[SYCL][ESIMD] Test fix for proper intrinsic generation

SYCL/ESIMD/regression/iselect.cpp

@@ -0,0 +1,101 @@
+//==--- iselect.cpp - DPC++ ESIMD feature test ---------==//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+// This is basic test for testing proper intrinsics generation.
+
+// REQUIRES: gpu
+// UNSUPPORTED: cuda || hip
+// RUN: %clangxx -fsycl %s -o %t.out
+// RUN: %GPU_RUN_PLACEHOLDER %t.out
+
+#include <sycl/ext/intel/esimd.hpp>
+#include <sycl/sycl.hpp>
+
+#include <iostream>
+#include <typeinfo>
+
+namespace esimd = sycl::ext::intel::esimd;
+template <typename DataT>
+using shared_allocator = sycl::usm_allocator<DataT, sycl::usm::alloc::shared>;
+template <typename DataT>
+using shared_vector = std::vector<DataT, shared_allocator<DataT>>;
+
+int test(sycl::queue queue) {
+
+  shared_allocator<int32_t> allocator1(queue);
+  shared_vector<int32_t> input(16, allocator1);
+  shared_vector<int32_t> output(16, allocator1);
+  shared_allocator<uint16_t> allocator2(queue);
+  shared_vector<uint16_t> offsets(8, allocator2);
+
+  for (int i = 0; i < input.size(); ++i) {
+    input[i] = i;
+  }
+
+  offsets[0] = 3;
+  offsets[1] = 5;
+  offsets[2] = 9;
+  offsets[3] = 2;
+  offsets[4] = 7;
+  offsets[5] = 8;
+  offsets[6] = 0;
+  offsets[7] = 1;
+  {
+    queue
+        .submit([&](sycl::handler &cgh) {
+          int32_t *in_ptr = input.data();
+          int32_t *out_ptr = output.data();
+          uint16_t *off_ptr = offsets.data();
+          cgh.single_task([=]() SYCL_ESIMD_KERNEL {
+            __ESIMD_NS::simd<int32_t, 16> input_load_vec;
+            input_load_vec.copy_from(in_ptr);
+
+            esimd::simd<uint16_t, 8> offset;
+            offset.copy_from(off_ptr);
+
+            esimd::simd<int, 8> data = input_load_vec.iselect(offset);
+            data += 3;
+            input_load_vec.template iupdate(offset, data,
+                                            esimd::simd_mask<8>(1));
+            input_load_vec.copy_to(out_ptr);
+          });
+        })
+        .wait_and_throw();
+  }
+  int32_t expected_values[] = {3,  4,  5,  6,  4,  8,  6,  10,
+                               11, 12, 10, 11, 12, 13, 14, 15};
+  int error = 0;
+  for (int i = 0; i < 16; i++) {
+    if (output[i] != expected_values[i]) {
+      ++error;
+      std::cout << "Output[" << i << "] = " << output[i] << ", expected "
+                << expected_values[i] << std::endl;
+    }
+  }
+
+  return error;
+}
+
+int main(int, char **) {
+  sycl::queue queue;
+
+  auto dev = queue.get_device();
+  std::cout << "Running on " << dev.get_info<sycl::info::device::name>()
+            << "\n";
+
+  int test_result = 0;
+
+  test_result = test(queue);
+
+  if (test_result == 0) {
+    std::cout << "Pass!!" << std::endl;
+  } else {
+    std::cout << "Fail!!" << std::endl;
+  }
+
+  return test_result;
+}