Data preprocess conflict in multi XPU training

zhouyu5 commented 6 days ago

🐛 Describe the bug

While performing multi xpu training, it will report the following error message:

Reproduce code:

cd benchmark/multi_gpu/training
mpirun -np 2 python training_benchmark_xpu.py --dataset ogbn-products --model edge_cnn --num-epochs 1

Root cause analysis The dataset will download and transform repeatedly on each process, thus causing the conflic. Please check the code. In fact, the dataset download and transform process only need to be done on main process, that is the rank=0 process.
```
assert args.dataset in supported_sets.keys(), \
    f"Dataset {args.dataset} isn't supported."
data, num_classes = get_dataset(args.dataset, args.root)
```
suggest fix First download/preprocess data on main process, and after that, read the processed data on other process. I would like to submit a PR to solve this problem.

Versions

Collecting environment information... PyTorch version: 2.1.0.post2+cxx11.abi Is debug build: False CUDA used to build PyTorch: None ROCM used to build PyTorch: N/A

OS: Ubuntu 22.04.4 LTS (x86_64) GCC version: (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 Clang version: Could not collect CMake version: version 3.22.1 Libc version: glibc-2.35

Python version: 3.10.12 (main, Nov 20 2023, 15:14:05) [GCC 11.4.0] (64-bit runtime) Python platform: Linux-5.15.0-73-generic-x86_64-with-glibc2.35 Is CUDA available: False CUDA runtime version: No CUDA CUDA_MODULE_LOADING set to: N/A GPU models and configuration: No CUDA Nvidia driver version: No CUDA cuDNN version: No CUDA 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: 52 bits physical, 57 bits virtual Byte Order: Little Endian CPU(s): 224 On-line CPU(s) list: 0-223 Vendor ID: GenuineIntel Model name: Intel(R) Xeon(R) Platinum 8480+ CPU family: 6 Model: 143 Thread(s) per core: 2 Core(s) per socket: 56 Socket(s): 2 Stepping: 8 CPU max MHz: 3800.0000 CPU min MHz: 800.0000 BogoMIPS: 4000.00 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 monitor 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 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: 5.3 MiB (112 instances) L1i cache: 3.5 MiB (112 instances) L2 cache: 224 MiB (112 instances) L3 cache: 210 MiB (2 instances) NUMA node(s): 2 NUMA node0 CPU(s): 0-55,112-167 NUMA node1 CPU(s): 56-111,168-223 Vulnerability Itlb multihit: Not affected Vulnerability L1tf: Not affected Vulnerability Mds: Not affected Vulnerability Meltdown: Not affected Vulnerability Mmio stale data: Not affected Vulnerability Retbleed: 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 IBRS, IBPB conditional, RSB filling, PBRSB-eIBRS SW sequence Vulnerability Srbds: Not affected Vulnerability Tsx async abort: Not affected

Versions of relevant libraries: [pip3] intel-extension-for-pytorch==2.1.30+xpu [pip3] numpy==1.26.4 [pip3] torch==2.1.0.post2+cxx11.abi [pip3] torch-geometric==2.6.0 [pip3] torchaudio==2.1.0.post2+cxx11.abi [pip3] torchvision==0.16.0.post2+cxx11.abi [conda] Could not collect

zhouyu5 commented 6 days ago

@rusty1s @DamianSzwichtenberg Could you please share some of your thoughts? Thanks.

DamianSzwichtenberg commented 5 days ago

@zhouyu5 your proposition for a fix seems reasonable, please go ahead. 😉

zhouyu5 commented 5 days ago

@zhouyu5 your proposition for a fix seems reasonable, please go ahead. 😉

Fixed it in PR #9456 , please have a review. @DamianSzwichtenberg

pyg-team / pytorch_geometric

Data preprocess conflict in multi XPU training #9455

🐛 Describe the bug

Versions