Fix round robin sharding (#121022)

Fix round robin sharding when there are no test times and sort_by_time=False

Adds more tests to test_test_selections for sort_by_time=False
Adds more checks to test_split_shards_random for serial/parallel ordering + ordering of tests
Refactoring of dup code

Tested locally by running `python test/run_test.py --shard 3 5` with no test times downloaded and checked that it wasn't an empty list.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/121022
Approved by: https://github.com/huydhn, https://github.com/osalpekar
This commit is contained in:
Catherine Lee
2024-03-08 17:01:34 +00:00
committed by PyTorch MergeBot
parent 9d83f9dc0e
commit effdea5fc6
3 changed files with 219 additions and 79 deletions

View File

@ -1,3 +1,4 @@
import functools
import pathlib
import random
import sys
@ -72,6 +73,109 @@ class TestCalculateShards(unittest.TestCase):
self.assertAlmostEqual(expected[0], actual[0])
self.assertListEqual(expected[1], actual[1])
def test_no_times(self) -> None:
# Check that round robin sharding is used when no times are provided
expected_shards = [
(
0.0,
[
ShardedTest(
test="super_long_test", shard=1, num_shards=1, time=None
),
ShardedTest(test="long_test2", shard=1, num_shards=1, time=None),
ShardedTest(test="normal_test2", shard=1, num_shards=1, time=None),
ShardedTest(test="short_test1", shard=1, num_shards=1, time=None),
ShardedTest(test="short_test3", shard=1, num_shards=1, time=None),
ShardedTest(test="short_test5", shard=1, num_shards=1, time=None),
],
),
(
0.0,
[
ShardedTest(test="long_test1", shard=1, num_shards=1, time=None),
ShardedTest(test="normal_test1", shard=1, num_shards=1, time=None),
ShardedTest(test="normal_test3", shard=1, num_shards=1, time=None),
ShardedTest(test="short_test2", shard=1, num_shards=1, time=None),
ShardedTest(test="short_test4", shard=1, num_shards=1, time=None),
],
),
]
self.assert_shards_equal(
expected_shards,
calculate_shards(2, self.tests, {}, {}, sort_by_time=False),
)
def test_some_times_with_not_sort_by_time(self) -> None:
expected_shards = [
(
300.0,
[
ShardedTest(test="test_1", shard=1, num_shards=1, time=None),
ShardedTest(test="test_3", shard=1, num_shards=1, time=300),
ShardedTest(test="test_4", shard=1, num_shards=1, time=None),
],
),
(
400.0,
[
ShardedTest(test="test_2", shard=1, num_shards=1, time=400),
ShardedTest(test="test_5", shard=1, num_shards=1, time=None),
],
),
]
self.assert_shards_equal(
expected_shards,
calculate_shards(
2,
[
TestRun("test_1"),
TestRun("test_2"),
TestRun("test_3"),
TestRun("test_4"),
TestRun("test_5"),
],
{"test_2": 400, "test_3": 300},
{},
sort_by_time=False,
),
)
def test_serial_parallel_interleaving(self) -> None:
expected_shards = [
(
400.0,
[
ShardedTest(test="test_1", shard=1, num_shards=1, time=None),
ShardedTest(test="test_4", shard=1, num_shards=1, time=None),
ShardedTest(test="test_5", shard=1, num_shards=1, time=None),
ShardedTest(test="test_2", shard=1, num_shards=1, time=400),
],
),
(
300.0,
[
ShardedTest(test="test_3", shard=1, num_shards=1, time=300),
],
),
]
self.assert_shards_equal(
expected_shards,
calculate_shards(
2,
[
TestRun("test_1"),
TestRun("test_2"),
TestRun("test_3"),
TestRun("test_4"),
TestRun("test_5"),
],
{"test_2": 400, "test_3": 300},
{},
must_serial=lambda x: x in ["test_1", "test_3"],
sort_by_time=False,
),
)
def test_calculate_2_shards_with_complete_test_times(self) -> None:
expected_shards = [
(
@ -213,8 +317,7 @@ class TestCalculateShards(unittest.TestCase):
22.0,
[
ShardedTest(test="long_test1", shard=1, num_shards=1, time=22),
ShardedTest(test="normal_test2", shard=1, num_shards=1, time=None),
ShardedTest(test="short_test5", shard=1, num_shards=1, time=None),
ShardedTest(test="short_test2", shard=1, num_shards=1, time=None),
],
),
(
@ -228,7 +331,8 @@ class TestCalculateShards(unittest.TestCase):
1.0,
[
ShardedTest(test="short_test1", shard=1, num_shards=1, time=1),
ShardedTest(test="short_test2", shard=1, num_shards=1, time=None),
ShardedTest(test="normal_test2", shard=1, num_shards=1, time=None),
ShardedTest(test="short_test5", shard=1, num_shards=1, time=None),
],
),
(
@ -330,23 +434,27 @@ class TestCalculateShards(unittest.TestCase):
for _ in range(100):
num_shards = random.randint(1, 10)
num_tests = random.randint(1, 100)
test_names = [str(i) for i in range(num_tests)]
tests = [TestRun(x) for x in test_names]
serial = [x for x in test_names if random.randint(0, 1) == 0]
has_times = [x for x in test_names if random.randint(0, 1) == 0]
random_times: Dict[str, float] = {
str(i): random.randint(0, THRESHOLD * 10) for i in range(num_tests)
i: random.randint(0, THRESHOLD * 10) for i in has_times
}
serial = [str(i) for i in range(num_tests) if random.randint(0, 1) == 0]
sort_by_time = random.randint(0, 1) == 0
shards = calculate_shards(
num_shards,
[TestRun(t) for t in random_times.keys()],
tests,
random_times,
None,
must_serial=lambda x: x in serial,
sort_by_time=random.randint(0, 1) == 0,
sort_by_time=sort_by_time,
)
times = [x[0] for x in shards]
max_diff = max(times) - min(times)
self.assertTrue(max_diff <= THRESHOLD)
self.assertTrue(max_diff <= THRESHOLD + (num_tests - len(has_times)) * 60)
all_sharded_tests: Dict[str, List[ShardedTest]] = defaultdict(list)
for _, sharded_tests in shards:
@ -354,19 +462,58 @@ class TestCalculateShards(unittest.TestCase):
all_sharded_tests[sharded_test.name].append(sharded_test)
# Check that all test files are represented in the shards
self.assertListEqual(
sorted(random_times.keys()), sorted(all_sharded_tests.keys())
)
self.assertListEqual(sorted(test_names), sorted(all_sharded_tests.keys()))
# Check that for each test file, the pytest shards' times adds up to
# original and all shards are present
for test, sharded_tests in all_sharded_tests.items():
if random_times.get(test) is None:
self.assertTrue(len(sharded_tests) == 1)
self.assertTrue(sharded_tests[0].time is None)
else:
# x.time is not None because of the above check
self.assertAlmostEqual(
random_times[test], sum(x.time or 0 for x in sharded_tests)
random_times[test], sum(x.time for x in sharded_tests) # type: ignore[misc]
)
self.assertListEqual(
list(range(sharded_tests[0].num_shards)),
sorted(x.shard - 1 for x in sharded_tests),
)
# Tests without times are serial
true_serial = set(serial + [x for x in test_names if x not in has_times])
# Check that sort_by_time is respected
if sort_by_time:
def comparator(a: ShardedTest, b: ShardedTest) -> int:
# serial comes first
if a.name in true_serial and b.name not in true_serial:
return -1
if a.name not in true_serial and b.name in true_serial:
return 1
# known test times come first
if a.time is not None and b.time is None:
return -1
if a.time is None and b.time is not None:
return 1
if a.time == b.time:
return 0
# not None due to the above checks
return -1 if a.time > b.time else 1 # type: ignore[operator]
else:
def comparator(a: ShardedTest, b: ShardedTest) -> int:
# serial comes first
if a.name in true_serial and b.name not in true_serial:
return -1
if a.name not in true_serial and b.name in true_serial:
return 1
return test_names.index(a.name) - test_names.index(b.name)
for _, sharded_tests in shards:
self.assertListEqual(
sorted(sharded_tests, key=functools.cmp_to_key(comparator)),
sharded_tests,
)
def test_calculate_2_shards_against_optimal_shards(self) -> None:
random.seed(120)

View File

@ -293,8 +293,8 @@ class ShardedTest:
def __str__(self) -> str:
return f"{self.test} {self.shard}/{self.num_shards}"
def get_time(self) -> float:
return self.time or 0
def get_time(self, default: float = 0) -> float:
return self.time if self.time is not None else default
def get_pytest_args(self) -> List[str]:
filter = self.test.get_pytest_filter()

View File

@ -20,6 +20,7 @@ IS_ROCM = os.path.exists("/opt/rocm")
NUM_PROCS = 1 if IS_MEM_LEAK_CHECK else 2
NUM_PROCS_FOR_SHARDING_CALC = NUM_PROCS if not IS_ROCM or IS_MEM_LEAK_CHECK else 2
THRESHOLD = 60 * 10 # 10 minutes
DEFAULT_TIME = 60 # if no test times available for the test, assume it takes 60s
# See Note [ROCm parallel CI testing]
# Special logic for ROCm GHA runners to query number of GPUs available.
@ -48,12 +49,13 @@ class ShardJob:
self.serial: List[ShardedTest] = []
self.parallel: List[ShardedTest] = []
def get_total_time(self) -> float:
def get_total_time(self, default: float = 0.0) -> float:
"""Default is the value for which to substitute if a test has no time"""
procs = [0.0 for _ in range(NUM_PROCS_FOR_SHARDING_CALC)]
for test in self.parallel:
min_index = procs.index(min(procs))
procs[min_index] += test.get_time()
time = max(procs) + sum(test.get_time() for test in self.serial)
procs[min_index] += test.get_time(default)
time = max(procs) + sum(test.get_time(default) for test in self.serial)
return time
def convert_to_tuple(self) -> Tuple[float, List[ShardedTest]]:
@ -86,6 +88,8 @@ def get_duration(
test_file_times: Dict[str, float],
test_class_times: Dict[str, Dict[str, float]],
) -> Optional[float]:
"""Calculate the time for a TestRun based on the given test_file_times and
test_class_times. Returns None if the time is unknown."""
file_duration = test_file_times.get(test.test_file, None)
if test.is_full_file():
return file_duration
@ -123,65 +127,46 @@ def get_duration(
def shard(
sharded_jobs: List[ShardJob],
tests: Sequence[TestRun],
test_file_times: Dict[str, float],
test_class_times: Dict[str, Dict[str, float]],
pytest_sharded_tests: Sequence[ShardedTest],
estimated_time_limit: Optional[float] = None,
sort_by_time: bool = True,
serial: bool = False,
) -> None:
if len(sharded_jobs) == 0:
assert len(tests) == 0, "No shards provided but there are tests to shard"
return
# Modifies sharded_jobs in place
known_tests = tests
unknown_tests = []
if sort_by_time:
known_tests = [
x
for x in tests
if get_duration(x, test_file_times, test_class_times) is not None
]
unknown_tests = [x for x in tests if x not in known_tests]
if len(sharded_jobs) == 0:
assert (
unknown_tests == [] or serial
), f"Attmempting to parallelize unknown tests {unknown_tests}"
del tests
len(pytest_sharded_tests) == 0
), "No shards provided but there are tests to shard"
return
known_tests = get_with_pytest_shard(known_tests, test_file_times, test_class_times)
def _get_min_sharded_job(sharded_jobs: List[ShardJob]) -> ShardJob:
return min(sharded_jobs, key=lambda j: j.get_total_time(default=DEFAULT_TIME))
if sort_by_time:
known_tests = sorted(known_tests, key=lambda j: j.get_time(), reverse=True)
def _shard_serial(tests: List[ShardedTest], sharded_jobs: List[ShardJob]) -> None:
def _shard_serial(
tests: Sequence[ShardedTest], sharded_jobs: List[ShardJob]
) -> None:
assert estimated_time_limit is not None, "Estimated time limit must be provided"
new_sharded_jobs = sharded_jobs
for test in tests:
if (
len(sharded_jobs) > 1
and sharded_jobs[-1].get_total_time() > estimated_time_limit
and sharded_jobs[-1].get_total_time(default=DEFAULT_TIME)
> estimated_time_limit
):
new_sharded_jobs = sharded_jobs[:-1]
min_sharded_job = min(new_sharded_jobs, key=lambda j: j.get_total_time())
min_sharded_job = _get_min_sharded_job(new_sharded_jobs)
min_sharded_job.serial.append(test)
def _shard_parallel(tests: List[ShardedTest], sharded_jobs: List[ShardJob]) -> None:
def _shard_parallel(
tests: Sequence[ShardedTest], sharded_jobs: List[ShardJob]
) -> None:
for test in tests:
min_sharded_job = min(sharded_jobs, key=lambda j: j.get_total_time())
min_sharded_job = _get_min_sharded_job(sharded_jobs)
min_sharded_job.parallel.append(test)
if serial:
_shard_serial(known_tests, sharded_jobs)
_shard_serial(pytest_sharded_tests, sharded_jobs)
else:
_shard_parallel(known_tests, sharded_jobs)
# Round robin the unknown jobs starting with the smallest shard
num_shards = len(sharded_jobs)
index = min(range(num_shards), key=lambda i: sharded_jobs[i].get_total_time())
for unknown_test in unknown_tests:
sharded_jobs[index].serial.append(ShardedTest(unknown_test, 1, 1, None))
index = (index + 1) % num_shards
_shard_parallel(pytest_sharded_tests, sharded_jobs)
return
@ -196,22 +181,36 @@ def calculate_shards(
) -> List[Tuple[float, List[ShardedTest]]]:
must_serial = must_serial or (lambda x: True)
test_class_times = test_class_times or {}
# Divide tests into pytest shards
if sort_by_time:
known_tests = [
x
for x in tests
if get_duration(x, test_file_times, test_class_times) is not None
]
unknown_tests = [x for x in tests if x not in known_tests]
pytest_sharded_tests = sorted(
get_with_pytest_shard(known_tests, test_file_times, test_class_times),
key=lambda j: j.get_time(),
reverse=True,
) + get_with_pytest_shard(unknown_tests, test_file_times, test_class_times)
else:
pytest_sharded_tests = get_with_pytest_shard(
tests, test_file_times, test_class_times
)
del tests
serial_tests = [
test
for test in tests
if get_duration(test, test_file_times, test_class_times) is None
or must_serial(test.test_file)
for test in pytest_sharded_tests
if must_serial(test.name) or test.time is None
]
parallel_tests = [test for test in tests if test not in serial_tests]
parallel_tests = [test for test in pytest_sharded_tests if test not in serial_tests]
serial_time = sum(
get_duration(test, test_file_times, test_class_times) or 0
for test in serial_tests
)
parallel_time = sum(
get_duration(test, test_file_times, test_class_times) or 0
for test in parallel_tests
)
serial_time = sum(test.get_time(DEFAULT_TIME) for test in serial_tests)
parallel_time = sum(test.get_time(DEFAULT_TIME) for test in parallel_tests)
total_time = serial_time + parallel_time / NUM_PROCS_FOR_SHARDING_CALC
estimated_time_per_shard = total_time / num_shards
# Separate serial tests from parallel tests as much as possible to maximize
@ -237,20 +236,14 @@ def calculate_shards(
sharded_jobs = [ShardJob() for _ in range(num_shards)]
shard(
sharded_jobs[:num_serial_shards],
serial_tests,
test_file_times,
test_class_times,
sharded_jobs=sharded_jobs[:num_serial_shards],
pytest_sharded_tests=serial_tests,
estimated_time_limit=estimated_time_limit,
sort_by_time=sort_by_time,
serial=True,
)
shard(
sharded_jobs,
parallel_tests,
test_file_times,
test_class_times,
sort_by_time=sort_by_time,
sharded_jobs=sharded_jobs,
pytest_sharded_tests=parallel_tests,
serial=False,
)