Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Program does not exit when using torch.distributed.tensor.distribute_module. #139060

Closed
dest1n1s opened this issue Oct 28, 2024 · 2 comments
Closed

Program does not exit when using torch.distributed.tensor.distribute_module. #139060

dest1n1s opened this issue Oct 28, 2024 · 2 comments
Labels
module: dtensor distributed tensor tag oncall: distributed Add this issue/PR to distributed oncall triage queue triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module

Comments

@dest1n1s
Copy link

dest1n1s commented Oct 28, 2024 •
edited by pytorch-bot bot
Loading

🐛 Describe the bug

The following program (mainly based on the example provided in PyTorch DTensor) does not exit automatically after outputting the shape of y:

import torch
import torch.nn as nn
from torch.distributed.tensor import (
    Replicate,
    Shard,
    distribute_module,
    distribute_tensor,
    init_device_mesh,
)


class MyModule(nn.Module):
    def __init__(self) -> None:
        super().__init__()
        self.fc1 = nn.Linear(8, 8)
        self.fc2 = nn.Linear(8, 8)
        self.relu = nn.ReLU()

    def forward(self, input):
        return self.relu(self.fc1(input) + self.fc2(input))


mesh = init_device_mesh("cuda", (4,))


def shard_params(mod_name, mod, mesh):
    col_linear_placement = [Shard(0)]
    # shard fc1 and fc2
    if isinstance(mod, nn.Linear):
        for name, param in mod.named_parameters():
            dist_param = nn.Parameter(
                distribute_tensor(param, mesh, col_linear_placement)
            )
            mod.register_parameter(name, dist_param)


sharded_module = distribute_module(MyModule(), mesh, partition_fn=shard_params)

x = torch.randn(8, 8).cuda(0)
x = distribute_tensor(x, mesh, [Replicate()])
y = sharded_module(x)
print(y.shape)

The command used to run this program is:

torchrun --standalone --nnodes=1 --nproc-per-node=4 test.py

The full output is:

W1028 10:48:46.760000 292341 site-packages/torch/distributed/run.py:793]                                                
W1028 10:48:46.760000 292341 site-packages/torch/distributed/run.py:793] *****************************************      
W1028 10:48:46.760000 292341 site-packages/torch/distributed/run.py:793] Setting OMP_NUM_THREADS environment variable fo
r each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal p
erformance in your application as needed.                                                                               
W1028 10:48:46.760000 292341 site-packages/torch/distributed/run.py:793] *****************************************      
torch.Size([8, 8])

It seems that the program should exit after the final output.

Versions

Collecting environment information...
PyTorch version: 2.5.0+cu124
Is debug build: False
CUDA used to build PyTorch: 12.4
ROCM used to build PyTorch: N/A

OS: Ubuntu 22.04.5 LTS (x86_64)
GCC version: (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0
Clang version: Could not collect
CMake version: Could not collect
Libc version: glibc-2.35

Python version: 3.12.5 | packaged by conda-forge | (main, Aug  8 2024, 18:36:51) [GCC 12.4.0] (64-bit runtime)
Python platform: Linux-5.15.0-124-generic-x86_64-with-glibc2.35
Is CUDA available: True
CUDA runtime version: Could not collect
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration: 
GPU 0: NVIDIA GeForce RTX 4090 D
GPU 1: NVIDIA GeForce RTX 4090 D
GPU 2: NVIDIA GeForce RTX 4090 D
GPU 3: NVIDIA GeForce RTX 4090 D

Nvidia driver version: 560.35.03
cuDNN version: Could not collect
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True

CPU:
Architecture:                         x86_64
CPU op-mode(s):                       32-bit, 64-bit
Address sizes:                        46 bits physical, 57 bits virtual
Byte Order:                           Little Endian
CPU(s):                               128
On-line CPU(s) list:                  0-127
Vendor ID:                            GenuineIntel
Model name:                           Intel(R) Xeon(R) Gold 6430
CPU family:                           6
Model:                                143
Thread(s) per core:                   2
Core(s) per socket:                   32
Socket(s):                            2
Stepping:                             8
CPU max MHz:                          3400.0000
CPU min MHz:                          800.0000
BogoMIPS:                             4200.00
Flags:                                fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf tsc_known_freq pni pclmulqdq dtes64 ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cat_l2 cdp_l3 invpcid_single intel_ppin cdp_l2 ssbd mba ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb intel_pt avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local split_lock_detect avx_vnni avx512_bf16 wbnoinvd dtherm ida arat pln pts hwp hwp_act_window hwp_epp hwp_pkg_req avx512vbmi umip pku ospke waitpkg avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg tme avx512_vpopcntdq la57 rdpid bus_lock_detect cldemote movdiri movdir64b enqcmd fsrm md_clear serialize tsxldtrk pconfig arch_lbr amx_bf16 avx512_fp16 amx_tile amx_int8 flush_l1d arch_capabilities
Virtualization:                       VT-x
L1d cache:                            3 MiB (64 instances)
L1i cache:                            2 MiB (64 instances)
L2 cache:                             128 MiB (64 instances)
L3 cache:                             120 MiB (2 instances)
NUMA node(s):                         2
NUMA node0 CPU(s):                    0-31,64-95
NUMA node1 CPU(s):                    32-63,96-127
Vulnerability Gather data sampling:   Not affected
Vulnerability Itlb multihit:          Not affected
Vulnerability L1tf:                   Not affected
Vulnerability Mds:                    Not affected
Vulnerability Meltdown:               Not affected
Vulnerability Mmio stale data:        Not affected
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed:               Not affected
Vulnerability Spec rstack overflow:   Not affected
Vulnerability Spec store bypass:      Mitigation; Speculative Store Bypass disabled via prctl and seccomp
Vulnerability Spectre v1:             Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:             Mitigation; Enhanced / Automatic IBRS; IBPB conditional; RSB filling; PBRSB-eIBRS SW sequence; BHI BHI_DIS_S
Vulnerability Srbds:                  Not affected
Vulnerability Tsx async abort:        Not affected

Versions of relevant libraries:
[pip3] numpy==1.26.4
[pip3] nvidia-cublas-cu12==12.4.5.8
[pip3] nvidia-cuda-cupti-cu12==12.4.127
[pip3] nvidia-cuda-nvrtc-cu12==12.4.127
[pip3] nvidia-cuda-runtime-cu12==12.4.127
[pip3] nvidia-cudnn-cu12==9.1.0.70
[pip3] nvidia-cufft-cu12==11.2.1.3
[pip3] nvidia-curand-cu12==10.3.5.147
[pip3] nvidia-cusolver-cu12==11.6.1.9
[pip3] nvidia-cusparse-cu12==12.3.1.170
[pip3] nvidia-nccl-cu12==2.21.5
[pip3] nvidia-nvjitlink-cu12==12.4.127
[pip3] nvidia-nvtx-cu12==12.4.127
[pip3] optree==0.12.1
[pip3] torch==2.5.0
[pip3] triton==3.1.0
[conda] numpy                     1.26.4                   pypi_0    pypi
[conda] nvidia-cublas-cu12        12.4.5.8                 pypi_0    pypi
[conda] nvidia-cuda-cupti-cu12    12.4.127                 pypi_0    pypi
[conda] nvidia-cuda-nvrtc-cu12    12.4.127                 pypi_0    pypi
[conda] nvidia-cuda-runtime-cu12  12.4.127                 pypi_0    pypi
[conda] nvidia-cudnn-cu12         9.1.0.70                 pypi_0    pypi
[conda] nvidia-cufft-cu12         11.2.1.3                 pypi_0    pypi
[conda] nvidia-curand-cu12        10.3.5.147               pypi_0    pypi
[conda] nvidia-cusolver-cu12      11.6.1.9                 pypi_0    pypi
[conda] nvidia-cusparse-cu12      12.3.1.170               pypi_0    pypi
[conda] nvidia-nccl-cu12          2.21.5                   pypi_0    pypi
[conda] nvidia-nvjitlink-cu12     12.4.127                 pypi_0    pypi
[conda] nvidia-nvtx-cu12          12.4.127                 pypi_0    pypi
[conda] optree                    0.12.1                   pypi_0    pypi
[conda] torch                     2.5.0                    pypi_0    pypi
[conda] triton                    3.1.0                    pypi_0    pypi

cc @H-Huang @awgu @kwen2501 @wanchaol @fegin @fduwjj @wz337 @wconstab @d4l3k @c-p-i-o @tianyu-l @XilunWu
@awgu awgu added the module: dtensor distributed tensor tag label Oct 28, 2024
@malfet malfet added the oncall: distributed Add this issue/PR to distributed oncall triage queue label Oct 28, 2024
@XilunWu
Copy link
Contributor

XilunWu commented Oct 28, 2024

I can confirm the hang on the test script. However, changing x = torch.randn(8, 8).cuda(0) to x = torch.randn(8, 8).cuda() will resolve the hang so it's very likely caused by multiple processes trying to access one device in user code rather than DTensor's behavior change. I think it's recommended that each process (i.e. rank) only access its own device, right? cc @kwen2501

@XilunWu XilunWu added the triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module label Oct 28, 2024
@awgu
Copy link
Contributor

awgu commented Oct 30, 2024

yes I think the .cuda(0) does not look right
let us close this for now unless @dest1n1s has specific reason why the input should be on cuda:0 for all ranks

@awgu awgu closed this as completed Oct 30, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
module: dtensor distributed tensor tag oncall: distributed Add this issue/PR to distributed oncall triage queue triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
4 participants
@malfet @XilunWu @dest1n1s @awgu