pytorch - 💡(How to fix) Fix torch.compile overflows torch.sinh near float32 boundary while eager returns finite values

🐛 Describe the bug

torch.compile with the Inductor backend produces different overflow behavior from eager execution for torch.sinh on large float32 inputs near the overflow boundary.

Eager execution returns finite float32 values around 1e38, while the compiled version returns inf / -inf for the same inputs.

import warnings
warnings.filterwarnings("ignore")

import numpy as np
import torch

def fn(x):
    return torch.sinh(x)

x = torch.tensor([88.85, 89.2, 88.95, -89.0], dtype=torch.float32)

eager = fn(x).cpu().numpy()

torch._dynamo.reset()
compiled = torch.compile(fn, backend="inductor", fullgraph=True)(x).cpu().numpy()

print(f"input: {x.cpu().numpy()}")
print(f"eager: {eager}")
print(f"comp : {compiled}")

assert np.isfinite(eager[0]) and eager[0] > 1e38
assert np.isinf(comp[0])

Error logs

input: [ 88.85 89.2 88.95 -89. ] eager: [ 1.9321198e+38 2.7418043e+38 2.1353193e+38 -2.2448064e+38] comp : [ inf inf inf -inf]

Versions

PyTorch version: 2.13.0.dev20260513+cpu Is debug build: False CUDA used to build PyTorch: Could not collect ROCM used to build PyTorch: N/A

OS: Ubuntu 24.04.3 LTS (x86_64) GCC version: (Ubuntu 13.3.0-6ubuntu2~24.04.1) 13.3.0 Clang version: 18.1.3 (1ubuntu1) CMake version: Could not collect Libc version: glibc-2.39

Python version: 3.11.15 (main, Mar 11 2026, 17:20:07) [GCC 14.3.0] (64-bit runtime) Python platform: Linux-6.17.0-20-generic-x86_64-with-glibc2.39 Is CUDA available: False CUDA runtime version: Could not collect CUDA_MODULE_LOADING set to: N/A GPU models and configuration: GPU 0: NVIDIA RTX 6000 Ada Generation GPU 1: NVIDIA RTX 6000 Ada Generation

Nvidia driver version: 570.211.01 cuDNN version: Could not collect Is XPU available: False HIP runtime version: N/A MIOpen runtime version: N/A Is XNNPACK available: True Caching allocator config: N/A

Versions of relevant libraries: [pip3] numpy==2.4.4 [pip3] torch==2.13.0.dev20260513+cpu [conda] numpy 2.4.4 pypi_0 pypi [conda] torch 2.13.0.dev20260513+cpu pypi_0 pypi

cc @chauhang @penguinwu @voznesenskym @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @kadeng @muchulee8 @amjames @aakhundov @coconutruben @jataylo