From 8aa2667ae61ffecbee883009d5fa7f18bef595db Mon Sep 17 00:00:00 2001
From: Frank Dong <123416088+frank-dong-ms@users.noreply.github.com>
Date: Mon, 17 Jun 2024 05:52:13 -0700
Subject: [PATCH] add bf16 for Tile CUDA executor (#20854)

### Description
add bf16 for Tile CUDA executor



### Motivation and Context
required change to support phimm model for ORT training
---
 docs/OperatorKernels.md                        | 2 +-
 onnxruntime/core/providers/cuda/tensor/tile.cc | 3 ++-
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/docs/OperatorKernels.md b/docs/OperatorKernels.md
index 67bfe48327e14..5f19c16cba616 100644
--- a/docs/OperatorKernels.md
+++ b/docs/OperatorKernels.md
@@ -827,7 +827,7 @@ Do not modify directly.*
 |||[6, 12]|**T** = tensor(double), tensor(float), tensor(float16)|
 |ThresholdedRelu|*in* X:**T**<br> *out* Y:**T**|10+|**T** = tensor(double), tensor(float), tensor(float16)|
 |||1+|**T** = tensor(double), tensor(float), tensor(float16)|
-|Tile|*in* input:**T**<br> *in* repeats:**T1**<br> *out* output:**T**<br><br>or<br><br>*in* input:**T**<br> *in* tiles:**T**<br> *in* axis:**T**<br> *out* output:**T**|13+|**T** = tensor(double), tensor(float), tensor(float16), tensor(int32), tensor(int64)<br/> **T1** = tensor(int64)|
+|Tile|*in* input:**T**<br> *in* repeats:**T1**<br> *out* output:**T**<br><br>or<br><br>*in* input:**T**<br> *in* tiles:**T**<br> *in* axis:**T**<br> *out* output:**T**|13+|**T** = tensor(bfloat16), tensor(double), tensor(float), tensor(float16), tensor(int32), tensor(int64)<br/> **T1** = tensor(int64)|
 |||[6, 12]|**T** = tensor(double), tensor(float), tensor(float16), tensor(int32), tensor(int64)<br/> **T1** = tensor(int64)|
 |TopK|*in* X:**T**<br> *in* K:**tensor(int64)**<br> *out* Values:**T**<br> *out* Indices:**I**<br><br>or<br><br>*in* X:**T**<br> *out* Values:**T**<br> *out* Indices:**I**|11+|**I** = tensor(int64)<br/> **T** = tensor(double), tensor(float), tensor(float16), tensor(int32), tensor(int64)|
 |||10|**I** = tensor(int64)<br/> **T** = tensor(double), tensor(float), tensor(float16), tensor(int32), tensor(int64)|
diff --git a/onnxruntime/core/providers/cuda/tensor/tile.cc b/onnxruntime/core/providers/cuda/tensor/tile.cc
index e671b2cdc0277..01522c71dc51b 100644
--- a/onnxruntime/core/providers/cuda/tensor/tile.cc
+++ b/onnxruntime/core/providers/cuda/tensor/tile.cc
@@ -36,7 +36,8 @@ ONNX_OPERATOR_KERNEL_EX(
                               DataTypeImpl::GetTensorType<double>(),
                               DataTypeImpl::GetTensorType<int32_t>(),
                               DataTypeImpl::GetTensorType<int64_t>(),
-                              DataTypeImpl::GetTensorType<MLFloat16>()})
+                              DataTypeImpl::GetTensorType<MLFloat16>(),
+                              DataTypeImpl::GetTensorType<BFloat16>()})
         .TypeConstraint("T1", DataTypeImpl::GetTensorType<int64_t>()),
     Tile);