PerformanceAnalysis/test/test_ad.cpp at master · CODARcode/PerformanceAnalysis · GitHub

History

executable file

412 lines (339 loc) · 12.7 KB

Raw

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

#include<chimbuko_config.h>

#ifdef USE_MPI

#include <mpi.h>

#endif

#include "chimbuko/modules/performance_analysis/chimbuko.hpp"

#include <gtest/gtest.h>

#include <chrono>

#include <thread>

using namespace chimbuko;

using namespace chimbuko::modules::performance_analysis;

#ifdef USE_MPI

class ADTest : public ::testing::Test

{

protected:

int world_rank, world_size;

virtual void SetUp()

{

MPI_Comm_rank(MPI_COMM_WORLD, &world_rank);

MPI_Comm_size(MPI_COMM_WORLD, &world_size);

}

virtual void TearDown()

{

}

template <typename K, typename V>

void show_map(const std::unordered_map<K, V>& m)

{

for (const auto& it : m)

{

std::cout << it.first << " : " << it.second << std::endl;

}

};

TEST_F(ADTest, EnvTest)

{

EXPECT_EQ(4, world_size);

}

TEST_F(ADTest, BpfileTest)

{

const std::vector<int> N_STEPS{7, 7, 7, 7};

//Below indexed by [rank][step] for 4 ranks

const std::vector<std::vector<size_t>> N_FUNC{

{57077, 51845, 60561, 63278, 64628, 66484, 42233},

{41215, 51375, 56620, 57683, 58940, 61010, 41963},

{41108, 50820, 57237, 56590, 63458, 63931, 42486},

{40581, 51127, 59175, 58158, 60465, 62516, 41238}

};

const std::vector<std::vector<size_t>> N_COMM{

{89349, 107207, 121558, 123381, 128682, 131611, 83326},

{78250, 94855 , 106301, 107222, 111581, 114273, 77817},

{77020, 93355 , 105238, 106608, 114192, 118530, 79135},

{77332, 93027 , 107856, 107628, 112019, 116468, 74517}

};

const std::vector<std::vector<unsigned long>> N_OUTLIERS{

{137, 82, 94, 95, 90, 91, 51},

{90, 60, 78, 75, 54, 68, 37},

{94, 79, 80, 87, 92, 92, 47},

{93, 81, 100, 81, 78, 85, 61}

};

ChimbukoParams params;

params.base_params.rank = world_rank;

params.trace_data_dir = "./data";

params.trace_inputFile = "tau-metrics-" + std::to_string(world_rank) + ".bp";

params.trace_engineType = "BPFile";

params.base_params.prov_outputpath = "./temp"; //have to output somewhere if provdb not in use

params.base_params.algo_params.sstd_sigma = 6.0;

params.base_params.algo_params.algorithm = "sstd";

params.base_params.only_one_frame = true; //just analyze first IO frame

params.anom_win_size = 0;

Chimbuko driver;

int step;

unsigned long n_outliers = 0, frames = 0;

unsigned long long n_func_events = 0, n_comm_events = 0, n_counter_events = 0;

try{

driver.initialize(params);

}catch(const std::exception &e){

std::cerr << "Caught exception during driver init: " << e.what() << std::endl;

throw e;

}

std::vector<std::vector<unsigned long> > N_FUNC_ACTUAL(4);

std::vector<std::vector<unsigned long> > N_COMM_ACTUAL(4);

std::vector<std::vector<unsigned long> > N_OUTLIERS_ACTUAL(4);

while ( driver.get_status() )

{

try{

frames = driver.run();

}catch(const std::exception &e){

std::cerr << "Caught exception in driver run: " << e.what() << std::endl;

throw e;

}

if(frames == 0) break; //last call to run will not analyze any frames

//if (driver.getStep() == -1)

// break;

step = driver.getStep()-1; //this is the index of the step just completed; the current step is the one after!

std::cout << "Checking results for step " << step << std::endl;

//FIXME: Add an expect for the counter events

EXPECT_EQ(N_FUNC[world_rank][step], driver.getRunStats().n_func_events);

EXPECT_EQ(N_COMM[world_rank][step], driver.getRunStats().n_comm_events);

EXPECT_EQ(N_OUTLIERS[world_rank][step], driver.getBaseRunStats().n_outliers);

N_FUNC_ACTUAL[world_rank].push_back(driver.getRunStats().n_func_events);

N_COMM_ACTUAL[world_rank].push_back(driver.getRunStats().n_comm_events);

N_OUTLIERS_ACTUAL[world_rank].push_back(driver.getBaseRunStats().n_outliers);

}

int nstep = step+1;

EXPECT_EQ(N_STEPS[world_rank], nstep);

if(!world_rank) std::cout << "Steps performed: " << nstep << std::endl;

MPI_Barrier(MPI_COMM_WORLD);

//For gathering actual event counts, check

int nsteps_check[4] = {0,0,0,0};

nsteps_check[world_rank] = nstep;

MPI_Allreduce(MPI_IN_PLACE, nsteps_check, 4, MPI_INT, MPI_SUM, MPI_COMM_WORLD);

bool fail = false;

for(int i=0;i<4;i++)

if(nsteps_check[i] != nstep){

fail = true;

break;

}

if(!fail){

//Fill in zeroes for other ranks

for(int i=0;i<4;i++){

if(i!=world_rank){

N_FUNC_ACTUAL[i].resize(nstep, 0);

N_COMM_ACTUAL[i].resize(nstep, 0);

N_OUTLIERS_ACTUAL[i].resize(nstep, 0);

}

//Use global sum to sync all data to root

for(int i=0;i<4;i++){

MPI_Allreduce(MPI_IN_PLACE, N_FUNC_ACTUAL[i].data(), nstep, MPI_UNSIGNED_LONG, MPI_SUM, MPI_COMM_WORLD);

MPI_Allreduce(MPI_IN_PLACE, N_COMM_ACTUAL[i].data(), nstep, MPI_UNSIGNED_LONG, MPI_SUM, MPI_COMM_WORLD);

MPI_Allreduce(MPI_IN_PLACE, N_OUTLIERS_ACTUAL[i].data(), nstep, MPI_UNSIGNED_LONG, MPI_SUM, MPI_COMM_WORLD);

}

if(world_rank == 0){

std::cout << "Actual value tables:" << std::endl;

std::cout << "FUNC:" << std::endl;

for(int i=0;i<4;i++){

for(int j=0;j<nstep;j++)

std::cout << N_FUNC_ACTUAL[i][j] << (j==nstep-1 ? "" :", ");

std::cout << std::endl;

}

std::cout << "COMM:" << std::endl;

for(int i=0;i<4;i++){

for(int j=0;j<nstep;j++)

std::cout << N_COMM_ACTUAL[i][j] << (j==nstep-1 ? "" :", ");

std::cout << std::endl;

}

std::cout << "OUTLIERS:" << std::endl;

for(int i=0;i<4;i++){

for(int j=0;j<nstep;j++)

std::cout << N_OUTLIERS_ACTUAL[i][j] << (j==nstep-1 ? "" :", ");

std::cout << std::endl;

}

//Check an end-to-end run without breaking up into steps

driver.finalize();

params.base_params.only_one_frame = false;

driver.initialize(params);

try{

frames = driver.run();

}catch(const std::exception &e){

std::cerr << "Caught exception in driver run: " << e.what() << std::endl;

throw e;

}

EXPECT_EQ(N_STEPS[world_rank], frames);

unsigned long func_tot = 0, comm_tot=0, outlier_tot = 0;

for(int step=0;step<N_STEPS[world_rank];step++){

func_tot += N_FUNC[world_rank][step];

comm_tot += N_COMM[world_rank][step];

outlier_tot += N_OUTLIERS[world_rank][step];

}

EXPECT_EQ(func_tot, driver.getRunStats().n_func_events);

EXPECT_EQ(comm_tot, driver.getRunStats().n_comm_events);

EXPECT_EQ(outlier_tot, driver.getBaseRunStats().n_outliers);

}

TEST_F(ADTest, BpfileWithNetTest)

{

const std::vector<int> N_STEPS{7, 7, 7, 7};

const std::vector<std::vector<size_t>> N_FUNC{

{57077, 51845, 60561, 63278, 64628, 66484, 42233},

{41215, 51375, 56620, 57683, 58940, 61010, 41963},

{41108, 50820, 57237, 56590, 63458, 63931, 42486},

{40581, 51127, 59175, 58158, 60465, 62516, 41238}

}; //[rank][step]

const std::vector<std::vector<size_t>> N_COMM{

{89349, 107207, 121558, 123381, 128682, 131611, 83326},

{78250, 94855, 106301, 107222, 111581, 114273, 77817},

{77020, 93355, 105238, 106608, 114192, 118530, 79135},

{77332, 93027, 107856, 107628, 112019, 116468, 74517}

}; //[rank][step]

const std::vector<std::vector<unsigned long>> N_OUTLIERS{

{137, 52, 67, 71, 64, 74, 30},

{102, 64, 85, 80, 55, 61, 33},

{76, 50, 56, 63, 83, 76, 52},

{77, 55, 94, 61, 49, 68, 33}

}; //[rank][step]

ChimbukoParams params;

params.base_params.rank = world_rank;

params.base_params.verbose = world_rank == 0;

params.trace_data_dir = "./data";

params.trace_inputFile = "tau-metrics-" + std::to_string(world_rank) + ".bp";

params.trace_engineType = "BPFile";

params.base_params.prov_outputpath = "./temp";

params.anom_win_size = 0;

params.base_params.pserver_addr = "tcp://localhost:5559"; //connect to the pserver

#ifdef _PERF_METRIC

params.base_params.perf_outputpath = "./perf";

params.base_params.perf_step = 1;

#endif

params.base_params.algo_params.sstd_sigma = 6.0;

params.base_params.algo_params.algorithm = "sstd";

params.base_params.only_one_frame = true;

Chimbuko driver;

int step;

unsigned long n_outliers = 0, frames = 0;

unsigned long long n_func_events = 0, n_comm_events = 0, n_counter_events = 0;

try{

driver.initialize(params);

}catch(const std::exception &e){

std::cerr << "Caught exception during driver init: " << e.what() << std::endl;

throw e;

}

std::vector<std::vector<unsigned long> > N_FUNC_ACTUAL(4);

std::vector<std::vector<unsigned long> > N_COMM_ACTUAL(4);

std::vector<std::vector<unsigned long> > N_OUTLIERS_ACTUAL(4);

while ( driver.get_status() )

{

// for the test purpose, we serialize the excution for run()

int token = 1;

if (world_rank != 0) {

MPI_Recv(&token, 1, MPI_INT, world_rank-1, 0, MPI_COMM_WORLD, MPI_STATUS_IGNORE);

}

try{

frames = driver.run();

}catch(const std::exception &e){

std::cerr << "Caught exception in driver run: " << e.what() << std::endl;

throw e;

}

MPI_Send(&token, 1, MPI_INT, (world_rank+1)%world_size, 0, MPI_COMM_WORLD);

if (world_rank == 0) {

MPI_Recv(&token, 1, MPI_INT, world_size-1, 0, MPI_COMM_WORLD, MPI_STATUS_IGNORE);

}

if (frames > 0)

{

step = driver.getStep() - 1; //the step that just completed; the current value is for the next step

if(!world_rank) std::cout << "Checking data for step " << step << std::endl;

//FIXME: Add expect for counter events

EXPECT_EQ(N_FUNC[world_rank][step], driver.getRunStats().n_func_events);

EXPECT_EQ(N_COMM[world_rank][step], driver.getRunStats().n_comm_events);

EXPECT_EQ(N_OUTLIERS[world_rank][step], driver.getBaseRunStats().n_outliers);

N_FUNC_ACTUAL[world_rank].push_back(driver.getRunStats().n_func_events);

N_COMM_ACTUAL[world_rank].push_back(driver.getRunStats().n_comm_events);

N_OUTLIERS_ACTUAL[world_rank].push_back(driver.getBaseRunStats().n_outliers);

}

int nstep = step+1;

EXPECT_EQ(N_STEPS[world_rank], nstep);

if(!world_rank) std::cout << "Steps performed: " << nstep << std::endl;

MPI_Barrier(MPI_COMM_WORLD);

//For gathering actual event counts, check

int nsteps_check[4] = {0,0,0,0};

nsteps_check[world_rank] = nstep;

MPI_Allreduce(MPI_IN_PLACE, nsteps_check, 4, MPI_INT, MPI_SUM, MPI_COMM_WORLD);

bool fail = false;

for(int i=0;i<4;i++)

if(nsteps_check[i] != nstep){

fail = true;

break;

}

if(!fail){

//Fill in zeroes for other ranks

for(int i=0;i<4;i++){

if(i!=world_rank){

N_FUNC_ACTUAL[i].resize(nstep, 0);

N_COMM_ACTUAL[i].resize(nstep, 0);

N_OUTLIERS_ACTUAL[i].resize(nstep, 0);

}

//Use global sum to sync all data to root

for(int i=0;i<4;i++){

MPI_Allreduce(MPI_IN_PLACE, N_FUNC_ACTUAL[i].data(), nstep, MPI_UNSIGNED_LONG, MPI_SUM, MPI_COMM_WORLD);

MPI_Allreduce(MPI_IN_PLACE, N_COMM_ACTUAL[i].data(), nstep, MPI_UNSIGNED_LONG, MPI_SUM, MPI_COMM_WORLD);

MPI_Allreduce(MPI_IN_PLACE, N_OUTLIERS_ACTUAL[i].data(), nstep, MPI_UNSIGNED_LONG, MPI_SUM, MPI_COMM_WORLD);

}

if(world_rank == 0){

std::cout << "Actual value tables:" << std::endl;

std::cout << "FUNC:" << std::endl;

for(int i=0;i<4;i++){

for(int j=0;j<nstep;j++)

std::cout << N_FUNC_ACTUAL[i][j] << (j==nstep-1 ? "" :", ");

std::cout << std::endl;

}

std::cout << "COMM:" << std::endl;

for(int i=0;i<4;i++){

for(int j=0;j<nstep;j++)

std::cout << N_COMM_ACTUAL[i][j] << (j==nstep-1 ? "" :", ");

std::cout << std::endl;

}

std::cout << "OUTLIERS:" << std::endl;

for(int i=0;i<4;i++){

for(int j=0;j<nstep;j++)

std::cout << N_OUTLIERS_ACTUAL[i][j] << (j==nstep-1 ? "" :", ");

std::cout << std::endl;

}

TEST_F(ADTest, ReadExecDataTest)

{

//TODO: Update to check provenance data dump

#if 0

const std::vector<int> N_STEPS{6, 6, 6, 6};

const std::vector<std::vector<unsigned long>> N_OUTLIERS{

{137, 52, 67, 71, 64, 74, 30},

{102, 64, 85, 80, 55, 61, 33},

{76, 50, 56, 63, 83, 76, 52},

{77, 55, 94, 61, 49, 68, 33}

}; //[rank][step]

std::string execOutput = "temp/0/" + std::to_string(world_rank);

for (int i = 0; i < N_STEPS[world_rank]; i++)

{

std::ifstream f;

f.open(execOutput + "/" + std::to_string(i) + ".json");

ASSERT_TRUE(f.is_open());

nlohmann::json j;

std::string s;

f.seekg(0, std::ios::end);

s.reserve(f.tellg());

f.seekg(0, std::ios::beg);

s.assign(std::istreambuf_iterator<char>(f), std::istreambuf_iterator<char>());

j = nlohmann::json::parse(s);

// if (world_rank == 0 && i == 0)

// {

// std::cout << j.dump(2) << std::endl;

// }

EXPECT_EQ(N_OUTLIERS[world_rank][i], j["exec"].size());

}

#endif

}

#endif

PerformanceAnalysis/test/test_ad.cpp at master · CODARcode/PerformanceAnalysis · GitHub

Uh oh!

Expand file tree

Breadcrumbs

test_ad.cpp

Latest commit

History

Breadcrumbs

test_ad.cpp

File metadata and controls