Dtype Selective Build Performs *Increases* Binary Size for Transformer Models

[BujSet]

🐛 Describe the bug

Summary

When evaluating the efficacy of the dtype selective build process, I found that for some of the transformer models, using the dtype selective build process resulted in an binary that was larger than if dtype weren't considered (only operator selective build). This seems like an error because in theory, the binary size should only ever stay the same or decrease when using the dtype selective in comparison to operator selective build.

Steps to Repro

Checkout changes here: https://github.com/BujSet/executorch/tree/dtype_selection_size_errors_test

After installing executorch on that patch, you can run:

bash examples/dtype_selective_build/test_selective_build.sh cmake

This creates a file called results.txt that lists the binary size results w/ and w/o dtype selective build. You should see something like:

Model,UseNoOps,IncludeAllOps,ModelDtypeSelect,StrippedBinarySize,CompilationTime(sec),NumOps,OpsWith1Dtype,OpsWith2Dtypes,OpsWith3+Dtypes
phi_4_mini,False,True,OFF,3803560,9.312259892
qwen2_5,False,True,OFF,3803560,9.672060556
phi_4_mini,False,False,ON,698792,44.351337928,
phi_4_mini,False,False,OFF,633256,10.269945404
qwen2_5,False,False,ON,698792,50.225504398,
qwen2_5,False,False,OFF,633256,8.811934295

When IncludeAllOps is set to True, the binary is built to include all operators and dtypes (which is why phi_4_mini and qwen2_5 report the same binary size of 3803560. When ModelDtypeSelect is set to OFF, the build process defaults to operator selective build. So we should expect that when ModelDtypeSelect is ON, the binary size should be smaller. But instead, we see:

phi_4_mini: 698792 > 633256
qwen2_5: 698792 > 633256

Versions

Collecting environment information...
PyTorch version: 2.9.0.dev20250714+cpu
Is debug build: False
CUDA used to build PyTorch: None
ROCM used to build PyTorch: N/A

OS: Ubuntu 24.04.2 LTS (x86_64)
GCC version: (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0
Clang version: Could not collect
CMake version: version 3.31.6
Libc version: glibc-2.39

Python version: 3.10.0 (default, Mar  3 2022, 09:58:08) [GCC 7.5.0] (64-bit runtime)
Python platform: Linux-6.6.87.2-microsoft-standard-WSL2-x86_64-with-glibc2.39
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:                        46 bits physical, 48 bits virtual
Byte Order:                           Little Endian
CPU(s):                               10
On-line CPU(s) list:                  0-9
Vendor ID:                            GenuineIntel
Model name:                           Intel(R) Core(TM) Ultra 7 165U
CPU family:                           6
Model:                                170
Thread(s) per core:                   2
Core(s) per socket:                   5
Socket(s):                            1
Stepping:                             4
BogoMIPS:                             5376.00
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 tsc_reliable nonstop_tsc cpuid tsc_known_freq pni pclmulqdq vmx ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch ssbd ibrs ibpb stibp ibrs_enhanced tpr_shadow ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves avx_vnni vnmi umip waitpkg gfni vaes vpclmulqdq rdpid movdiri movdir64b fsrm md_clear serialize flush_l1d arch_capabilities
Virtualization:                       VT-x
Hypervisor vendor:                    Microsoft
Virtualization type:                  full
L1d cache:                            240 KiB (5 instances)
L1i cache:                            320 KiB (5 instances)
L2 cache:                             10 MiB (5 instances)
L3 cache:                             12 MiB (1 instance)
NUMA node(s):                         1
NUMA node0 CPU(s):                    0-9
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:               Mitigation; Enhanced IBRS
Vulnerability Spec rstack overflow:   Not affected
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 / 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] executorch==0.7.0a0+8f97a34
[pip3] flake8==6.1.0
[pip3] flake8-breakpoint==1.1.0
[pip3] flake8-bugbear==24.4.26
[pip3] flake8-comprehensions==3.14.0
[pip3] flake8-plugin-utils==1.3.3
[pip3] flake8-pyi==23.5.0
[pip3] mypy==1.14.1
[pip3] mypy_extensions==1.1.0
[pip3] numpy==2.2.6
[pip3] nvidia-cublas-cu12==12.6.4.1
[pip3] nvidia-cuda-cupti-cu12==12.6.80
[pip3] nvidia-cuda-nvrtc-cu12==12.6.77
[pip3] nvidia-cuda-runtime-cu12==12.6.77
[pip3] nvidia-cudnn-cu12==9.5.1.17
[pip3] nvidia-cufft-cu12==11.3.0.4
[pip3] nvidia-curand-cu12==10.3.7.77
[pip3] nvidia-cusolver-cu12==11.7.1.2
[pip3] nvidia-cusparse-cu12==12.5.4.2
[pip3] nvidia-cusparselt-cu12==0.6.3
[pip3] nvidia-nccl-cu12==2.26.2
[pip3] nvidia-nvjitlink-cu12==12.6.85
[pip3] nvidia-nvtx-cu12==12.6.77
[pip3] pytorch_tokenizers==0.1.0
[pip3] torch==2.9.0.dev20250714+cpu
[pip3] torchao==0.12.0+gitaee079503
[pip3] torchaudio==2.8.0.dev20250714+cpu
[pip3] torchdata==0.11.0
[pip3] torchsr==1.0.4
[pip3] torchtune==0.6.1
[pip3] torchvision==0.24.0.dev20250714+cpu
[pip3] triton==3.3.1
[conda] executorch                0.8.0a0+c2de265          pypi_0    pypi
[conda] numpy                     2.2.6                    pypi_0    pypi
[conda] nvidia-cublas-cu12        12.6.4.1                 pypi_0    pypi
[conda] nvidia-cuda-cupti-cu12    12.6.80                  pypi_0    pypi
[conda] nvidia-cuda-nvrtc-cu12    12.6.77                  pypi_0    pypi
[conda] nvidia-cuda-runtime-cu12  12.6.77                  pypi_0    pypi
[conda] nvidia-cudnn-cu12         9.5.1.17                 pypi_0    pypi
[conda] nvidia-cufft-cu12         11.3.0.4                 pypi_0    pypi
[conda] nvidia-curand-cu12        10.3.7.77                pypi_0    pypi
[conda] nvidia-cusolver-cu12      11.7.1.2                 pypi_0    pypi
[conda] nvidia-cusparse-cu12      12.5.4.2                 pypi_0    pypi
[conda] nvidia-cusparselt-cu12    0.6.3                    pypi_0    pypi
[conda] nvidia-nccl-cu12          2.26.2                   pypi_0    pypi
[conda] nvidia-nvjitlink-cu12     12.6.85                  pypi_0    pypi
[conda] nvidia-nvtx-cu12          12.6.77                  pypi_0    pypi
[conda] pytorch-tokenizers        0.1.0                    pypi_0    pypi
[conda] torch                     2.9.0.dev20250714+cpu          pypi_0    pypi
[conda] torchao                   0.12.0+gitaee079503          pypi_0    pypi
[conda] torchaudio                2.8.0.dev20250714+cpu          pypi_0    pypi
[conda] torchdata                 0.11.0                   pypi_0    pypi
[conda] torchfix                  0.6.0                    pypi_0    pypi
[conda] torchsr                   1.0.4                    pypi_0    pypi
[conda] torchtune                 0.6.1                    pypi_0    pypi
[conda] torchvision               0.24.0.dev20250714+cpu          pypi_0    pypi
[conda] triton                    3.3.1                    pypi_0    pypi