MPRS with DL2 memory use increases over time vs. using older DL

🐛 Describe the bug

MPRS version:

    train_dp = IterableWrapper(train_ds).batch(batch_size=batch_size).collate(collate_fn=encode_processor)

    rs = MultiProcessingReadingService(num_workers=num_workers)
    train_dl = DataLoader2(train_dp, reading_service=rs)
    for batch_idx, batch in enumerate(train_dl):
        print(batch_idx)
        if batch_idx > 500:
            break

Running with mprof run --multiprocess --include-children <script.py> results in increasing memory usage over time:

vs. older DataLoader:

    train_dl = DataLoader(train_ds, batch_size=batch_size, num_workers=num_workers, collate_fn=encode_processor)
    for batch_idx, batch in enumerate(train_dl):
        print(batch_idx)
        if batch_idx > 500:
            break

Running with mprof run --multiprocess --include-children <script.py> results in reasonable memory usage over time:

Any possible recommendations for debugging this further. For context train_ds is a MapDataPipe generating some strings for each index.

Versions

PyTorch version: 2.0.1+cu117
Is debug build: False
CUDA used to build PyTorch: 11.7
ROCM used to build PyTorch: N/A

OS: Ubuntu 22.04.2 LTS (x86_64)
GCC version: (Ubuntu 11.3.0-1ubuntu1~22.04) 11.3.0
Clang version: Could not collect
CMake version: version 3.26.3
Libc version: glibc-2.35

Python version: 3.10.11 (main, Apr 20 2023, 19:02:41) [GCC 11.2.0] (64-bit runtime)
Python platform: Linux-5.19.0-1022-gcp-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 A100-SXM4-40GB
Nvidia driver version: 530.30.02
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, 48 bits virtual
Byte Order: Little Endian
CPU(s): 12
On-line CPU(s) list: 0-11
Vendor ID: GenuineIntel
Model name: Intel(R) Xeon(R) CPU @ 2.20GHz
CPU family: 6
Model: 85
Thread(s) per core: 2
Core(s) per socket: 6
Socket(s): 1
Stepping: 7
BogoMIPS: 4400.42
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single ssbd ibrs ibpb stibp ibrs_enhanced fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm mpx avx512f avx512dq rdseed adx smap clflushopt clwb avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves arat avx512_vnni md_clear arch_capabilities
Hypervisor vendor: KVM
Virtualization type: full
L1d cache: 192 KiB (6 instances)
L1i cache: 192 KiB (6 instances)
L2 cache: 6 MiB (6 instances)
L3 cache: 38.5 MiB (1 instance)
NUMA node(s): 1
NUMA node0 CPU(s): 0-11
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Mitigation; Clear CPU buffers; SMT Host state unknown
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Vulnerable: Clear CPU buffers attempted, no microcode; SMT Host state unknown
Vulnerability Retbleed: Mitigation; Enhanced IBRS
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl
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: Mitigation; Clear CPU buffers; SMT Host state unknown

Versions of relevant libraries:
[pip3] mypy-extensions==1.0.0
[pip3] numpy==1.24.3
[pip3] pytorch-lightning==2.0.2
[pip3] torch==2.0.1
[pip3] torchdata==0.6.1
[pip3] torchmetrics==0.11.4
[pip3] torchvision==0.15.2
[pip3] triton==2.0.0
[conda] numpy 1.24.3 pypi_0 pypi
[conda] pytorch-lightning 2.0.2 pypi_0 pypi
[conda] torch 2.0.1 pypi_0 pypi
[conda] torchdata 0.6.1 pypi_0 pypi
[conda] torchmetrics 0.11.4 pypi_0 pypi
[conda] torchvision 0.15.2 pypi_0 pypi
[conda] triton 2.0.0 pypi_0 pypi

No project files or tests are named. Start by reproducing the supplied Python MPRS/DataLoader2 loop with mprof, then compare its memory profile with the older DataLoader example. Done means identifying the source of the increasing memory use and documenting a reproducible finding or fix.