TensorRT-LLMs/cpp/tensorrt_llm/kernels/fmhaDispatcher.h
Yihan Wang 9df4dad3b6
[None][fix] Introduce inline namespace to avoid symbol collision (#9541)
Signed-off-by: Yihan Wang <yihwang@nvidia.com>
2025-12-12 23:32:15 +08:00

71 lines
2.2 KiB
C++

/*
* Copyright (c) 2020-2024, NVIDIA CORPORATION. All rights reserved.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#pragma once
#include "tensorrt_llm/common/config.h"
#include "tensorrt_llm/common/opUtils.h"
#include "tensorrt_llm/kernels/contextFusedMultiHeadAttention/fmhaRunner.h"
#include "tensorrt_llm/kernels/contextFusedMultiHeadAttention/fused_multihead_attention_common.h"
#include "tensorrt_llm/kernels/trtllmGenKernels/fmha/fmhaRunner.h"
using tensorrt_llm::common::op::UniqPtrWNullCopy;
TRTLLM_NAMESPACE_BEGIN
namespace kernels
{
////////////////////////////////////////////////////////////////////////////////////////////////////
class FmhaDispatcher
{
public:
// Constructor.
FmhaDispatcher(kernels::MHARunnerFixedParams fixedParams);
// Deconstructor.
~FmhaDispatcher() = default;
// Check if any fmha kernel meets the requirements.
bool isSupported();
// Does FMHA need a separate Q and Kv input ?
bool isSeparateQAndKvInput() const
{
return mFixedParams.attentionInputLayout != kernels::AttentionInputLayout::PACKED_QKV;
}
// Run the fmha kernel.
void run(tensorrt_llm::kernels::MHARunnerParams runnerParams);
private:
// The fixed fmha parameters.
kernels::MHARunnerFixedParams mFixedParams;
// Whether to enable trtllm-gen kernels.
bool mUseTllmGen;
// Runner for fmha v2 kernels (for SM <= 90)
UniqPtrWNullCopy<kernels::FusedMHARunnerV2> mFMHARunner;
// Runner for trtllm-gen fmha kernels (for SM == 100)
UniqPtrWNullCopy<kernels::TllmGenFmhaRunner> mTllmGenFMHARunner;
};
////////////////////////////////////////////////////////////////////////////////////////////////////
} // namespace kernels
TRTLLM_NAMESPACE_END