[Performance] in windows v1.19.2 - non-deterministic but in linux deterministic

[parkkyungjun]

Describe the issue

Hello, I have a question about onnxruntime.

On Windows, the inference results differ in version 1.19.2, but they are consistent in version 1.20.0. (in cuda)

On Linux, the results are consistent in both versions 1.19.2 and 1.20.0.

Could you help me understand why this might be?

To reproduce

just inference any file in cuda

Urgency

No response

Platform

Windows

OS Version

windows10

ONNX Runtime Installation

Released Package

ONNX Runtime Version or Commit ID

1.19.2

ONNX Runtime API

Python

Architecture

X64

Execution Provider

CUDA

Execution Provider Library Version

No response

Model File

No response

Is this a quantized model?

No