Implement a simple vector add test with alpaka

Implement a simple vector add test with alpaka and cms::alpakatools utilities.

HeterogeneousCore/AlpakaInterface/test/BuildFile.xml

@@ -4,3 +4,10 @@
   <use name="HeterogeneousCore/AlpakaInterface"/>
   <flags ALPAKA_BACKENDS="1"/>
 </bin>
+
+<bin name="alpakaTestKernel" file="alpaka/testKernel.dev.cc">
+  <use name="alpaka"/>
+  <use name="catch2"/>
+  <use name="HeterogeneousCore/AlpakaInterface"/>
+  <flags ALPAKA_BACKENDS="1"/>
+</bin>

HeterogeneousCore/AlpakaInterface/test/alpaka/testKernel.dev.cc

@@ -0,0 +1,93 @@
+#include <random>
+#include <cmath>
+
+#include <alpaka/alpaka.hpp>
+
+#define CATCH_CONFIG_MAIN
+#include <catch.hpp>
+
+#include "HeterogeneousCore/AlpakaInterface/interface/config.h"
+#include "HeterogeneousCore/AlpakaInterface/interface/memory.h"
+#include "HeterogeneousCore/AlpakaInterface/interface/vec.h"
+#include "HeterogeneousCore/AlpakaInterface/interface/workdivision.h"
+
+// each test binary is built for a single Alpaka backend
+using namespace ALPAKA_ACCELERATOR_NAMESPACE;
+
+static constexpr auto s_tag = "[" ALPAKA_TYPE_ALIAS_NAME(alpakaTestKernel) "]";
+
+struct VectorAddKernel {
+  template <typename TAcc, typename T>
+  ALPAKA_FN_ACC void operator()(
+      TAcc const& acc, T const* __restrict__ in1, T const* __restrict__ in2, T* __restrict__ out, size_t size) const {
+    for (auto index : cms::alpakatools::elements_with_stride(acc, size)) {
+      out[index] = in1[index] + in2[index];
+    }
+  }
+};
+
+TEST_CASE("Standard checks of " ALPAKA_TYPE_ALIAS_NAME(alpakaTestKernel), s_tag) {
+  SECTION("Kernel1D") {
+    // get the list of devices on the current platform
+    auto const& devices = cms::alpakatools::devices<Platform>();
+    if (devices.empty()) {
+      std::cout << "No devices on platform\n";
+      return;
+    }
+
+    // random number generator with a gaussian distribution
+    std::random_device rd{};
+    std::mt19937 gen{rd()};
+    std::normal_distribution<> dist{0., 1.};
+
+    // tolerance
+    constexpr float epsilon = 0.0001;
+
+    // buffer size
+    constexpr size_t size = 1024 * 1024;
+
+    // create a default queue on the first device
+    auto const& device = devices[0];
+    auto queue = Queue(device);
+
+    // allocate buffers for the input data on the host and fill them with random data
+    auto in1_h = cms::alpakatools::make_host_buffer<float[]>(queue, size);
+    auto in2_h = cms::alpakatools::make_host_buffer<float[]>(queue, size);
+    alpaka::wait(queue);
+    for (size_t i = 0; i < size; ++i) {
+      in1_h[i] = dist(gen);
+      in2_h[i] = dist(gen);
+    }
+
+    // allocate buffers for the input data on the device
+    auto in1_d = cms::alpakatools::make_device_buffer<float[]>(queue, size);
+    auto in2_d = cms::alpakatools::make_device_buffer<float[]>(queue, size);
+
+    // copy the input data to the device; the size is known from the buffer objects
+    alpaka::memcpy(queue, in1_d, in1_h);
+    alpaka::memcpy(queue, in2_d, in2_h);
+
+    // allocate a buffer for the results on the device, and fill it with zeros
+    auto out_d = cms::alpakatools::make_device_buffer<float[]>(queue, size);
+    alpaka::memset(queue, out_d, 0.);
+
+    // launch the 1-dimensional kernel
+    auto div = cms::alpakatools::make_workdiv<Acc1D>(16, 1024);
+    alpaka::exec<Acc1D>(queue, div, VectorAddKernel{}, in1_d.data(), in2_d.data(), out_d.data(), size);
+
+    // copy the results from the device to the host
+    auto out_h = cms::alpakatools::make_host_buffer<float[]>(queue, size);
+    alpaka::memcpy(queue, out_h, out_d);
+
+    // wait for all the operations to complete
+    alpaka::wait(queue);
+
+    // check the results
+    for (size_t i = 0; i < size; ++i) {
+      //REQUIRE(std::abs(out_h[i] - (in1_h[i] + in2_h[i])) < epsilon);
+      float sum = in1_h[i] + in2_h[i];
+      REQUIRE(out_h[i] < sum + epsilon);
+      REQUIRE(out_h[i] > sum - epsilon);
+    }
+  }
+}