Test gathering data on the non-home node with asynchronous partitioning.

tests/Makefile.am

@@ -327,6 +327,7 @@ myPROGRAMS +=				\
 	datawizard/temporary_partition		\
 	datawizard/partitioned_initialization	\
 	datawizard/partitioned_acquire		\
+	datawizard/partitioned_gather		\
 	datawizard/temporary_partition_implicit	\
 	datawizard/temporary_partition_read	\
 	datawizard/redux_acquire		\
@@ -535,6 +536,10 @@ datawizard_lazy_allocation_SOURCES =		\
 	datawizard/lazy_allocation.c		\
 	vector/memset.c
 
+datawizard_partitioned_gather_SOURCES =		\
+	datawizard/partitioned_gather.c		\
+	vector/memset.c
+
 datawizard_wt_host_SOURCES =		\
 	datawizard/wt_host.c		\
 	variable/increment.c

tests/datawizard/partitioned_gather.c

@@ -0,0 +1,108 @@
+/* StarPU --- Runtime system for heterogeneous multicore architectures.
+ *
+ * Copyright (C) 2010-2024  University of Bordeaux, CNRS (LaBRI UMR 5800), Inria
+ *
+ * StarPU is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU Lesser General Public License as published by
+ * the Free Software Foundation; either version 2.1 of the License, or (at
+ * your option) any later version.
+ *
+ * StarPU is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
+ *
+ * See the GNU Lesser General Public License in COPYING.LGPL for more details.
+ */
+
+#include <starpu.h>
+#include "../helper.h"
+#include "../vector/memset.h"
+
+#define SIZE (1<<20)
+#define NPARTS 16
+
+/*
+ * Test gathering data on the non-home node with asynchronous partitioning.
+ */
+
+int main(void)
+{
+	int ret;
+	starpu_data_handle_t handle, handles[NPARTS];
+	int i;
+	char d[SIZE];
+	struct starpu_conf conf;
+
+	starpu_conf_init(&conf);
+	starpu_conf_noworker(&conf);
+	conf.ncpus = -1;
+	conf.ncuda = 1;
+	conf.nopencl = 1;
+	conf.nmpi_ms = -1;
+	conf.ntcpip_ms = -1;
+
+	ret = starpu_init(&conf);
+	if (ret == -ENODEV) return STARPU_TEST_SKIPPED;
+	STARPU_CHECK_RETURN_VALUE(ret, "starpu_init");
+	if (starpu_cuda_worker_get_count() == 0 && starpu_opencl_worker_get_count() == 0)
+	{
+		FPRINTF(stderr, "This application requires a GPU\n");
+		starpu_shutdown();
+		return STARPU_TEST_SKIPPED;
+	}
+
+	memset(d, 0, SIZE*sizeof(char));
+	starpu_vector_data_register(&handle, STARPU_MAIN_RAM, (uintptr_t) &d, SIZE, sizeof(char));
+
+	/* Fork */
+	struct starpu_data_filter f =
+	{
+		.filter_func = starpu_vector_filter_block,
+		.nchildren = NPARTS
+	};
+	starpu_data_partition_plan(handle, &f, handles);
+
+	/* We want to operate on a GPU */
+	memset_cl.cpu_funcs[0] = NULL;
+	memset_check_content_cl.cpu_funcs[0] = NULL;
+
+	/* We will work with this GPU */
+	int gpu;
+	int node;
+
+	if (starpu_cuda_worker_get_count() > 0)
+	{
+		/* We have a CUDA GPU, ignore any OpenCL GPU */
+		memset_cl.opencl_funcs[0] = NULL;
+		memset_check_content_cl.opencl_funcs[0] = NULL;
+		gpu = starpu_worker_get_by_type(STARPU_CUDA_WORKER, 0);
+	}
+	else
+	{
+		gpu = starpu_worker_get_by_type(STARPU_OPENCL_WORKER, 0);
+	}
+
+	node = starpu_worker_get_memory_node(gpu);
+
+	/* Prefetch the whole piece on the GPU, we should be able to use this to gather */
+	starpu_data_acquire_on_node(handle, node, STARPU_R);
+	starpu_data_release_on_node(handle, node);
+
+	/* Memset on the pieces on the GPU */
+	for (i = 0; i < NPARTS; i++)
+	{
+		starpu_task_insert(&memset_cl, STARPU_W, handles[i], 0);
+	}
+
+	/* Check that the whole data on the GPU is correct */
+	starpu_task_insert(&memset_check_content_cl, STARPU_R, handle, 0);
+
+	/* Clean */
+	starpu_data_partition_clean(handle, NPARTS, handles);
+
+	starpu_data_unregister(handle);
+
+	starpu_shutdown();
+
+	return 0;
+}