Files
DeepSpeed/tests/onebit/test_mpi_perf.py
2023-03-30 17:14:38 -07:00

77 lines
2.2 KiB
Python

# Copyright (c) Microsoft Corporation.
# SPDX-License-Identifier: Apache-2.0
# DeepSpeed Team
from mpi4py import MPI
import torch
import deepspeed
from deepspeed.runtime.comm.mpi import MpiBackend
# Configure wall clock timer
from deepspeed.utils.timer import SynchronizedWallClockTimer
from deepspeed.accelerator import get_accelerator
from statistics import mean
timers = SynchronizedWallClockTimer()
comm = MPI.COMM_WORLD
size = comm.Get_size()
rank = comm.Get_rank()
deepspeed.init_distributed(dist_backend=get_accelerator().communication_backend_name())
# Change cuda_aware to True to test out CUDA-Aware MPI communication
backend = MpiBackend(cuda_aware=False)
local_rank = rank % get_accelerator().device_count()
device = torch.device(get_accelerator().device_name(), local_rank)
tensor_size = 300 * 2**20
server_size = int(tensor_size / size)
if tensor_size % (8 * size) != 0:
right_tensor_size = tensor_size + (8 * size - (tensor_size % (8 * size)))
else:
right_tensor_size = tensor_size
right_server_size = right_tensor_size // size
# Adding bias to the initialization of the gradient we are communicating
# In order to get rid of the case where some elements in the gradient are too small
a = (torch.rand(tensor_size, device=device) - 0.5) + 0.01 * rank
worker_error = torch.zeros(right_tensor_size, device=device)
server_error = torch.zeros(right_server_size, device=device)
warmup = 10
iters = 10
# Warmup
for i in range(warmup):
backend.compressed_allreduce(a, worker_error, server_error, local_rank)
time_list = []
for i in range(iters):
timers('compressed_allreduce').start()
backend.compressed_allreduce(a, worker_error, server_error, local_rank)
timers('compressed_allreduce').stop()
time_list.append(timers('compressed_allreduce').elapsed())
timer_names = ['compressed_allreduce']
timers.log(names=timer_names, normalizer=1, memory_breakdown=None)
places = 2
convert = 1e3
float_size = 4
if rank == 0:
for i in range(iters):
lat = time_list[i]
print("latency = ", lat * convert)
minlat = round(min(time_list) * convert)
maxlat = round(max(time_list) * convert)
meanlat = round(mean(time_list) * convert, places)
print("min, max, and mean = {} ms, {} ms, {} ms".format(minlat, maxlat, meanlat))