COB-61: fixed test

RTCP/Cobalt/GPUProc/test/t_gpu_utils.cc

@@ -61,99 +61,99 @@ TEST_FIXTURE(CreateFixture, CreatePtx)
 {
   createPTX(srcFile);
 }
-//
-// TEST_FIXTURE(CreateFixture, CreatePtxExtraDef)
-// {
-//   CompileDefinitions defs;
-//   defs["FOO"] = "42";
-//   createPTX(srcFile, defs);
-// }
-//
-// TEST_FIXTURE(CreateFixture, CreatePtxWrongExtraDef)
-// {
-//   CompileDefinitions defs;
-//   defs["FOO"] = "24";
-//   CHECK_THROW(createPTX(srcFile, defs), GPUProcException);
-// }
-//
-// TEST_FIXTURE(CreateFixture, CreatePtxExtraFlag)
-// {
-//   CompileFlags flags;
-//   flags.insert("--source-in-ptx");
-//   createPTX(srcFile, defaultCompileDefinitions(), flags);
-// }
-//
-// TEST_FIXTURE(CreateFixture, CreatePtxWrongExtraFlag)
-// {
-//   CompileFlags flags;
-//   flags.insert("--yaddayadda");
-//   CHECK_THROW(createPTX(srcFile, defaultCompileDefinitions(), flags),
-//               GPUProcException);
-// }
-//
-// TEST_FIXTURE(CreateFixture, CreateModule)
-// {
-//   gpu::Device device(gpu::Platform().devices()[0]);
-//   createModule(gpu::Context(device), srcFile, createPTX(srcFile));
-// }
-//
-// TEST_FIXTURE(CreateFixture, CreateModuleHighestArch)
-// {
-//   // Highest known architecture is 3.5.
-//   // Only perform this test if we do NOT have a device with that capability.
-//   gpu::Device device(gpu::Platform().devices()[0]);
-//
-//   if (device.getComputeCapabilityMajor() > 3) return;
-//   if (device.getComputeCapabilityMajor() == 3 &&
-//       device.getComputeCapabilityMinor() >= 5) return;
-//
-//   CHECK_THROW(createModule(gpu::Context(device),
-//                            srcFile,
-//                            createPTX(srcFile,
-//                                      defaultCompileDefinitions(),
-//                                      defaultCompileFlags(),
-//                                      vector<gpu::Device>())),
-//               gpu::GPUException);
-// }
-//
-// TEST(DumpBuffer)
-// {
-//   typedef unsigned element_type;
-//   const char* dumpFile("tDevMem.dat");
-//   const size_t num_elements(1024);
-//   const size_t num_bytes(num_elements * sizeof(element_type));
-//   vector<element_type> input(num_elements), output(num_elements);
-//
-//   // Initialize input vector
-//   for(size_t i = 0; i < num_elements; i++)
-//     input[i] = element_type(i);
-//
-//   // Initialize GPU device, context, and stream.
-//   gpu::Device device(gpu::Platform().devices()[0]);
-//   gpu::Context ctx(device);
-//   gpu::Stream stream(ctx);
-//
-//   // Allocate memory on host and device
-//   gpu::HostMemory hostMem(ctx, num_bytes);
-//   gpu::DeviceMemory devMem(ctx, num_bytes);
-//
-//   // Copy input vector to host memory
-//   copy(input.begin(), input.end(), hostMem.get<element_type>());
-//
-//   // Transfer to device memory
-//   stream.writeBuffer(devMem, hostMem);
-//
-//   // Dump device memory to file
-//   dumpBuffer(devMem, dumpFile);
-//
-//   // Read raw data back from file into output vector and remove file
-//   ifstream ifs(dumpFile, ios::binary);
-//   ifs.read(reinterpret_cast<char*>(&output[0]), num_bytes);
-//   remove(dumpFile);
-//
-//   // Compare input and output vector.
-//   CHECK_ARRAY_EQUAL(input, output, num_elements);
-// }
+
+TEST_FIXTURE(CreateFixture, CreatePtxExtraDef)
+{
+  CompileDefinitions defs;
+  defs["FOO"] = "42";
+  createPTX(srcFile, defs);
+}
+
+TEST_FIXTURE(CreateFixture, CreatePtxWrongExtraDef)
+{
+  CompileDefinitions defs;
+  defs["FOO"] = "24";
+  CHECK_THROW(createPTX(srcFile, defs), GPUProcException);
+}
+
+TEST_FIXTURE(CreateFixture, CreatePtxExtraFlag)
+{
+  CompileFlags flags;
+  flags.insert("--source-in-ptx");
+  createPTX(srcFile, defaultCompileDefinitions(), flags);
+}
+
+TEST_FIXTURE(CreateFixture, CreatePtxWrongExtraFlag)
+{
+  CompileFlags flags;
+  flags.insert("--yaddayadda");
+  CHECK_THROW(createPTX(srcFile, defaultCompileDefinitions(), flags),
+              GPUProcException);
+}
+
+TEST_FIXTURE(CreateFixture, CreateModule)
+{
+  gpu::Device device(gpu::Platform().devices()[0]);
+  createModule(gpu::Context(device), srcFile, createPTX(srcFile));
+}
+
+TEST_FIXTURE(CreateFixture, CreateModuleHighestArch)
+{
+  // Highest known architecture is 3.5.
+  // Only perform this test if we do NOT have a device with that capability.
+  gpu::Device device(gpu::Platform().devices()[0]);
+
+  if (device.getComputeCapabilityMajor() > 3) return;
+  if (device.getComputeCapabilityMajor() == 3 &&
+      device.getComputeCapabilityMinor() >= 5) return;
+
+  CHECK_THROW(createModule(gpu::Context(device),
+                           srcFile,
+                           createPTX(srcFile,
+                                     defaultCompileDefinitions(),
+                                     defaultCompileFlags(),
+                                     vector<gpu::Device>())),
+              gpu::GPUException);
+}
+
+TEST(DumpBuffer)
+{
+  typedef unsigned element_type;
+  const char* dumpFile("tDevMem.dat");
+  const size_t num_elements(1024);
+  const size_t num_bytes(num_elements * sizeof(element_type));
+  vector<element_type> input(num_elements), output(num_elements);
+
+  // Initialize input vector
+  for(size_t i = 0; i < num_elements; i++)
+    input[i] = element_type(i);
+
+  // Initialize GPU device, context, and stream.
+  gpu::Device device(gpu::Platform().devices()[0]);
+  gpu::Context ctx(device);
+  gpu::Stream stream(ctx);
+
+  // Allocate memory on host and device
+  gpu::HostMemory hostMem(ctx, num_bytes);
+  gpu::DeviceMemory devMem(ctx, num_bytes);
+
+  // Copy input vector to host memory
+  copy(input.begin(), input.end(), hostMem.get<element_type>());
+
+  // Transfer to device memory
+  stream.writeBuffer(devMem, hostMem);
+
+  // Dump device memory to file
+  dumpBuffer(devMem, dumpFile);
+
+  // Read raw data back from file into output vector and remove file
+  ifstream ifs(dumpFile, ios::binary);
+  ifs.read(reinterpret_cast<char*>(&output[0]), num_bytes);
+  remove(dumpFile);
+
+  // Compare input and output vector.
+  CHECK_ARRAY_EQUAL(input, output, num_elements);
+}
 
 int main()
 {