_sub_data(args->dataA, 2, u12a->i, u12a->i); 
 			task12->buffers[0].mode = STARPU_R;
 			task12->buffers[1].handle = 
-				get_sub_data(args->dataA, 2, u12a->k, u12a->i); 
+				starpu_get_sub_data(args->dataA, 2, u12a->k, u12a->i); 
 			task12->buffers[1].mode = STARPU_RW;
 
 			task21->buffers[0].handle = 
-				get_sub_data(args->dataA, 2, u21a->i, u21a->i);
+				starpu_get_sub_data(args->dataA, 2, u21a->i, u21a->i);
 			task21->buffers[0].mode = STARPU_R;
 			task21->buffers[1].handle = 
-				get_sub_data(args->dataA, 2, u21a->i, u21a->k);
+				starpu_get_sub_data(args->dataA, 2, u21a->i, u21a->k);
 			task21->buffers[1].mode = STARPU_RW;
 		
 			starpu_submit_task(task12);
@@ -517,7 +517,7 @@ void dw_callback_codelet_update_u22(void *argcb)
 			task->cl = &cl11;
 			task->cl_arg = u11arg;
 
-			task->buffers[0].handle = get_sub_data(args->dataA, 2, args->k + 1, args->k + 1);
+			task->buffers[0].handle = starpu_get_sub_data(args->dataA, 2, args->k + 1, args->k + 1);
 			task->buffers[0].mode = STARPU_RW;
 	
 		u11arg->dataA = args->dataA;
@@ -566,13 +566,13 @@ void dw_callback_codelet_update_u12_21(void *argcb)
 				u22a->nblocks = nblocks;
 				u22a->remaining = remaining;
 
-				task22->buffers[0].handle = get_sub_data(args->dataA, 2, u22a->i, u22a->k);
+				task22->buffers[0].handle = starpu_get_sub_data(args->dataA, 2, u22a->i, u22a->k);
 				task22->buffers[0].mode = STARPU_R;
 
-				task22->buffers[1].handle = get_sub_data(args->dataA, 2, u22a->k, u22a->j);
+				task22->buffers[1].handle = starpu_get_sub_data(args->dataA, 2, u22a->k, u22a->j);
 				task22->buffers[1].mode = STARPU_R;
 
-				task22->buffers[2].handle = get_sub_data(args->dataA, 2, u22a->i, u22a->j);
+				task22->buffers[2].handle = starpu_get_sub_data(args->dataA, 2, u22a->i, u22a->j);
 				task22->buffers[2].mode = STARPU_RW;
 				
 				/* schedule that codelet */
@@ -605,7 +605,7 @@ void dw_codelet_facto(starpu_data_handle dataA, unsigned nblocks)
 		task->cl = &cl11;
 		task->cl_arg = args;
 
-		task->buffers[0].handle = get_sub_data(dataA, 2, 0, 0);
+		task->buffers[0].handle = starpu_get_sub_data(dataA, 2, 0, 0);
 		task->buffers[0].mode = STARPU_RW;
 
 	/* schedule the codelet */
@@ -657,7 +657,7 @@ void dw_codelet_facto_v2(starpu_data_handle dataA, unsigned nblocks)
 		task->cl = &cl11;
 		task->cl_arg = args;
 
-		task->buffers[0].handle = get_sub_data(dataA, 2, 0, 0); 
+		task->buffers[0].handle = starpu_get_sub_data(dataA, 2, 0, 0); 
 		task->buffers[0].mode = STARPU_RW;
 
 	/* schedule the codelet */

examples/heat/dw_factolu_grain.c

@@ -41,7 +41,7 @@ static struct starpu_task *create_task(starpu_tag_t id)
 }
 
 static starpu_codelet cl11 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dw_core_codelet_update_u11,
 #ifdef USE_CUDA
 	.cuda_func = dw_cublas_codelet_update_u11,
@@ -59,11 +59,11 @@ static struct starpu_task *create_task_11(starpu_data_handle dataA, unsigned k,
 	task->cl = &cl11;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k);
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k);
 	task->buffers[0].mode = STARPU_RW;
 
 	/* this is an important task */
-	task->priority = MAX_PRIO;
+	task->priority = STARPU_MAX_PRIO;
 
 	/* enforce dependencies ... */
 	if (k > 0) {
@@ -74,7 +74,7 @@ static struct starpu_task *create_task_11(starpu_data_handle dataA, unsigned k,
 }
 
 static starpu_codelet cl12 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dw_core_codelet_update_u12,
 #ifdef USE_CUDA
 	.cuda_func = dw_cublas_codelet_update_u12,
@@ -92,13 +92,13 @@ static void create_task_12(starpu_data_handle dataA, unsigned k, unsigned i, uns
 	task->cl = &cl12;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k); 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k); 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, i, k); 
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, i, k); 
 	task->buffers[1].mode = STARPU_RW;
 
 	if (i == k+1) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */
@@ -113,7 +113,7 @@ static void create_task_12(starpu_data_handle dataA, unsigned k, unsigned i, uns
 }
 
 static starpu_codelet cl21 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dw_core_codelet_update_u21,
 #ifdef USE_CUDA
 	.cuda_func = dw_cublas_codelet_update_u21,
@@ -129,13 +129,13 @@ static void create_task_21(starpu_data_handle dataA, unsigned k, unsigned j, uns
 	task->cl = &cl21;
 	
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k); 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k); 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, k, j); 
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, k, j); 
 	task->buffers[1].mode = STARPU_RW;
 
 	if (j == k+1) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */
@@ -150,7 +150,7 @@ static void create_task_21(starpu_data_handle dataA, unsigned k, unsigned j, uns
 }
 
 static starpu_codelet cl22 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dw_core_codelet_update_u22,
 #ifdef USE_CUDA
 	.cuda_func = dw_cublas_codelet_update_u22,
@@ -168,15 +168,15 @@ static void create_task_22(starpu_data_handle dataA, unsigned k, unsigned i, uns
 	task->cl = &cl22;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, i, k); 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, i, k); 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, k, j); 
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, k, j); 
 	task->buffers[1].mode = STARPU_R;
-	task->buffers[2].handle = get_sub_data(dataA, 2, i, j); 
+	task->buffers[2].handle = starpu_get_sub_data(dataA, 2, i, j); 
 	task->buffers[2].mode = STARPU_RW;
 
 	if ( (i == k + 1) && (j == k +1) ) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */

examples/heat/dw_factolu_kernels.c

@@ -104,17 +104,17 @@ void display_stat_heat(void)
 
 static inline void dw_common_core_codelet_update_u22(void *descr[], int s, __attribute__((unused)) void *_args)
 {
-	float *left 	= (float *)GET_BLAS_PTR(descr[0]);
-	float *right 	= (float *)GET_BLAS_PTR(descr[1]);
-	float *center 	= (float *)GET_BLAS_PTR(descr[2]);
+	float *left 	= (float *)STARPU_GET_BLAS_PTR(descr[0]);
+	float *right 	= (float *)STARPU_GET_BLAS_PTR(descr[1]);
+	float *center 	= (float *)STARPU_GET_BLAS_PTR(descr[2]);
 
-	unsigned dx = GET_BLAS_NX(descr[2]);
-	unsigned dy = GET_BLAS_NY(descr[2]);
-	unsigned dz = GET_BLAS_NY(descr[0]);
+	unsigned dx = STARPU_GET_BLAS_NX(descr[2]);
+	unsigned dy = STARPU_GET_BLAS_NY(descr[2]);
+	unsigned dz = STARPU_GET_BLAS_NY(descr[0]);
 
-	unsigned ld12 = GET_BLAS_LD(descr[0]);
-	unsigned ld21 = GET_BLAS_LD(descr[1]);
-	unsigned ld22 = GET_BLAS_LD(descr[2]);
+	unsigned ld12 = STARPU_GET_BLAS_LD(descr[0]);
+	unsigned ld21 = STARPU_GET_BLAS_LD(descr[1]);
+	unsigned ld22 = STARPU_GET_BLAS_LD(descr[2]);
 
 #ifdef USE_CUDA
 	cublasStatus status;
@@ -171,14 +171,14 @@ static inline void dw_common_codelet_update_u12(void *descr[], int s, __attribut
 	float *sub11;
 	float *sub12;
 
-	sub11 = (float *)GET_BLAS_PTR(descr[0]);	
-	sub12 = (float *)GET_BLAS_PTR(descr[1]);
+	sub11 = (float *)STARPU_GET_BLAS_PTR(descr[0]);	
+	sub12 = (float *)STARPU_GET_BLAS_PTR(descr[1]);
 
-	unsigned ld11 = GET_BLAS_LD(descr[0]);
-	unsigned ld12 = GET_BLAS_LD(descr[1]);
+	unsigned ld11 = STARPU_GET_BLAS_LD(descr[0]);
+	unsigned ld12 = STARPU_GET_BLAS_LD(descr[1]);
 
-	unsigned nx12 = GET_BLAS_NX(descr[1]);
-	unsigned ny12 = GET_BLAS_NY(descr[1]);
+	unsigned nx12 = STARPU_GET_BLAS_NX(descr[1]);
+	unsigned ny12 = STARPU_GET_BLAS_NY(descr[1]);
 	
 #ifdef USE_CUDA
 	cublasStatus status;
@@ -234,14 +234,14 @@ static inline void dw_common_codelet_update_u21(void *descr[], int s, __attribut
 	float *sub11;
 	float *sub21;
 
-	sub11 = (float *)GET_BLAS_PTR(descr[0]);
-	sub21 = (float *)GET_BLAS_PTR(descr[1]);
+	sub11 = (float *)STARPU_GET_BLAS_PTR(descr[0]);
+	sub21 = (float *)STARPU_GET_BLAS_PTR(descr[1]);
 
-	unsigned ld11 = GET_BLAS_LD(descr[0]);
-	unsigned ld21 = GET_BLAS_LD(descr[1]);
+	unsigned ld11 = STARPU_GET_BLAS_LD(descr[0]);
+	unsigned ld21 = STARPU_GET_BLAS_LD(descr[1]);
 
-	unsigned nx21 = GET_BLAS_NX(descr[1]);
-	unsigned ny21 = GET_BLAS_NY(descr[1]);
+	unsigned nx21 = STARPU_GET_BLAS_NX(descr[1]);
+	unsigned ny21 = STARPU_GET_BLAS_NY(descr[1]);
 	
 #ifdef USE_CUDA
 	cublasStatus status;
@@ -309,10 +309,10 @@ static inline void dw_common_codelet_update_u11(void *descr[], int s, __attribut
 {
 	float *sub11;
 
-	sub11 = (float *)GET_BLAS_PTR(descr[0]); 
+	sub11 = (float *)STARPU_GET_BLAS_PTR(descr[0]); 
 
-	unsigned long nx = GET_BLAS_NX(descr[0]);
-	unsigned long ld = GET_BLAS_LD(descr[0]);
+	unsigned long nx = STARPU_GET_BLAS_NX(descr[0]);
+	unsigned long ld = STARPU_GET_BLAS_LD(descr[0]);
 
 	unsigned long z;
 

examples/heat/dw_factolu_tag.c

@@ -43,7 +43,7 @@ static struct starpu_task *create_task(starpu_tag_t id)
 }
 
 static starpu_codelet cl11 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dw_core_codelet_update_u11,
 #ifdef USE_CUDA
 	.cuda_func = dw_cublas_codelet_update_u11,
@@ -61,12 +61,12 @@ static struct starpu_task *create_task_11(starpu_data_handle dataA, unsigned k)
 	task->cl = &cl11;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k);
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k);
 	task->buffers[0].mode = STARPU_RW;
 
 	/* this is an important task */
 	if (!no_prio)
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 
 	/* enforce dependencies ... */
 	if (k > 0) {
@@ -77,7 +77,7 @@ static struct starpu_task *create_task_11(starpu_data_handle dataA, unsigned k)
 }
 
 static starpu_codelet cl12 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dw_core_codelet_update_u12,
 #ifdef USE_CUDA
 	.cuda_func = dw_cublas_codelet_update_u12,
@@ -95,13 +95,13 @@ static void create_task_12(starpu_data_handle dataA, unsigned k, unsigned i)
 	task->cl = &cl12;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k); 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k); 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, i, k); 
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, i, k); 
 	task->buffers[1].mode = STARPU_RW;
 
 	if (!no_prio && (i == k+1)) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */
@@ -116,7 +116,7 @@ static void create_task_12(starpu_data_handle dataA, unsigned k, unsigned i)
 }
 
 static starpu_codelet cl21 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dw_core_codelet_update_u21,
 #ifdef USE_CUDA
 	.cuda_func = dw_cublas_codelet_update_u21,
@@ -132,13 +132,13 @@ static void create_task_21(starpu_data_handle dataA, unsigned k, unsigned j)
 	task->cl = &cl21;
 	
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k); 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k); 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, k, j); 
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, k, j); 
 	task->buffers[1].mode = STARPU_RW;
 
 	if (!no_prio && (j == k+1)) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */
@@ -153,7 +153,7 @@ static void create_task_21(starpu_data_handle dataA, unsigned k, unsigned j)
 }
 
 static starpu_codelet cl22 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dw_core_codelet_update_u22,
 #ifdef USE_CUDA
 	.cuda_func = dw_cublas_codelet_update_u22,
@@ -171,15 +171,15 @@ static void create_task_22(starpu_data_handle dataA, unsigned k, unsigned i, uns
 	task->cl = &cl22;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, i, k); 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, i, k); 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, k, j); 
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, k, j); 
 	task->buffers[1].mode = STARPU_R;
-	task->buffers[2].handle = get_sub_data(dataA, 2, i, j); 
+	task->buffers[2].handle = starpu_get_sub_data(dataA, 2, i, j); 
 	task->buffers[2].mode = STARPU_RW;
 
 	if (!no_prio &&  (i == k + 1) && (j == k +1) ) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */

examples/heat/dw_sparse_cg.c

@@ -133,7 +133,7 @@ void init_cg(struct cg_problem *problem)
 
 	/* r = b  - A x */
 	struct starpu_task *task1 = create_task(1UL);
-	task1->cl->where = CORE;
+	task1->cl->where = STARPU_CORE;
 	task1->cl->core_func = core_codelet_func_1;
 	task1->cl->nbuffers = 4;
 		task1->buffers[0].handle = problem->ds_matrixA;
@@ -147,7 +147,7 @@ void init_cg(struct cg_problem *problem)
 
 	/* d = r */
 	struct starpu_task *task2 = create_task(2UL);
-	task2->cl->where = CORE;
+	task2->cl->where = STARPU_CORE;
 	task2->cl->core_func = core_codelet_func_2;
 	task2->cl->nbuffers = 2;
 		task2->buffers[0].handle = problem->ds_vecd;
@@ -159,7 +159,7 @@ void init_cg(struct cg_problem *problem)
 
 	/* delta_new = trans(r) r */
 	struct starpu_task *task3 = create_task(3UL);
-	task3->cl->where = CUDA|CORE;
+	task3->cl->where = STARPU_CUDA|STARPU_CORE;
 #ifdef USE_CUDA
 	task3->cl->cuda_func = cublas_codelet_func_3;
 #endif
@@ -194,7 +194,7 @@ void launch_new_cg_iteration(struct cg_problem *problem)
 
 	/* q = A d */
 	struct starpu_task *task4 = create_task(maskiter | 4UL);
-	task4->cl->where = CORE;
+	task4->cl->where = STARPU_CORE;
 	task4->cl->core_func = core_codelet_func_4;
 	task4->cl->nbuffers = 3;
 		task4->buffers[0].handle = problem->ds_matrixA;
@@ -206,7 +206,7 @@ void launch_new_cg_iteration(struct cg_problem *problem)
 
 	/* alpha = delta_new / ( trans(d) q )*/
 	struct starpu_task *task5 = create_task(maskiter | 5UL);
-	task5->cl->where = CUDA|CORE;
+	task5->cl->where = STARPU_CUDA|STARPU_CORE;
 #ifdef USE_CUDA
 	task5->cl->cuda_func = cublas_codelet_func_5;
 #endif
@@ -222,7 +222,7 @@ void launch_new_cg_iteration(struct cg_problem *problem)
 
 	/* x = x + alpha d */
 	struct starpu_task *task6 = create_task(maskiter | 6UL);
-	task6->cl->where = CUDA|CORE;
+	task6->cl->where = STARPU_CUDA|STARPU_CORE;
 #ifdef USE_CUDA
 	task6->cl->cuda_func = cublas_codelet_func_6;
 #endif
@@ -238,7 +238,7 @@ void launch_new_cg_iteration(struct cg_problem *problem)
 
 	/* r = r - alpha q */
 	struct starpu_task *task7 = create_task(maskiter | 7UL);
-	task7->cl->where = CUDA|CORE;
+	task7->cl->where = STARPU_CUDA|STARPU_CORE;
 #ifdef USE_CUDA
 	task7->cl->cuda_func = cublas_codelet_func_7;
 #endif
@@ -254,7 +254,7 @@ void launch_new_cg_iteration(struct cg_problem *problem)
 
 	/* update delta_* and compute beta */
 	struct starpu_task *task8 = create_task(maskiter | 8UL);
-	task8->cl->where = CUDA|CORE;
+	task8->cl->where = STARPU_CUDA|STARPU_CORE;
 #ifdef USE_CUDA
 	task8->cl->cuda_func = cublas_codelet_func_8;
 #endif
@@ -268,7 +268,7 @@ void launch_new_cg_iteration(struct cg_problem *problem)
 
 	/* d = r + beta d */
 	struct starpu_task *task9 = create_task(maskiter | 9UL);
-	task9->cl->where = CUDA|CORE;
+	task9->cl->where = STARPU_CUDA|STARPU_CORE;
 #ifdef USE_CUDA
 	task9->cl->cuda_func = cublas_codelet_func_9;
 #endif

examples/heat/lu_kernels_model.c

@@ -218,7 +218,7 @@ struct starpu_perfmodel_t model_11 = {
 		[STARPU_CORE_DEFAULT] = { .cost_model = task_11_cost_core },
 		[STARPU_CUDA_DEFAULT] = { .cost_model = task_11_cost_cuda }
 	},
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = "lu_model_11_atlas"
 #elif defined(GOTO)
@@ -234,7 +234,7 @@ struct starpu_perfmodel_t model_12 = {
 		[STARPU_CORE_DEFAULT] = { .cost_model = task_12_cost_core },
 		[STARPU_CUDA_DEFAULT] = { .cost_model = task_12_cost_cuda }
 	},
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = "lu_model_12_atlas"
 #elif defined(GOTO)
@@ -250,7 +250,7 @@ struct starpu_perfmodel_t model_21 = {
 		[STARPU_CORE_DEFAULT] = { .cost_model = task_21_cost_core },
 		[STARPU_CUDA_DEFAULT] = { .cost_model = task_21_cost_cuda }
 	},
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = "lu_model_21_atlas"
 #elif defined(GOTO)
@@ -266,7 +266,7 @@ struct starpu_perfmodel_t model_22 = {
 		[STARPU_CORE_DEFAULT] = { .cost_model = task_22_cost_core },
 		[STARPU_CUDA_DEFAULT] = { .cost_model = task_22_cost_cuda }
 	},
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = "lu_model_22_atlas"
 #elif defined(GOTO)

examples/incrementer/incrementer.c

@@ -48,7 +48,7 @@ int main(int argc, char **argv)
 	starpu_codelet cl =
 	{
 		/* CUBLAS stands for CUDA kernels controlled from the host */
-		.where = CORE|CUDA,
+		.where = STARPU_CORE|STARPU_CUDA,
 		.core_func = core_codelet,
 #ifdef USE_CUDA
 		.cuda_func = cuda_codelet,

examples/lu/xlu.c

@@ -47,7 +47,7 @@ static struct starpu_task *create_task(starpu_tag_t id)
 }
 
 static struct starpu_perfmodel_t STARPU_LU(model_11) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = STARPU_LU_STR(lu_model_11_atlas)
 #elif defined(GOTO)
@@ -58,7 +58,7 @@ static struct starpu_perfmodel_t STARPU_LU(model_11) = {
 };
 
 static starpu_codelet cl11 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = STARPU_LU(cpu_u11),
 #ifdef USE_CUDA
 	.cuda_func = STARPU_LU(cublas_u11),
@@ -76,12 +76,12 @@ static struct starpu_task *create_task_11(starpu_data_handle dataA, unsigned k)
 	task->cl = &cl11;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k);
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k);
 	task->buffers[0].mode = STARPU_RW;
 
 	/* this is an important task */
 	if (!no_prio)
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 
 	/* enforce dependencies ... */
 	if (k > 0) {
@@ -92,7 +92,7 @@ static struct starpu_task *create_task_11(starpu_data_handle dataA, unsigned k)
 }
 
 static struct starpu_perfmodel_t STARPU_LU(model_12) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = STARPU_LU_STR(lu_model_12_atlas)
 #elif defined(GOTO)
@@ -103,7 +103,7 @@ static struct starpu_perfmodel_t STARPU_LU(model_12) = {
 };
 
 static starpu_codelet cl12 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = STARPU_LU(cpu_u12),
 #ifdef USE_CUDA
 	.cuda_func = STARPU_LU(cublas_u12),
@@ -121,13 +121,13 @@ static void create_task_12(starpu_data_handle dataA, unsigned k, unsigned j)
 	task->cl = &cl12;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k); 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k); 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, j, k); 
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, j, k); 
 	task->buffers[1].mode = STARPU_RW;
 
 	if (!no_prio && (j == k+1)) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */
@@ -142,7 +142,7 @@ static void create_task_12(starpu_data_handle dataA, unsigned k, unsigned j)
 }
 
 static struct starpu_perfmodel_t STARPU_LU(model_21) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = STARPU_LU_STR(lu_model_21_atlas)
 #elif defined(GOTO)
@@ -153,7 +153,7 @@ static struct starpu_perfmodel_t STARPU_LU(model_21) = {
 };
 
 static starpu_codelet cl21 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = STARPU_LU(cpu_u21),
 #ifdef USE_CUDA
 	.cuda_func = STARPU_LU(cublas_u21),
@@ -169,13 +169,13 @@ static void create_task_21(starpu_data_handle dataA, unsigned k, unsigned i)
 	task->cl = &cl21;
 	
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, k); 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, k); 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, k, i); 
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, k, i); 
 	task->buffers[1].mode = STARPU_RW;
 
 	if (!no_prio && (i == k+1)) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */
@@ -190,7 +190,7 @@ static void create_task_21(starpu_data_handle dataA, unsigned k, unsigned i)
 }
 
 static struct starpu_perfmodel_t STARPU_LU(model_22) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = STARPU_LU_STR(lu_model_22_atlas)
 #elif defined(GOTO)
@@ -201,7 +201,7 @@ static struct starpu_perfmodel_t STARPU_LU(model_22) = {
 };
 
 static starpu_codelet cl22 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = STARPU_LU(cpu_u22),
 #ifdef USE_CUDA
 	.cuda_func = STARPU_LU(cublas_u22),
@@ -219,15 +219,15 @@ static void create_task_22(starpu_data_handle dataA, unsigned k, unsigned i, uns
 	task->cl = &cl22;
 
 	/* which sub-data is manipulated ? */
-	task->buffers[0].handle = get_sub_data(dataA, 2, k, i); /* produced by TAG21(k, i) */ 
+	task->buffers[0].handle = starpu_get_sub_data(dataA, 2, k, i); /* produced by TAG21(k, i) */ 
 	task->buffers[0].mode = STARPU_R;
-	task->buffers[1].handle = get_sub_data(dataA, 2, j, k); /* produced by TAG12(k, j) */
+	task->buffers[1].handle = starpu_get_sub_data(dataA, 2, j, k); /* produced by TAG12(k, j) */
 	task->buffers[1].mode = STARPU_R;
-	task->buffers[2].handle = get_sub_data(dataA, 2, j, i); /* produced by TAG22(k-1, i, j) */
+	task->buffers[2].handle = starpu_get_sub_data(dataA, 2, j, i); /* produced by TAG22(k-1, i, j) */
 	task->buffers[2].mode = STARPU_RW;
 
 	if (!no_prio &&  (i == k + 1) && (j == k +1) ) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */

examples/lu/xlu_kernels.c

@@ -24,17 +24,17 @@
 static inline void STARPU_LU(common_u22)(void *descr[],
 				int s, __attribute__((unused)) void *_args)
 {
-	TYPE *right 	= (TYPE *)GET_BLAS_PTR(descr[0]);
-	TYPE *left 	= (TYPE *)GET_BLAS_PTR(descr[1]);
-	TYPE *center 	= (TYPE *)GET_BLAS_PTR(descr[2]);
+	TYPE *right 	= (TYPE *)STARPU_GET_BLAS_PTR(descr[0]);
+	TYPE *left 	= (TYPE *)STARPU_GET_BLAS_PTR(descr[1]);
+	TYPE *center 	= (TYPE *)STARPU_GET_BLAS_PTR(descr[2]);
 
-	unsigned dx = GET_BLAS_NX(descr[2]);
-	unsigned dy = GET_BLAS_NY(descr[2]);
-	unsigned dz = GET_BLAS_NY(descr[0]);
+	unsigned dx = STARPU_GET_BLAS_NX(descr[2]);
+	unsigned dy = STARPU_GET_BLAS_NY(descr[2]);
+	unsigned dz = STARPU_GET_BLAS_NY(descr[0]);
 
-	unsigned ld12 = GET_BLAS_LD(descr[0]);
-	unsigned ld21 = GET_BLAS_LD(descr[1]);
-	unsigned ld22 = GET_BLAS_LD(descr[2]);
+	unsigned ld12 = STARPU_GET_BLAS_LD(descr[0]);
+	unsigned ld21 = STARPU_GET_BLAS_LD(descr[1]);
+	unsigned ld22 = STARPU_GET_BLAS_LD(descr[2]);
 
 #ifdef USE_CUDA
 	cublasStatus status;
@@ -91,14 +91,14 @@ static inline void STARPU_LU(common_u12)(void *descr[],
 	TYPE *sub11;
 	TYPE *sub12;
 
-	sub11 = (TYPE *)GET_BLAS_PTR(descr[0]);	
-	sub12 = (TYPE *)GET_BLAS_PTR(descr[1]);
+	sub11 = (TYPE *)STARPU_GET_BLAS_PTR(descr[0]);	
+	sub12 = (TYPE *)STARPU_GET_BLAS_PTR(descr[1]);
 
-	unsigned ld11 = GET_BLAS_LD(descr[0]);
-	unsigned ld12 = GET_BLAS_LD(descr[1]);
+	unsigned ld11 = STARPU_GET_BLAS_LD(descr[0]);
+	unsigned ld12 = STARPU_GET_BLAS_LD(descr[1]);
 
-	unsigned nx12 = GET_BLAS_NX(descr[1]);
-	unsigned ny12 = GET_BLAS_NY(descr[1]);
+	unsigned nx12 = STARPU_GET_BLAS_NX(descr[1]);
+	unsigned ny12 = STARPU_GET_BLAS_NY(descr[1]);
 
 #ifdef USE_CUDA
 	cublasStatus status;
@@ -153,14 +153,14 @@ static inline void STARPU_LU(common_u21)(void *descr[],
 	TYPE *sub11;
 	TYPE *sub21;
 
-	sub11 = (TYPE *)GET_BLAS_PTR(descr[0]);
-	sub21 = (TYPE *)GET_BLAS_PTR(descr[1]);
+	sub11 = (TYPE *)STARPU_GET_BLAS_PTR(descr[0]);
+	sub21 = (TYPE *)STARPU_GET_BLAS_PTR(descr[1]);
 
-	unsigned ld11 = GET_BLAS_LD(descr[0]);
-	unsigned ld21 = GET_BLAS_LD(descr[1]);
+	unsigned ld11 = STARPU_GET_BLAS_LD(descr[0]);
+	unsigned ld21 = STARPU_GET_BLAS_LD(descr[1]);
 
-	unsigned nx21 = GET_BLAS_NX(descr[1]);
-	unsigned ny21 = GET_BLAS_NY(descr[1]);
+	unsigned nx21 = STARPU_GET_BLAS_NX(descr[1]);
+	unsigned ny21 = STARPU_GET_BLAS_NY(descr[1]);
 	
 #ifdef USE_CUDA
 	cublasStatus status;
@@ -212,10 +212,10 @@ static inline void STARPU_LU(common_u11)(void *descr[],
 {
 	TYPE *sub11;
 
-	sub11 = (TYPE *)GET_BLAS_PTR(descr[0]); 
+	sub11 = (TYPE *)STARPU_GET_BLAS_PTR(descr[0]); 
 
-	unsigned long nx = GET_BLAS_NX(descr[0]);
-	unsigned long ld = GET_BLAS_LD(descr[0]);
+	unsigned long nx = STARPU_GET_BLAS_NX(descr[0]);
+	unsigned long ld = STARPU_GET_BLAS_LD(descr[0]);
 
 	unsigned long z;
 
@@ -284,10 +284,10 @@ static inline void STARPU_LU(common_u11_pivot)(void *descr[],
 {
 	TYPE *sub11;
 
-	sub11 = (TYPE *)GET_BLAS_PTR(descr[0]); 
+	sub11 = (TYPE *)STARPU_GET_BLAS_PTR(descr[0]); 
 
-	unsigned long nx = GET_BLAS_NX(descr[0]);
-	unsigned long ld = GET_BLAS_LD(descr[0]);
+	unsigned long nx = STARPU_GET_BLAS_NX(descr[0]);
+	unsigned long ld = STARPU_GET_BLAS_LD(descr[0]);
 
 	unsigned long z;
 
@@ -399,9 +399,9 @@ static inline void STARPU_LU(common_pivot)(void *descr[],
 {
 	TYPE *matrix;
 
-	matrix = (TYPE *)GET_BLAS_PTR(descr[0]); 
-	unsigned long nx = GET_BLAS_NX(descr[0]);
-	unsigned long ld = GET_BLAS_LD(descr[0]);
+	matrix = (TYPE *)STARPU_GET_BLAS_PTR(descr[0]); 
+	unsigned long nx = STARPU_GET_BLAS_NX(descr[0]);
+	unsigned long ld = STARPU_GET_BLAS_LD(descr[0]);
 
 	unsigned row, rowaux;
 

examples/lu/xlu_pivot.c

@@ -50,7 +50,7 @@ static struct starpu_task *create_task(starpu_tag_t id)
 
 
 static struct starpu_perfmodel_t STARPU_LU(model_pivot) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = STARPU_LU_STR(lu_model_pivot_atlas)
 #elif defined(GOTO)
@@ -61,7 +61,7 @@ static struct starpu_perfmodel_t STARPU_LU(model_pivot) = {
 };
 
 static starpu_codelet cl_pivot = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = STARPU_LU(cpu_pivot),
 #ifdef USE_CUDA
 	.cuda_func = STARPU_LU(cublas_pivot),
@@ -87,7 +87,7 @@ static void create_task_pivot(starpu_data_handle *dataAp, unsigned nblocks,
 
 	/* this is an important task */
 	if (!no_prio && (i == k+1))
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 
 	/* enforce dependencies ... */
 	if (k == 0) {
@@ -117,7 +117,7 @@ static void create_task_pivot(starpu_data_handle *dataAp, unsigned nblocks,
 }
 
 static struct starpu_perfmodel_t STARPU_LU(model_11_pivot) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = STARPU_LU_STR(lu_model_11_pivot_atlas)
 #elif defined(GOTO)
@@ -128,7 +128,7 @@ static struct starpu_perfmodel_t STARPU_LU(model_11_pivot) = {
 };
 
 static starpu_codelet cl11_pivot = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = STARPU_LU(cpu_u11_pivot),
 #ifdef USE_CUDA
 	.cuda_func = STARPU_LU(cublas_u11_pivot),
@@ -153,7 +153,7 @@ static struct starpu_task *create_task_11_pivot(starpu_data_handle *dataAp, unsi
 
 	/* this is an important task */
 	if (!no_prio)
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 
 	/* enforce dependencies ... */
 	if (k > 0) {
@@ -164,7 +164,7 @@ static struct starpu_task *create_task_11_pivot(starpu_data_handle *dataAp, unsi
 }
 
 static struct starpu_perfmodel_t STARPU_LU(model_12) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = STARPU_LU_STR(lu_model_12_atlas)
 #elif defined(GOTO)
@@ -175,7 +175,7 @@ static struct starpu_perfmodel_t STARPU_LU(model_12) = {
 };
 
 static starpu_codelet cl12 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = STARPU_LU(cpu_u12),
 #ifdef USE_CUDA
 	.cuda_func = STARPU_LU(cublas_u12),
@@ -202,7 +202,7 @@ static void create_task_12(starpu_data_handle *dataAp, unsigned nblocks, unsigne
 	task->buffers[1].mode = STARPU_RW;
 
 	if (!no_prio && (j == k+1)) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */
@@ -220,7 +220,7 @@ static void create_task_12(starpu_data_handle *dataAp, unsigned nblocks, unsigne
 }
 
 static struct starpu_perfmodel_t STARPU_LU(model_21) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = STARPU_LU_STR(lu_model_21_atlas)
 #elif defined(GOTO)
@@ -231,7 +231,7 @@ static struct starpu_perfmodel_t STARPU_LU(model_21) = {
 };
 
 static starpu_codelet cl21 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = STARPU_LU(cpu_u21),
 #ifdef USE_CUDA
 	.cuda_func = STARPU_LU(cublas_u21),
@@ -254,7 +254,7 @@ static void create_task_21(starpu_data_handle *dataAp, unsigned nblocks, unsigne
 	task->buffers[1].mode = STARPU_RW;
 
 	if (!no_prio && (i == k+1)) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	task->cl_arg = (void *)(task->tag_id);
@@ -274,7 +274,7 @@ static void create_task_21(starpu_data_handle *dataAp, unsigned nblocks, unsigne
 }
 
 static struct starpu_perfmodel_t STARPU_LU(model_22) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = STARPU_LU_STR(lu_model_22_atlas)
 #elif defined(GOTO)
@@ -285,7 +285,7 @@ static struct starpu_perfmodel_t STARPU_LU(model_22) = {
 };
 
 static starpu_codelet cl22 = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = STARPU_LU(cpu_u22),
 #ifdef USE_CUDA
 	.cuda_func = STARPU_LU(cublas_u22),
@@ -314,7 +314,7 @@ static void create_task_22(starpu_data_handle *dataAp, unsigned nblocks, unsigne
 	task->buffers[2].mode = STARPU_RW;
 
 	if (!no_prio &&  (i == k + 1) && (j == k +1) ) {
-		task->priority = MAX_PRIO;
+		task->priority = STARPU_MAX_PRIO;
 	}
 
 	/* enforce dependencies ... */
@@ -415,7 +415,7 @@ starpu_data_handle get_block_with_striding(starpu_data_handle *dataAp,
 			unsigned nblocks __attribute__((unused)), unsigned j, unsigned i)
 {
 	/* we use filters */
-	return get_sub_data(*dataAp, 2, j, i);
+	return starpu_get_sub_data(*dataAp, 2, j, i);
 }
 
 

examples/mult/dw_mult_no_stride.c

@@ -236,7 +236,7 @@ static starpu_codelet cl = {
 #endif
 
 	.model = &sgemm_model,
-	.where = CORE|CUDA|GORDON,
+	.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON,
 	.nbuffers = 3
 };
 

examples/mult/dw_mult_no_stride_no_tag.c

@@ -280,7 +280,7 @@ struct cb2_s {
 
 
 static starpu_codelet cl = {
-	.where = CORE|CUDA|GORDON,
+	.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON,
 	.core_func = core_mult,
 #ifdef USE_CUDA
 	.cuda_func = cublas_mult,

examples/mult/sgemm_kernels.c

@@ -25,17 +25,17 @@
 	float *subB;			\
 	float *subC;			\
 					\
-	subA = (float *)GET_BLAS_PTR(descr[0]);	\
-	subB = (float *)GET_BLAS_PTR(descr[1]);	\
-	subC = (float *)GET_BLAS_PTR(descr[2]);	\
+	subA = (float *)STARPU_GET_BLAS_PTR(descr[0]);	\
+	subB = (float *)STARPU_GET_BLAS_PTR(descr[1]);	\
+	subC = (float *)STARPU_GET_BLAS_PTR(descr[2]);	\
 					\
-	nxC = GET_BLAS_NX(descr[2]);		\
-	nyC = GET_BLAS_NY(descr[2]);		\
-	nyA = GET_BLAS_NY(descr[0]);		\
+	nxC = STARPU_GET_BLAS_NX(descr[2]);		\
+	nyC = STARPU_GET_BLAS_NY(descr[2]);		\
+	nyA = STARPU_GET_BLAS_NY(descr[0]);		\
 					\
-	ldA = GET_BLAS_LD(descr[0]);		\
-	ldB = GET_BLAS_LD(descr[1]);		\
-	ldC = GET_BLAS_LD(descr[2]);
+	ldA = STARPU_GET_BLAS_LD(descr[0]);		\
+	ldB = STARPU_GET_BLAS_LD(descr[1]);		\
+	ldC = STARPU_GET_BLAS_LD(descr[2]);
 
 
 

examples/mult/xgemm.c

@@ -172,7 +172,7 @@ static void unpartition_mult_data(void)
 }
 
 static struct starpu_perfmodel_t gemm_model = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 #ifdef ATLAS
 	.symbol = STARPU_GEMM_STR(gemm_atlas)
 #elif defined(GOTO)
@@ -183,7 +183,7 @@ static struct starpu_perfmodel_t gemm_model = {
 };
 
 static starpu_codelet cl = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = STARPU_GEMM(core_mult),
 #ifdef USE_CUDA
 	.cuda_func = STARPU_GEMM(cublas_mult),
@@ -212,12 +212,11 @@ static void launch_codelets(void)
 			task->callback_func = callback_func;
 			task->callback_arg = NULL;
 
-			task->buffers[0].handle = get_sub_data(A_handle, 1, tasky);
+			task->buffers[0].handle = starpu_get_sub_data(A_handle, 1, tasky);
 			task->buffers[0].mode = STARPU_R;
-			task->buffers[1].handle = get_sub_data(B_handle, 1, taskx);
+			task->buffers[1].handle = starpu_get_sub_data(B_handle, 1, taskx);
 			task->buffers[1].mode = STARPU_R;
-			task->buffers[2].handle = 
-				get_sub_data(C_handle, 2, taskx, tasky);
+			task->buffers[2].handle = starpu_get_sub_data(C_handle, 2, taskx, tasky);
 			task->buffers[2].mode = STARPU_RW;
 
 			starpu_submit_task(task);

examples/mult/xgemm_kernels.c

@@ -25,17 +25,17 @@
 	TYPE *subB;			\
 	TYPE *subC;			\
 					\
-	subA = (TYPE *)GET_BLAS_PTR(descr[0]);	\
-	subB = (TYPE *)GET_BLAS_PTR(descr[1]);	\
-	subC = (TYPE *)GET_BLAS_PTR(descr[2]);	\
+	subA = (TYPE *)STARPU_GET_BLAS_PTR(descr[0]);	\
+	subB = (TYPE *)STARPU_GET_BLAS_PTR(descr[1]);	\
+	subC = (TYPE *)STARPU_GET_BLAS_PTR(descr[2]);	\
 					\
-	nxC = GET_BLAS_NX(descr[2]);		\
-	nyC = GET_BLAS_NY(descr[2]);		\
-	nyA = GET_BLAS_NY(descr[0]);		\
+	nxC = STARPU_GET_BLAS_NX(descr[2]);		\
+	nyC = STARPU_GET_BLAS_NY(descr[2]);		\
+	nyA = STARPU_GET_BLAS_NY(descr[0]);		\
 					\
-	ldA = GET_BLAS_LD(descr[0]);		\
-	ldB = GET_BLAS_LD(descr[1]);		\
-	ldC = GET_BLAS_LD(descr[2]);
+	ldA = STARPU_GET_BLAS_LD(descr[0]);		\
+	ldB = STARPU_GET_BLAS_LD(descr[1]);		\
+	ldC = STARPU_GET_BLAS_LD(descr[2]);
 
 
 

examples/pastix-wrappers/starpu-blas-wrapper.c

@@ -40,7 +40,7 @@
 
 #include "starpu-blas-wrapper.h"
 
-extern struct data_interface_ops_t interface_blas_ops;
+extern struct starpu_data_interface_ops_t interface_blas_ops;
 
 static int core_sgemm = 0;
 static int cublas_sgemm = 0;

examples/ppm-downscaler/yuv-downscaler.c

@@ -68,14 +68,14 @@ static void ds_callback(void *arg)
 
 static void ds_kernel_cpu(void *descr[], __attribute__((unused)) void *arg)
 {
-	uint8_t *input = (uint8_t *)GET_BLAS_PTR(descr[0]);
-	unsigned input_ld = GET_BLAS_LD(descr[0]);
+	uint8_t *input = (uint8_t *)STARPU_GET_BLAS_PTR(descr[0]);
+	unsigned input_ld = STARPU_GET_BLAS_LD(descr[0]);
 
-	uint8_t *output = (uint8_t *)GET_BLAS_PTR(descr[1]);
-	unsigned output_ld = GET_BLAS_LD(descr[1]);
+	uint8_t *output = (uint8_t *)STARPU_GET_BLAS_PTR(descr[1]);
+	unsigned output_ld = STARPU_GET_BLAS_LD(descr[1]);
 
-	unsigned ncols = GET_BLAS_NX(descr[0]);
-	unsigned nlines = GET_BLAS_NY(descr[0]);
+	unsigned ncols = STARPU_GET_BLAS_NX(descr[0]);
+	unsigned nlines = STARPU_GET_BLAS_NY(descr[0]);
 
 	unsigned line, col;
 	for (line = 0; line < nlines; line+=FACTOR)
@@ -98,7 +98,7 @@ static void ds_kernel_cpu(void *descr[], __attribute__((unused)) void *arg)
 }
 
 static struct starpu_codelet_t ds_codelet = {
-	.where = CORE,
+	.where = STARPU_CORE,
 	.core_func = ds_kernel_cpu,
 	.nbuffers = 2, /* input -> output */
 	.model = NULL
@@ -229,11 +229,11 @@ int main(int argc, char **argv)
 				task->callback_func = ds_callback;
 
 				/* input */
-				task->buffers[0].handle = get_sub_data(frame_y_handle[frame], 1, blocky);
+				task->buffers[0].handle = starpu_get_sub_data(frame_y_handle[frame], 1, blocky);
 				task->buffers[0].mode = STARPU_R;
 
 				/* output */
-				task->buffers[1].handle = get_sub_data(new_frame_y_handle[frame], 1, blocky);
+				task->buffers[1].handle = starpu_get_sub_data(new_frame_y_handle[frame], 1, blocky);
 				task->buffers[1].mode = STARPU_W;
 
 			starpu_submit_task(task);
@@ -247,11 +247,11 @@ int main(int argc, char **argv)
 				task->callback_func = ds_callback;
 
 				/* input */
-				task->buffers[0].handle = get_sub_data(frame_u_handle[frame], 1, blocku);
+				task->buffers[0].handle = starpu_get_sub_data(frame_u_handle[frame], 1, blocku);
 				task->buffers[0].mode = STARPU_R;
 
 				/* output */
-				task->buffers[1].handle = get_sub_data(new_frame_u_handle[frame], 1, blocku);
+				task->buffers[1].handle = starpu_get_sub_data(new_frame_u_handle[frame], 1, blocku);
 				task->buffers[1].mode = STARPU_W;
 
 			starpu_submit_task(task);
@@ -265,11 +265,11 @@ int main(int argc, char **argv)
 				task->callback_func = ds_callback;
 
 				/* input */
-				task->buffers[0].handle = get_sub_data(frame_v_handle[frame], 1, blockv);
+				task->buffers[0].handle = starpu_get_sub_data(frame_v_handle[frame], 1, blockv);
 				task->buffers[0].mode = STARPU_R;
 
 				/* output */
-				task->buffers[1].handle = get_sub_data(new_frame_v_handle[frame], 1, blockv);
+				task->buffers[1].handle = starpu_get_sub_data(new_frame_v_handle[frame], 1, blockv);
 				task->buffers[1].mode = STARPU_W;
 
 			starpu_submit_task(task);

examples/starpufft/starpufftx1d.c

@@ -232,36 +232,36 @@ STARPUFFT(twist3_1d_kernel_cpu)(void *descr[], void *_args)
 }
 
 static struct starpu_perfmodel_t STARPUFFT(twist1_1d_model) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = TYPE"twist1_1d"
 };
 
 static struct starpu_perfmodel_t STARPUFFT(fft1_1d_model) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = TYPE"fft1_1d"
 };
 
 static struct starpu_perfmodel_t STARPUFFT(twist2_1d_model) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = TYPE"twist2_1d"
 };
 
 static struct starpu_perfmodel_t STARPUFFT(fft2_1d_model) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = TYPE"fft2_1d"
 };
 
 static struct starpu_perfmodel_t STARPUFFT(twist3_1d_model) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = TYPE"twist3_1d"
 };
 
 static starpu_codelet STARPUFFT(twist1_1d_codelet) = {
 	.where =
 #ifdef USE_CUDA
-		CUDA|
+		STARPU_CUDA|
 #endif
-		CORE,
+		STARPU_CORE,
 #ifdef USE_CUDA
 	.cuda_func = STARPUFFT(twist1_1d_kernel_gpu),
 #endif
@@ -273,10 +273,10 @@ static starpu_codelet STARPUFFT(twist1_1d_codelet) = {
 static starpu_codelet STARPUFFT(fft1_1d_codelet) = {
 	.where =
 #ifdef USE_CUDA
-		CUDA|
+		STARPU_CUDA|
 #endif
 #ifdef HAVE_FFTW
-		CORE|
+		STARPU_CORE|
 #endif
 		0,
 #ifdef USE_CUDA
@@ -290,7 +290,7 @@ static starpu_codelet STARPUFFT(fft1_1d_codelet) = {
 };
 
 static starpu_codelet STARPUFFT(twist2_1d_codelet) = {
-	.where = CORE,
+	.where = STARPU_CORE,
 	.core_func = STARPUFFT(twist2_1d_kernel_cpu),
 	.model = &STARPUFFT(twist2_1d_model),
 	.nbuffers = 1
@@ -299,10 +299,10 @@ static starpu_codelet STARPUFFT(twist2_1d_codelet) = {
 static starpu_codelet STARPUFFT(fft2_1d_codelet) = {
 	.where =
 #ifdef USE_CUDA
-		CUDA|
+		STARPU_CUDA|
 #endif
 #ifdef HAVE_FFTW
-		CORE|
+		STARPU_CORE|
 #endif
 		0,
 #ifdef USE_CUDA
@@ -316,7 +316,7 @@ static starpu_codelet STARPUFFT(fft2_1d_codelet) = {
 };
 
 static starpu_codelet STARPUFFT(twist3_1d_codelet) = {
-	.where = CORE,
+	.where = STARPU_CORE,
 	.core_func = STARPUFFT(twist3_1d_kernel_cpu),
 	.model = &STARPUFFT(twist3_1d_model),
 	.nbuffers = 1

examples/starpufft/starpufftx2d.c

@@ -272,36 +272,36 @@ STARPUFFT(twist3_2d_kernel_cpu)(void *descr[], void *_args)
 }
 
 struct starpu_perfmodel_t STARPUFFT(twist1_2d_model) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = TYPE"twist1_2d"
 };
 
 struct starpu_perfmodel_t STARPUFFT(fft1_2d_model) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = TYPE"fft1_2d"
 };
 
 struct starpu_perfmodel_t STARPUFFT(twist2_2d_model) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = TYPE"twist2_2d"
 };
 
 struct starpu_perfmodel_t STARPUFFT(fft2_2d_model) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = TYPE"fft2_2d"
 };
 
 struct starpu_perfmodel_t STARPUFFT(twist3_2d_model) = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = TYPE"twist3_2d"
 };
 
 static starpu_codelet STARPUFFT(twist1_2d_codelet) = {
 	.where =
 #ifdef USE_CUDA
-		CUDA|
+		STARPU_CUDA|
 #endif
-		CORE,
+		STARPU_CORE,
 #ifdef USE_CUDA
 	.cuda_func = STARPUFFT(twist1_2d_kernel_gpu),
 #endif
@@ -313,10 +313,10 @@ static starpu_codelet STARPUFFT(twist1_2d_codelet) = {
 static starpu_codelet STARPUFFT(fft1_2d_codelet) = {
 	.where =
 #ifdef USE_CUDA
-		CUDA|
+		STARPU_CUDA|
 #endif
 #ifdef HAVE_FFTW
-		CORE|
+		STARPU_CORE|
 #endif
 		0,
 #ifdef USE_CUDA
@@ -330,7 +330,7 @@ static starpu_codelet STARPUFFT(fft1_2d_codelet) = {
 };
 
 static starpu_codelet STARPUFFT(twist2_2d_codelet) = {
-	.where = CORE,
+	.where = STARPU_CORE,
 	.core_func = STARPUFFT(twist2_2d_kernel_cpu),
 	.model = &STARPUFFT(twist2_2d_model),
 	.nbuffers = 1
@@ -339,10 +339,10 @@ static starpu_codelet STARPUFFT(twist2_2d_codelet) = {
 static starpu_codelet STARPUFFT(fft2_2d_codelet) = {
 	.where =
 #ifdef USE_CUDA
-		CUDA|
+		STARPU_CUDA|
 #endif
 #ifdef HAVE_FFTW
-		CORE|
+		STARPU_CORE|
 #endif
 		0,
 #ifdef USE_CUDA
@@ -356,7 +356,7 @@ static starpu_codelet STARPUFFT(fft2_2d_codelet) = {
 };
 
 static starpu_codelet STARPUFFT(twist3_2d_codelet) = {
-	.where = CORE,
+	.where = STARPU_CORE,
 	.core_func = STARPUFFT(twist3_2d_kernel_cpu),
 	.model = &STARPUFFT(twist3_2d_model),
 	.nbuffers = 1

examples/strassen2/strassen2.c

@@ -98,28 +98,28 @@ extern void null_codelet(__attribute__((unused)) void *descr[],
 extern void display_perf(double timing, unsigned size);
 
 struct starpu_perfmodel_t strassen_model_mult = {
-        .type = HISTORY_BASED,
+        .type = STARPU_HISTORY_BASED,
         .symbol = "strassen_model_mult"
 };
 
 struct starpu_perfmodel_t strassen_model_add = {
-        .type = HISTORY_BASED,
+        .type = STARPU_HISTORY_BASED,
         .symbol = "strassen_model_add"
 };
 
 struct starpu_perfmodel_t strassen_model_sub = {
-        .type = HISTORY_BASED,
+        .type = STARPU_HISTORY_BASED,
         .symbol = "strassen_model_sub"
 };
 
 
 struct starpu_perfmodel_t strassen_model_self_add = {
-        .type = HISTORY_BASED,
+        .type = STARPU_HISTORY_BASED,
         .symbol = "strassen_model_self_add"
 };
 
 struct starpu_perfmodel_t strassen_model_self_sub = {
-        .type = HISTORY_BASED,
+        .type = STARPU_HISTORY_BASED,
         .symbol = "strassen_model_self_sub"
 };
 
@@ -207,7 +207,7 @@ enum operation {
 };
 
 static starpu_codelet cl_add = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.model = &strassen_model_add,
 	.core_func = add_core_codelet,
 #ifdef USE_CUDA
@@ -217,7 +217,7 @@ static starpu_codelet cl_add = {
 };
 
 static starpu_codelet cl_sub = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.model = &strassen_model_sub,
 	.core_func = sub_core_codelet,
 #ifdef USE_CUDA
@@ -227,7 +227,7 @@ static starpu_codelet cl_sub = {
 };
 
 static starpu_codelet cl_mult = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.model = &strassen_model_mult,
 	.core_func = mult_core_codelet,
 #ifdef USE_CUDA
@@ -273,7 +273,7 @@ struct starpu_task *compute_add_sub_op(starpu_data_handle C, enum operation op,
 }
 
 static starpu_codelet cl_self_add = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.model = &strassen_model_self_add,
 	.core_func = self_add_core_codelet,
 #ifdef USE_CUDA
@@ -283,7 +283,7 @@ static starpu_codelet cl_self_add = {
 };
 
 static starpu_codelet cl_self_sub = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.model = &strassen_model_self_sub,
 	.core_func = self_sub_core_codelet,
 #ifdef USE_CUDA
@@ -343,7 +343,7 @@ void cleanup_callback(void *_arg)
 }
 
 static starpu_codelet cleanup_codelet = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.model = NULL,
 	.core_func = null_codelet,
 #ifdef USE_CUDA
@@ -402,20 +402,20 @@ void strassen_mult(struct strassen_iter *iter)
 		return;
 	}
 
-        starpu_data_handle A11 = get_sub_data(iter->A, 2, 0, 0);
-        starpu_data_handle A12 = get_sub_data(iter->A, 2, 1, 0);
-        starpu_data_handle A21 = get_sub_data(iter->A, 2, 0, 1);
-        starpu_data_handle A22 = get_sub_data(iter->A, 2, 1, 1);
+        starpu_data_handle A11 = starpu_get_sub_data(iter->A, 2, 0, 0);
+        starpu_data_handle A12 = starpu_get_sub_data(iter->A, 2, 1, 0);
+        starpu_data_handle A21 = starpu_get_sub_data(iter->A, 2, 0, 1);
+        starpu_data_handle A22 = starpu_get_sub_data(iter->A, 2, 1, 1);
 
-        starpu_data_handle B11 = get_sub_data(iter->B, 2, 0, 0);
-        starpu_data_handle B12 = get_sub_data(iter->B, 2, 1, 0);
-        starpu_data_handle B21 = get_sub_data(iter->B, 2, 0, 1);
-        starpu_data_handle B22 = get_sub_data(iter->B, 2, 1, 1);
+        starpu_data_handle B11 = starpu_get_sub_data(iter->B, 2, 0, 0);
+        starpu_data_handle B12 = starpu_get_sub_data(iter->B, 2, 1, 0);
+        starpu_data_handle B21 = starpu_get_sub_data(iter->B, 2, 0, 1);
+        starpu_data_handle B22 = starpu_get_sub_data(iter->B, 2, 1, 1);
 
-        starpu_data_handle C11 = get_sub_data(iter->C, 2, 0, 0);
-        starpu_data_handle C12 = get_sub_data(iter->C, 2, 1, 0);
-        starpu_data_handle C21 = get_sub_data(iter->C, 2, 0, 1);
-        starpu_data_handle C22 = get_sub_data(iter->C, 2, 1, 1);
+        starpu_data_handle C11 = starpu_get_sub_data(iter->C, 2, 0, 0);
+        starpu_data_handle C12 = starpu_get_sub_data(iter->C, 2, 1, 0);
+        starpu_data_handle C21 = starpu_get_sub_data(iter->C, 2, 0, 1);
+        starpu_data_handle C22 = starpu_get_sub_data(iter->C, 2, 1, 1);
 
 	unsigned size = starpu_get_blas_nx(A11);
 
@@ -719,7 +719,7 @@ static void dummy_codelet_func(__attribute__((unused))void *descr[],
 }
 
 static starpu_codelet dummy_codelet = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.model = NULL,
 	.core_func = dummy_codelet_func,
 	#ifdef USE_CUDA

examples/strassen2/strassen2_kernels.c

@@ -53,15 +53,15 @@ void display_perf(double timing, unsigned size)
 
 static void mult_common_codelet(void *descr[], int s, __attribute__((unused))  void *arg)
 {
-	float *center 	= (float *)GET_BLAS_PTR(descr[0]);
-	float *left 	= (float *)GET_BLAS_PTR(descr[1]);
-	float *right 	= (float *)GET_BLAS_PTR(descr[2]);
+	float *center 	= (float *)STARPU_GET_BLAS_PTR(descr[0]);
+	float *left 	= (float *)STARPU_GET_BLAS_PTR(descr[1]);
+	float *right 	= (float *)STARPU_GET_BLAS_PTR(descr[2]);
 
-	unsigned n = GET_BLAS_NX(descr[0]);
+	unsigned n = STARPU_GET_BLAS_NX(descr[0]);
 
-	unsigned ld21 = GET_BLAS_LD(descr[1]);
-	unsigned ld12 = GET_BLAS_LD(descr[2]);
-	unsigned ld22 = GET_BLAS_LD(descr[0]);
+	unsigned ld21 = STARPU_GET_BLAS_LD(descr[1]);
+	unsigned ld12 = STARPU_GET_BLAS_LD(descr[2]);
+	unsigned ld22 = STARPU_GET_BLAS_LD(descr[0]);
 
 	double flop = 2.0*n*n*n;
 
@@ -106,15 +106,15 @@ static void add_sub_common_codelet(void *descr[], int s, __attribute__((unused))
 {
 	/* C = A op B */
 
-	float *C 	= (float *)GET_BLAS_PTR(descr[0]);
-	float *A 	= (float *)GET_BLAS_PTR(descr[1]);
-	float *B 	= (float *)GET_BLAS_PTR(descr[2]);
+	float *C 	= (float *)STARPU_GET_BLAS_PTR(descr[0]);
+	float *A 	= (float *)STARPU_GET_BLAS_PTR(descr[1]);
+	float *B 	= (float *)STARPU_GET_BLAS_PTR(descr[2]);
 
-	unsigned n = GET_BLAS_NX(descr[0]);
+	unsigned n = STARPU_GET_BLAS_NX(descr[0]);
 
-	unsigned ldA = GET_BLAS_LD(descr[1]);
-	unsigned ldB = GET_BLAS_LD(descr[2]);
-	unsigned ldC = GET_BLAS_LD(descr[0]);
+	unsigned ldA = STARPU_GET_BLAS_LD(descr[1]);
+	unsigned ldB = STARPU_GET_BLAS_LD(descr[2]);
+	unsigned ldC = STARPU_GET_BLAS_LD(descr[0]);
 
 	double flop = 2.0*n*n;
 
@@ -188,13 +188,13 @@ static void self_add_sub_common_codelet(void *descr[], int s, __attribute__((unu
 {
 	/* C +=/-= A */
 
-	float *C 	= (float *)GET_BLAS_PTR(descr[0]);
-	float *A 	= (float *)GET_BLAS_PTR(descr[1]);
+	float *C 	= (float *)STARPU_GET_BLAS_PTR(descr[0]);
+	float *A 	= (float *)STARPU_GET_BLAS_PTR(descr[1]);
 
-	unsigned n = GET_BLAS_NX(descr[0]);
+	unsigned n = STARPU_GET_BLAS_NX(descr[0]);
 
-	unsigned ldA = GET_BLAS_LD(descr[1]);
-	unsigned ldC = GET_BLAS_LD(descr[0]);
+	unsigned ldA = STARPU_GET_BLAS_LD(descr[1]);
+	unsigned ldC = STARPU_GET_BLAS_LD(descr[0]);
 
 	double flop = 1.0*n*n;
 

examples/tag_example/tag_example.c

@@ -199,7 +199,7 @@ int main(int argc __attribute__((unused)) , char **argv __attribute__((unused)))
 
 	fprintf(stderr, "ITER: %d\n", nk);
 
-	cl.where = CORE|CUDA|GORDON;
+	cl.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON;
 	cl.core_func = core_codelet;
 	cl.cuda_func = core_codelet;
 #ifdef USE_GORDON

examples/tag_example/tag_example2.c

@@ -123,7 +123,7 @@ int main(int argc __attribute__((unused)) , char **argv __attribute__((unused)))
 #ifdef USE_GORDON
 	cl.gordon_func = gordon_null_kernel;
 #endif
-	cl.where = CORE|CUDA|GORDON;
+	cl.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON;
 	cl.nbuffers = 0;
 
 	fprintf(stderr, "ITER : %d\n", nk);

examples/tag_example/tag_restartable.c

@@ -128,7 +128,7 @@ int main(int argc __attribute__((unused)) , char **argv __attribute__((unused)))
 #ifdef USE_GORDON
 	cl.gordon_func = gordon_null_kernel;
 #endif
-	cl.where = CORE|CUDA|GORDON;
+	cl.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON;
 	cl.nbuffers = 0;
 
 	fprintf(stderr, "ITER : %d\n", nk);

include/starpu-data-filters.h

@@ -33,14 +33,14 @@ typedef struct starpu_filter_t {
 void starpu_partition_data(starpu_data_handle initial_data, starpu_filter *f); 
 void starpu_unpartition_data(starpu_data_handle root_data, uint32_t gathering_node);
 
-struct data_interface_ops_t;
+struct starpu_data_interface_ops_t;
 void starpu_data_create_children(starpu_data_handle handle, unsigned nchildren,
-		 struct data_interface_ops_t *children_interface_ops);
+		 struct starpu_data_interface_ops_t *children_interface_ops);
 
 starpu_data_handle starpu_data_get_child(starpu_data_handle handle, unsigned i);
 
 /* unsigned list */
-starpu_data_handle get_sub_data(starpu_data_handle root_data, unsigned depth, ... );
+starpu_data_handle starpu_get_sub_data(starpu_data_handle root_data, unsigned depth, ... );
 
 /* starpu_filter * list */
 void starpu_map_filters(starpu_data_handle root_data, unsigned nfilters, ...);

include/starpu-data-interfaces.h

@@ -45,11 +45,11 @@ uintptr_t starpu_get_blas_local_ptr(starpu_data_handle handle);
 size_t starpu_get_blas_elemsize(starpu_data_handle handle);
 
 /* helper methods */
-#define GET_BLAS_PTR(interface)	(((starpu_blas_interface_t *)(interface))->ptr)
-#define GET_BLAS_NX(interface)	(((starpu_blas_interface_t *)(interface))->nx)
-#define GET_BLAS_NY(interface)	(((starpu_blas_interface_t *)(interface))->ny)
-#define GET_BLAS_LD(interface)	(((starpu_blas_interface_t *)(interface))->ld)
-#define GET_BLAS_ELEMSIZE(interface)	(((starpu_blas_interface_t *)(interface))->elemsize)
+#define STARPU_GET_BLAS_PTR(interface)	(((starpu_blas_interface_t *)(interface))->ptr)
+#define STARPU_GET_BLAS_NX(interface)	(((starpu_blas_interface_t *)(interface))->nx)
+#define STARPU_GET_BLAS_NY(interface)	(((starpu_blas_interface_t *)(interface))->ny)
+#define STARPU_GET_BLAS_LD(interface)	(((starpu_blas_interface_t *)(interface))->ld)
+#define STARPU_GET_BLAS_ELEMSIZE(interface)	(((starpu_blas_interface_t *)(interface))->elemsize)
 
 
 /* BLOCK interface for 3D dense blocks */

include/starpu-data.h

@@ -40,8 +40,8 @@ typedef struct starpu_buffer_descr_t {
 	starpu_access_mode mode;
 } starpu_buffer_descr;
 
-struct data_interface_ops_t;
-starpu_data_handle starpu_data_state_create(struct data_interface_ops_t *interface_ops);
+struct starpu_data_interface_ops_t;
+starpu_data_handle starpu_data_state_create(struct starpu_data_interface_ops_t *interface_ops);
 
 void starpu_unpartition_data(starpu_data_handle root_data, uint32_t gathering_node);
 void starpu_delete_data(starpu_data_handle state);

include/starpu-perfmodel.h

@@ -43,7 +43,7 @@ enum starpu_perf_archtype {
 	STARPU_GORDON_DEFAULT = STARPU_CUDA_DEFAULT + STARPU_MAXCUDADEVS
 };
 
-#define NARCH_VARIATIONS	(STARPU_GORDON_DEFAULT+1)
+#define STARPU_NARCH_VARIATIONS	(STARPU_GORDON_DEFAULT+1)
 
 struct starpu_regression_model_t {
 	/* sum of ln(measured) */
@@ -78,7 +78,7 @@ struct starpu_per_arch_perfmodel_t {
 #endif
 };
 
-typedef enum {PER_ARCH, COMMON, HISTORY_BASED, REGRESSION_BASED} starpu_perfmodel_type;
+typedef enum {STARPU_PER_ARCH, STARPU_COMMON, STARPU_HISTORY_BASED, STARPU_REGRESSION_BASED} starpu_perfmodel_type;
 
 struct starpu_perfmodel_t {
 	/* which model is used for that task ? */
@@ -88,7 +88,7 @@ struct starpu_perfmodel_t {
 	double (*cost_model)(struct starpu_buffer_descr_t *);
 
 	/* per-architecture model */
-	struct starpu_per_arch_perfmodel_t per_arch[NARCH_VARIATIONS];
+	struct starpu_per_arch_perfmodel_t per_arch[STARPU_NARCH_VARIATIONS];
 	
 	const char *symbol;
 	enum {

include/starpu-task.h

@@ -27,14 +27,14 @@
 
 #include <starpu-data.h>
 
-#define CORE	((1ULL)<<1)
-#define CUDA	((1ULL)<<3)
-#define SPU	((1ULL)<<4)
-#define GORDON	((1ULL)<<5)
+#define STARPU_CORE	((1ULL)<<1)
+#define STARPU_CUDA	((1ULL)<<3)
+#define STARPU_SPU	((1ULL)<<4)
+#define STARPU_GORDON	((1ULL)<<5)
 
-#define MIN_PRIO        (-4)
-#define MAX_PRIO        5
-#define DEFAULT_PRIO	0
+#define STARPU_MIN_PRIO        (-4)
+#define STARPU_MAX_PRIO        5
+#define STARPU_DEFAULT_PRIO	0
 
 #ifdef __cplusplus
 extern "C" {
@@ -129,7 +129,7 @@ struct starpu_task {
 	.cl_arg_size = 0,				\
 	.callback_func = NULL,				\
 	.callback_arg = NULL,				\
-	.priority = DEFAULT_PRIO,			\
+	.priority = STARPU_DEFAULT_PRIO,                \
 	.use_tag = 0,					\
 	.synchronous = 0,				\
 	.execute_on_a_specific_worker = 0,		\

src/core/jobs.h

@@ -45,10 +45,10 @@ struct worker_s;
 typedef void (*cl_func)(void **, void *);
 typedef void (*callback)(void *);
 
-#define CORE_MAY_PERFORM(j)	((j)->task->cl->where & CORE)
-#define CUDA_MAY_PERFORM(j)     ((j)->task->cl->where & CUDA)
-#define SPU_MAY_PERFORM(j)	((j)->task->cl->where & SPU)
-#define GORDON_MAY_PERFORM(j)	((j)->task->cl->where & GORDON)
+#define STARPU_CORE_MAY_PERFORM(j)	((j)->task->cl->where & STARPU_CORE)
+#define STARPU_CUDA_MAY_PERFORM(j)     ((j)->task->cl->where & STARPU_CUDA)
+#define STARPU_SPU_MAY_PERFORM(j)	((j)->task->cl->where & STARPU_SPU)
+#define STARPU_GORDON_MAY_PERFORM(j)	((j)->task->cl->where & STARPU_GORDON)
 
 /* a job is the internal representation of a task */
 LIST_TYPE(job,

src/core/perfmodel/perfmodel.c

@@ -66,10 +66,10 @@ static double common_task_expected_length(struct starpu_perfmodel_t *model, uint
 		float alpha;
 		exp = model->cost_model(task->buffers);
 		switch (who) {
-			case CORE:
+			case STARPU_CORE:
 				alpha = CORE_ALPHA;
 				break;
-			case CUDA:
+			case STARPU_CUDA:
 				alpha = CUDA_ALPHA;
 				break;
 			default:
@@ -93,16 +93,16 @@ double job_expected_length(uint32_t who, struct job_s *j, enum starpu_perf_archt
 
 	if (model) {
 		switch (model->type) {
-			case PER_ARCH:
+			case STARPU_PER_ARCH:
 				return per_arch_task_expected_length(model, arch, task);
 
-			case COMMON:
+			case STARPU_COMMON:
 				return common_task_expected_length(model, who, task);
 
-			case HISTORY_BASED:
+			case STARPU_HISTORY_BASED:
 				return history_based_job_expected_length(model, arch, j);
 
-			case REGRESSION_BASED:
+			case STARPU_REGRESSION_BASED:
 				return regression_based_job_expected_length(model, arch, j);
 
 			default:

src/core/perfmodel/perfmodel_history.c

@@ -126,7 +126,7 @@ static void parse_per_arch_model_file(FILE *f, struct starpu_per_arch_perfmodel_
 static void parse_model_file(FILE *f, struct starpu_perfmodel_t *model, unsigned scan_history)
 {
 	unsigned arch;
-	for (arch = 0; arch < NARCH_VARIATIONS; arch++)
+	for (arch = 0; arch < STARPU_NARCH_VARIATIONS; arch++)
 		parse_per_arch_model_file(f, &model->per_arch[arch], scan_history);
 }
 
@@ -166,7 +166,7 @@ static void dump_model_file(FILE *f, struct starpu_perfmodel_t *model)
 	fprintf(f, "#################\n");
 
 	unsigned arch;
-	for (arch = 0; arch < NARCH_VARIATIONS; arch++)
+	for (arch = 0; arch < STARPU_NARCH_VARIATIONS; arch++)
 	{
 		char archname[32];
 		starpu_perfmodel_get_arch_name(arch, archname, 32);
@@ -185,7 +185,7 @@ static void initialize_per_arch_model(struct starpu_per_arch_perfmodel_t *per_ar
 static void initialize_model(struct starpu_perfmodel_t *model)
 {
 	unsigned arch;
-	for (arch = 0; arch < NARCH_VARIATIONS; arch++)
+	for (arch = 0; arch < STARPU_NARCH_VARIATIONS; arch++)
 		initialize_per_arch_model(&model->per_arch[arch]);
 }
 
@@ -532,7 +532,7 @@ void _starpu_update_perfmodel_history(job_t j, enum starpu_perf_archtype arch, u
 	{
 		struct starpu_per_arch_perfmodel_t *per_arch_model = &model->per_arch[arch];
 
-		if (model->type == HISTORY_BASED || model->type == REGRESSION_BASED)
+		if (model->type == STARPU_HISTORY_BASED || model->type == STARPU_REGRESSION_BASED)
 		{
 			uint32_t key = j->footprint;
 			struct starpu_history_entry_t *entry;

src/core/policies/deque-modeling-policy-data-aware.c

@@ -185,7 +185,7 @@ static int dmda_push_prio_task(struct jobq_s *q, job_t j)
 
 static int dmda_push_task(struct jobq_s *q, job_t j)
 {
-	if (j->task->priority == MAX_PRIO)
+	if (j->task->priority == STARPU_MAX_PRIO)
 		return _dmda_push_task(q, j, 1);
 
 	return _dmda_push_task(q, j, 0);

src/core/policies/deque-modeling-policy.c

@@ -146,7 +146,7 @@ static int dm_push_prio_task(struct jobq_s *q, job_t j)
 
 static int dm_push_task(struct jobq_s *q, job_t j)
 {
-	if (j->task->priority == MAX_PRIO)
+	if (j->task->priority == STARPU_MAX_PRIO)
 		return _dm_push_task(q, j, 1);
 
 	return _dm_push_task(q, j, 0);

src/core/task.c

@@ -38,7 +38,7 @@ void starpu_task_init(struct starpu_task *task)
 	task->callback_func = NULL;
 	task->callback_arg = NULL;
 
-	task->priority = DEFAULT_PRIO;
+	task->priority = STARPU_DEFAULT_PRIO;
 	task->use_tag = 0;
 	task->synchronous = 0;
 

src/core/topology.c

@@ -204,8 +204,8 @@ static int _starpu_init_machine_config(struct machine_config_s *config,
 		enum starpu_perf_archtype arch = STARPU_CUDA_DEFAULT + devid;
 		config->workers[config->nworkers + cudagpu].id = devid;
 		config->workers[config->nworkers + cudagpu].perf_arch = arch; 
-		config->workers[config->nworkers + cudagpu].worker_mask = CUDA;
-		config->worker_mask |= CUDA;
+		config->workers[config->nworkers + cudagpu].worker_mask = STARPU_CUDA;
+		config->worker_mask |= STARPU_CUDA;
 	}
 
 	config->nworkers += config->ncudagpus;
@@ -272,8 +272,8 @@ static int _starpu_init_machine_config(struct machine_config_s *config,
 		config->workers[config->nworkers + core].arch = STARPU_CORE_WORKER;
 		config->workers[config->nworkers + core].perf_arch = STARPU_CORE_DEFAULT;
 		config->workers[config->nworkers + core].id = core;
-		config->workers[config->nworkers + core].worker_mask = CORE;
-		config->worker_mask |= CORE;
+		config->workers[config->nworkers + core].worker_mask = STARPU_CORE;
+		config->worker_mask |= STARPU_CORE;
 	}
 
 	config->nworkers += config->ncores;

src/core/workers.c

@@ -39,12 +39,12 @@ inline uint32_t _starpu_worker_exists(uint32_t task_mask)
 
 inline uint32_t may_submit_cuda_task(void)
 {
-	return (CUDA & config.worker_mask);
+	return (STARPU_CUDA & config.worker_mask);
 }
 
 inline uint32_t may_submit_core_task(void)
 {
-	return (CORE & config.worker_mask);
+	return (STARPU_CORE & config.worker_mask);
 }
 
 inline uint32_t _starpu_worker_may_execute_task(unsigned workerid, uint32_t where)

src/datawizard/coherency.h

@@ -92,7 +92,7 @@ struct starpu_data_state_t {
 	/* describe the actual data layout */
 	void *interface[STARPU_MAXNODES];
 
-	struct data_interface_ops_t *ops;
+	struct starpu_data_interface_ops_t *ops;
 
 	/* where is the data home ? -1 if none yet */
 	int data_home;

src/datawizard/hierarchy.c

@@ -144,7 +144,7 @@ starpu_data_handle starpu_data_get_child(starpu_data_handle handle, unsigned i)
 	return &handle->children[i];
 }
 
-starpu_data_handle get_sub_data(starpu_data_handle root_handle, unsigned depth, ... )
+starpu_data_handle starpu_get_sub_data(starpu_data_handle root_handle, unsigned depth, ... )
 {
 	STARPU_ASSERT(root_handle);
 	starpu_data_handle current_handle = root_handle;
@@ -328,7 +328,7 @@ void starpu_advise_if_data_is_important(starpu_data_handle handle, unsigned is_i
 
 }
 
-starpu_data_handle starpu_data_state_create(struct data_interface_ops_t *interface_ops)
+starpu_data_handle starpu_data_state_create(struct starpu_data_interface_ops_t *interface_ops)
 {
 	starpu_data_handle handle =
 		calloc(1, sizeof(struct starpu_data_state_t));
@@ -352,7 +352,7 @@ starpu_data_handle starpu_data_state_create(struct data_interface_ops_t *interfa
 /* TODO create an alternative version of that function which takes an array of
  * data interface ops in case each child may have its own interface type */
 void starpu_data_create_children(starpu_data_handle handle,
-		unsigned nchildren, struct data_interface_ops_t *children_interface_ops)
+		unsigned nchildren, struct starpu_data_interface_ops_t *children_interface_ops)
 {
 	handle->children = calloc(nchildren, sizeof(struct starpu_data_state_t));
 	STARPU_ASSERT(handle->children);

src/datawizard/interfaces/bcsr_filters.c

@@ -18,7 +18,7 @@
 #include <common/config.h>
 #include <datawizard/hierarchy.h>
 
-extern struct data_interface_ops_t interface_blas_ops;
+extern struct starpu_data_interface_ops_t interface_blas_ops;
 
 void starpu_canonical_block_filter_bcsr(starpu_filter *f __attribute__((unused)), starpu_data_handle root_handle)
 {

src/datawizard/interfaces/bcsr_interface.c

@@ -52,7 +52,7 @@ static void liberate_bcsr_buffer_on_node(void *interface, uint32_t node);
 static size_t bcsr_interface_get_size(starpu_data_handle handle);
 static uint32_t footprint_bcsr_interface_crc32(starpu_data_handle handle);
 
-struct data_interface_ops_t interface_bcsr_ops = {
+struct starpu_data_interface_ops_t interface_bcsr_ops = {
 	.register_data_handle = register_bcsr_handle,
 	.allocate_data_on_node = allocate_bcsr_buffer_on_node,
 	.liberate_data_on_node = liberate_bcsr_buffer_on_node,

src/datawizard/interfaces/blas_interface.c

@@ -61,7 +61,7 @@ static void display_blas_interface(starpu_data_handle handle, FILE *f);
 static int convert_blas_to_gordon(void *interface, uint64_t *ptr, gordon_strideSize_t *ss); 
 #endif
 
-struct data_interface_ops_t interface_blas_ops = {
+struct starpu_data_interface_ops_t interface_blas_ops = {
 	.register_data_handle = register_blas_handle,
 	.allocate_data_on_node = allocate_blas_buffer_on_node,
 	.liberate_data_on_node = liberate_blas_buffer_on_node,

src/datawizard/interfaces/block_interface.c

@@ -57,7 +57,7 @@ static void display_block_interface(starpu_data_handle handle, FILE *f);
 static int convert_block_to_gordon(void *interface, uint64_t *ptr, gordon_strideSize_t *ss);
 #endif
 
-struct data_interface_ops_t interface_block_ops = {
+struct starpu_data_interface_ops_t interface_block_ops = {
 	.register_data_handle = register_block_handle,
 	.allocate_data_on_node = allocate_block_buffer_on_node,
 	.liberate_data_on_node = liberate_block_buffer_on_node,

src/datawizard/interfaces/csr_interface.c

@@ -49,7 +49,7 @@ static void liberate_csr_buffer_on_node(void *interface, uint32_t node);
 static size_t csr_interface_get_size(starpu_data_handle handle);
 static uint32_t footprint_csr_interface_crc32(starpu_data_handle handle);
 
-struct data_interface_ops_t interface_csr_ops = {
+struct starpu_data_interface_ops_t interface_csr_ops = {
 	.register_data_handle = register_csr_handle,
 	.allocate_data_on_node = allocate_csr_buffer_on_node,
 	.liberate_data_on_node = liberate_csr_buffer_on_node,

src/datawizard/interfaces/data_interface.c

@@ -28,7 +28,7 @@ void *starpu_data_get_interface_on_node(starpu_data_handle handle, unsigned memo
 
 void register_data_handle(starpu_data_handle *handleptr, uint32_t home_node,
 				void *interface,
-				struct data_interface_ops_t *ops)
+				struct starpu_data_interface_ops_t *ops)
 {
 	starpu_data_handle handle =
 		starpu_data_state_create(ops);

src/datawizard/interfaces/data_interface.h

@@ -25,7 +25,7 @@
 #include <gordon.h>
 #endif
 
-struct data_interface_ops_t {
+struct starpu_data_interface_ops_t {
 	void (*register_data_handle)(starpu_data_handle handle,
 					uint32_t home_node, void *interface);
 	size_t (*allocate_data_on_node)(starpu_data_handle handle, uint32_t node);
@@ -44,6 +44,6 @@ struct data_interface_ops_t {
 
 void register_data_handle(starpu_data_handle *handleptr, uint32_t home_node,
 				void *interface,
-				struct data_interface_ops_t *ops);
+				struct starpu_data_interface_ops_t *ops);
 
 #endif // __DATA_INTERFACE_H__

src/datawizard/interfaces/vector_interface.c

@@ -60,7 +60,7 @@ static void display_vector_interface(starpu_data_handle handle, FILE *f);
 static int convert_vector_to_gordon(void *interface, uint64_t *ptr, gordon_strideSize_t *ss); 
 #endif
 
-struct data_interface_ops_t interface_vector_ops = {
+struct starpu_data_interface_ops_t interface_vector_ops = {
 	.register_data_handle = register_vector_handle,
 	.allocate_data_on_node = allocate_vector_buffer_on_node,
 	.liberate_data_on_node = liberate_vector_buffer_on_node,

src/datawizard/memalloc.h

@@ -38,7 +38,7 @@ LIST_TYPE(mem_chunk,
 	 * the starpu_data_interface_t) to stay on the safe side. We make a copy of
 	 * because when a data is deleted, the memory chunk remains.
 	 */
-	struct data_interface_ops_t *ops;
+	struct starpu_data_interface_ops_t *ops;
 	void *interface;
 	unsigned automatically_allocated;
 	unsigned data_was_deleted;

src/drivers/core/driver_core.c

@@ -170,7 +170,7 @@ void *_starpu_core_worker(void *arg)
 		jobq_unlock(queue);
 
 		/* can a core perform that task ? */
-		if (!CORE_MAY_PERFORM(j)) 
+		if (!STARPU_CORE_MAY_PERFORM(j)) 
 		{
 			/* put it and the end of the queue ... XXX */
 			push_task(j);

src/drivers/cuda/driver_cuda.c

@@ -252,7 +252,7 @@ void *_starpu_cuda_worker(void *arg)
 		jobq_unlock(queue);
 
 		/* can CUDA do that task ? */
-		if (!CUDA_MAY_PERFORM(j))
+		if (!STARPU_CUDA_MAY_PERFORM(j))
 		{
 			/* this is neither a cuda or a cublas task */
 			push_task(j);

src/util/malloc.c

@@ -41,7 +41,7 @@ static void malloc_pinned_codelet(void *buffers[] __attribute__((unused)), void
 }
 
 static starpu_codelet malloc_pinned_cl = {
-	.where = CUDA,
+	.where = STARPU_CUDA,
 	.cuda_func = malloc_pinned_codelet,
 	.model = NULL,
 	.nbuffers = 0
@@ -95,7 +95,7 @@ static void free_pinned_codelet(void *buffers[] __attribute__((unused)), void *a
 }
 
 static starpu_codelet free_pinned_cl = {
-	.where = CUDA,
+	.where = STARPU_CUDA,
 	.cuda_func = free_pinned_codelet,
 	.model = NULL,
 	.nbuffers = 0

src/util/starpu_cublas.c

@@ -34,13 +34,13 @@ static void shutdown_cublas_func(void *args __attribute__((unused)))
 void starpu_helper_init_cublas(void)
 {
 #ifdef USE_CUDA
-	starpu_execute_on_each_worker(init_cublas_func, NULL, CUDA);
+	starpu_execute_on_each_worker(init_cublas_func, NULL, STARPU_CUDA);
 #endif
 }
 
 void starpu_helper_shutdown_cublas(void)
 {
 #ifdef USE_CUDA
-	starpu_execute_on_each_worker(shutdown_cublas_func, NULL, CUDA);
+	starpu_execute_on_each_worker(shutdown_cublas_func, NULL, STARPU_CUDA);
 #endif
 }

tests/core/empty_task_sync_point.c

@@ -34,7 +34,7 @@ static void dummy_func(void *descr[] __attribute__ ((unused)), void *arg __attri
 
 static starpu_codelet dummy_codelet = 
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dummy_func,
 	.cuda_func = dummy_func,
 	.model = NULL,

tests/core/execute_on_a_specific_worker.c

@@ -71,7 +71,7 @@ static starpu_access_mode select_random_mode(void)
 
 
 static starpu_codelet cl = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = codelet_null,
 	.cuda_func = codelet_null,
 	.nbuffers = 1

tests/core/multithreaded.c

@@ -33,7 +33,7 @@ static void dummy_func(void *descr[] __attribute__ ((unused)), void *arg __attri
 
 static starpu_codelet dummy_codelet = 
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dummy_func,
 	.cuda_func = dummy_func,
 	.model = NULL,

tests/core/starpu_wait_all_tasks.c

@@ -28,7 +28,7 @@ static void dummy_func(void *descr[], void *arg)
 
 static starpu_codelet dummy_codelet = 
 {
-	.where = CORE|CUDA|GORDON,
+	.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON,
 	.core_func = dummy_func,
 	.cuda_func = dummy_func,
 #ifdef USE_GORDON

tests/core/starpu_wait_task.c

@@ -29,7 +29,7 @@ static void dummy_func(void *descr[] __attribute__ ((unused)), void *arg __attri
 
 static starpu_codelet dummy_codelet = 
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dummy_func,
 	.cuda_func = dummy_func,
 	.model = NULL,

tests/core/static_restartable.c

@@ -28,7 +28,7 @@ static void dummy_func(void *descr[] __attribute__ ((unused)), void *arg __attri
 
 static starpu_codelet dummy_codelet = 
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dummy_func,
 	.cuda_func = dummy_func,
 	.model = NULL,

tests/core/static_restartable_tag.c

@@ -29,7 +29,7 @@ static void dummy_func(void *descr[] __attribute__ ((unused)), void *arg __attri
 
 static starpu_codelet dummy_codelet = 
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dummy_func,
 	.cuda_func = dummy_func,
 	.model = NULL,

tests/core/static_restartable_using_initializer.c

@@ -31,7 +31,7 @@ static void dummy_func(void *descr[] __attribute__ ((unused)), void *arg __attri
 
 static starpu_codelet dummy_codelet = 
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dummy_func,
 	.cuda_func = dummy_func,
 	.model = NULL,

tests/core/tag-wait-api.c

@@ -26,7 +26,7 @@ static void dummy_func(void *descr[] __attribute__ ((unused)), void *arg __attri
 
 static starpu_codelet dummy_codelet = 
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dummy_func,
 	.cuda_func = dummy_func,
 	.model = NULL,

tests/datawizard/dining_philosophers.c

@@ -27,7 +27,7 @@ static void eat_kernel(void *descr[], void *arg)
 }
 
 static starpu_codelet eating_cl = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.cuda_func = eat_kernel,
 	.core_func = eat_kernel,
 	.nbuffers = 2

tests/datawizard/dsm_stress.c

@@ -74,7 +74,7 @@ static starpu_access_mode select_random_mode(void)
 
 
 static starpu_codelet cl = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = core_codelet_null,
 	.cuda_func = cuda_codelet_null,
 	.nbuffers = 2

tests/datawizard/readers_and_writers.c

@@ -24,7 +24,7 @@ static void dummy_kernel(void *descr[], void *arg)
 }
 
 static starpu_codelet rw_cl = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.cuda_func = dummy_kernel,
 	.core_func = dummy_kernel,
 	.nbuffers = 1

tests/datawizard/sync_and_notify_data.c

@@ -91,7 +91,7 @@ int main(int argc, char **argv)
 		{
 			/* increment a = v[0] */
 			starpu_codelet cl_inc_a = {
-				.where = CORE|CUDA|GORDON,
+				.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON,
 				.core_func = core_codelet_incA,
 #ifdef USE_CUDA
 				.cuda_func = cuda_codelet_incA,
@@ -127,7 +127,7 @@ int main(int argc, char **argv)
 		{
 			/* increment c = v[2] */
 			starpu_codelet cl_inc_c = {
-				.where = CORE|CUDA|GORDON,
+				.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON,
 				.core_func = core_codelet_incC,
 #ifdef USE_CUDA
 				.cuda_func = cuda_codelet_incC,

tests/datawizard/sync_with_data_with_mem.c

@@ -33,7 +33,7 @@ static void dummy_codelet(void *descr[], __attribute__ ((unused)) void *_args)
 }
 
 static starpu_codelet cl = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dummy_codelet,
 #ifdef USE_CUDA
 	.cuda_func = dummy_codelet,

tests/datawizard/sync_with_data_with_mem_non_blocking.c

@@ -33,7 +33,7 @@ static void dummy_codelet(void *descr[], __attribute__ ((unused)) void *_args)
 }
 
 static starpu_codelet cl = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dummy_codelet,
 #ifdef USE_CUDA
 	.cuda_func = dummy_codelet,

tests/datawizard/unpartition.c

@@ -32,7 +32,7 @@ static void dummy_codelet(void *descr[], __attribute__ ((unused)) void *_args)
 }
 
 static starpu_codelet cl = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dummy_codelet,
 #ifdef USE_CUDA
 	.cuda_func = dummy_codelet,
@@ -76,11 +76,11 @@ int main(int argc, char **argv)
 	{
 		starpu_map_filters(v_handle, 1, &f);
 	
-		ret = use_handle(get_sub_data(v_handle, 1, 0));
+		ret = use_handle(starpu_get_sub_data(v_handle, 1, 0));
 		if (ret == -ENODEV)
 			goto enodev;
 	
-		ret = use_handle(get_sub_data(v_handle, 1, 1));
+		ret = use_handle(starpu_get_sub_data(v_handle, 1, 1));
 		if (ret == -ENODEV)
 			goto enodev;
 	

tests/datawizard/write_only_tmp_buffer.c

@@ -51,7 +51,7 @@ static void display_var(void *descr[], __attribute__ ((unused)) void *_args)
 }
 
 static starpu_codelet cl = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = core_codelet_null,
 #ifdef USE_CUDA
 	.cuda_func = cuda_codelet_null,
@@ -60,7 +60,7 @@ static starpu_codelet cl = {
 };
 
 static starpu_codelet display_cl = {
-	.where = CORE,
+	.where = STARPU_CORE,
 	.core_func = display_var,
 	.nbuffers = 1
 };

tests/errorcheck/invalid_blocking_calls.c

@@ -38,7 +38,7 @@ static void wrong_func(void *descr[], void *arg)
 
 static starpu_codelet wrong_codelet = 
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = wrong_func,
 	.cuda_func = wrong_func,
 	.model = NULL,

tests/helper/execute_on_all.c

@@ -32,11 +32,11 @@ int main(int argc, char **argv)
 
 	int arg = 0x42;
 
-	starpu_execute_on_each_worker(func, &arg, CORE|CUDA);
+	starpu_execute_on_each_worker(func, &arg, STARPU_CORE|STARPU_CUDA);
 
-	starpu_execute_on_each_worker(func, &arg, CORE);
+	starpu_execute_on_each_worker(func, &arg, STARPU_CORE);
 	
-	starpu_execute_on_each_worker(func, &arg, CUDA);
+	starpu_execute_on_each_worker(func, &arg, STARPU_CUDA);
 
 	starpu_shutdown();
 

tests/helper/starpu_create_sync_task.c

@@ -25,7 +25,7 @@ static void dummy_func(void *descr[] __attribute__ ((unused)), void *arg __attri
 
 static starpu_codelet dummy_codelet =
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dummy_func,
 	.cuda_func = dummy_func,
 	.nbuffers = 0

tests/microbenchs/async-tasks-overhead.c

@@ -35,7 +35,7 @@ static void dummy_func(void *descr[] __attribute__ ((unused)), void *arg __attri
 
 static starpu_codelet dummy_codelet = 
 {
-	.where = CORE|CUDA|GORDON,
+	.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON,
 	.core_func = dummy_func,
 	.cuda_func = dummy_func,
 #ifdef USE_GORDON

tests/microbenchs/prefetch_data_on_node.c

@@ -75,7 +75,7 @@ static starpu_access_mode select_random_mode(void)
 
 
 static starpu_codelet cl = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = codelet_null,
 	.cuda_func = codelet_null,
 	.nbuffers = 1

tests/microbenchs/redundant_buffer.c

@@ -36,7 +36,7 @@ static void core_codelet_null(void *descr[], __attribute__ ((unused)) void *_arg
 }
 
 static starpu_codelet cl = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = core_codelet_null,
 	.cuda_func = cuda_codelet_null,
 	.nbuffers = 2

tests/microbenchs/sync-tasks-overhead.c

@@ -28,7 +28,7 @@ static void dummy_func(void *descr[] __attribute__ ((unused)), void *arg __attri
 
 static starpu_codelet dummy_codelet = 
 {
-	.where = CORE|CUDA|GORDON,
+	.where = STARPU_CORE|STARPU_CUDA|STARPU_GORDON,
 	.core_func = dummy_func,
 	.cuda_func = dummy_func,
 #ifdef USE_GORDON

tests/microbenchs/tasks-overhead.c

@@ -35,7 +35,7 @@ static void dummy_func(void *descr[] __attribute__ ((unused)), void *arg __attri
 
 static starpu_codelet dummy_codelet = 
 {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = dummy_func,
 	.cuda_func = dummy_func,
 	.model = NULL,

tests/overlap/overlap.c

@@ -54,12 +54,12 @@ static void codelet_sleep(void *descr[], __attribute__ ((unused)) void *_args)
 }
 
 static struct starpu_perfmodel_t model = {
-	.type = HISTORY_BASED,
+	.type = STARPU_HISTORY_BASED,
 	.symbol = NULL /* to be defined later */
 };
 
 static starpu_codelet cl = {
-	.where = CORE|CUDA,
+	.where = STARPU_CORE|STARPU_CUDA,
 	.core_func = codelet_sleep,
 	.cuda_func = codelet_sleep,
 	.nbuffers = 1,
@@ -99,7 +99,7 @@ int main(int argc, char **argv)
 		struct starpu_task *task = starpu_task_create();
 		task->cl = &cl;
 
-		task->buffers[0].handle = get_sub_data(handle, 1, iter);
+		task->buffers[0].handle = starpu_get_sub_data(handle, 1, iter);
 		task->buffers[0].mode = STARPU_R;
 
 		task->callback_func = callback;

tools/perfmodel-display.c

@@ -167,7 +167,7 @@ static void display_all_perf_models(struct starpu_perfmodel_t *model)
 	{
 		/* display all architectures */
 		unsigned archid;
-		for (archid = 0; archid < NARCH_VARIATIONS; archid++)
+		for (archid = 0; archid < STARPU_NARCH_VARIATIONS; archid++)
 		{
 			char archname[32];
 			starpu_perfmodel_get_arch_name(archid, archname, 32);