kanshan/TensorRT-LLMs
1
0
Fork 0
mirror of https://github.com/NVIDIA/TensorRT-LLM.git synced 2026-01-14 06:27:45 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity
be9cd719f7
TensorRT-LLMs/cpp/tensorrt_llm/plugins/gptAttentionCommon/gptAttentionCommon.h
Kaiyu Xie be9cd719f7
Update TensorRT-LLM (#2094) 
* Update TensorRT-LLM

---------

Co-authored-by: akhoroshev <arthoroshev@gmail.com>
Co-authored-by: Fabian Joswig <fjosw@users.noreply.github.com>
Co-authored-by: Tayef Shah <tayefshah@gmail.com>
Co-authored-by: lfz941 <linfanzai941@gmail.com>
2024-08-07 16:44:43 +08:00

449 lines
20 KiB
C++
Raw Blame History

/*
* SPDX-FileCopyrightText: Copyright (c) 1993-2022 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
* SPDX-License-Identifier: Apache-2.0
*
* 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/cublasMMWrapper.h"
#include "tensorrt_llm/common/quantization.h"
#include "tensorrt_llm/kernels/contextFusedMultiHeadAttention/fmhaRunner.h"
#include "tensorrt_llm/kernels/contextFusedMultiHeadAttention/fused_multihead_attention_common.h"
#include "tensorrt_llm/kernels/decoderMaskedMultiheadAttention/decoderXQARunner.h"
#include "tensorrt_llm/kernels/gptKernels.h"
#include "tensorrt_llm/kernels/kvCacheUtils.h"
#include "tensorrt_llm/plugins/common/plugin.h"
#include <cassert>
#include <set>
#include <string>
#include <vector>
namespace tensorrt_llm::plugins
{
class GPTAttentionPluginCommon : public BasePlugin
{
public:
GPTAttentionPluginCommon() = delete;
GPTAttentionPluginCommon(int layer_idx, int num_heads, int vision_start, int vision_length, int num_kv_heads,
int head_size, int unidirectional, float q_scaling, float qk_tanh_scale,
tensorrt_llm::kernels::PositionEmbeddingType position_embedding_type,
int rotary_embedding_dim, // for RoPE. Use 0 for non-RoPE
float rotary_embedding_base, tensorrt_llm::kernels::RotaryScalingType rotary_embedding_scale_type,
float rotary_embedding_scale, float rotary_embedding_short_m_scale, float rotary_embedding_long_m_scale,
int rotary_embedding_max_positions, int rotary_embedding_original_max_positions, int tp_size,
int tp_rank, // for ALiBi
bool unfuse_qkv_gemm, // for AutoPP
tensorrt_llm::kernels::ContextFMHAType context_fmha_type, bool enable_xqa, int kv_cache_quant_mode,
bool remove_input_padding, tensorrt_llm::kernels::AttentionMaskType mask_type,
tensorrt_llm::kernels::BlockSparseParams block_sparse_params, bool paged_kv_cache, int tokens_per_block,
nvinfer1::DataType type, int32_t max_context_length, bool qkv_bias_enabled, bool cross_attention = false,
int max_distance = 0, bool pos_shift_enabled = false, bool dense_context_fmha = false,
bool use_paged_context_fmha = false, bool use_fp8_context_fmha = false, bool use_cache = true,
bool is_spec_decoding_enabled = false, bool spec_decoding_is_generation_length_variable = false,
int32_t spec_decoding_max_generation_length = 1);
GPTAttentionPluginCommon(void const* data, size_t length);
~GPTAttentionPluginCommon() override = default;
template <typename T>
int enqueueImpl(nvinfer1::PluginTensorDesc const* inputDesc, nvinfer1::PluginTensorDesc const* outputDesc,
void const* const* inputs, void* const* outputs, void* workspace, cudaStream_t stream);
//! This is called on every trt Engine creation
int initialize() noexcept override;
//! This is called on every trt Engine destroy
void terminate() noexcept override;
//! This is called on every trt ExecutionContext creation by TRT
//! Note TRT does not call the initialize on cloned plugin, so clone internally should do initialization.
template <typename T>
T* cloneImpl() const noexcept;
//! This is called on evert trt Engine or ExecutionContext destroy.
//! None-cloned plugins will call terminate and then call destroy, while the cloned plugins will call destroy only
//! So plugin should put the resource release inside destroy.
void destroy() noexcept override;
size_t getCommonSerializationSize() const noexcept;
void serializeCommon(void* buffer) const noexcept;
int getHeadSize(bool checkInit = true) const;
protected:
int getMaxNumSeqLenTile(int batch_beam_size = 1) const;
size_t getWorkspaceSizeForContext(nvinfer1::DataType type, int32_t nbReq, int32_t max_input_length,
int32_t cross_qkv_length = 0, int32_t max_num_tokens = 0) const noexcept;
// total_num_seq is the sum of beam_width for multiple requests
size_t getWorkspaceSizeForGeneration(nvinfer1::DataType type, int32_t total_num_seq, int32_t max_kv_cache_length,
int32_t max_num_tokens) const noexcept;
template <typename T, typename KVCacheBuffer>
struct EnqueueContextParams
{
T const* attention_input;
T const* qkv_bias;
// Context FMHA custom mask buffer.
uint32_t const* fmha_custom_mask;
// Rotary inv_freq cache buffer to avoid re-computing.
float const* rotary_inv_freq;
// Rotary cos sin cache buffer to avoid re-computing.
float2 const* rotary_cos_sin;
int32_t input_seq_length; // padded input length
int32_t max_past_kv_len;
// By default, max_attention_window == cyclic_attention_window_size
// unless each layer has different cyclic kv cache length.
// Max cache capacity (used to allocate KV cache)
int32_t max_attention_window;
// Cyclic kv cache capacity (used to get the cyclic kv cache position for new tokens)
int32_t cyclic_attention_window_size;
int32_t sink_token_length;
int32_t const* q_seq_lengths;
int32_t const* kv_seq_lengths;
float const* kv_scale_orig_quant;
float const* kv_scale_quant_orig;
float const* attention_output_orig_quant;
T const* alibi_slopes;
void* context_buf;
void* key_value_cache;
kernels::KVBlockArray::DataType* block_offsets;
kernels::KVBlockArray::DataType* host_block_offsets;
void* host_primary_pool_pointer;
void* host_secondary_pool_pointer;
int32_t batch_size;
int32_t num_tokens;
int32_t max_blocks_per_sequence;
void* workspace;
// optional when relative position
T const* relative_attention_bias = nullptr;
int relative_attention_bias_stride = 0;
// optional when cross attention
T const* cross_qkv = nullptr;
int32_t cross_qkv_length = 0;
int32_t const* encoder_input_lengths = nullptr;
int32_t num_encoder_tokens = 0;
int64_t const* runtime_perf_knobs = nullptr;
std::string enqueueContextParamsToString() const
{
// variables from the params coming from the runtime
std::stringstream ss;
ss << "EnqueueContextParams ====================" << std::endl;
ss << "attention_input: " << attention_input << std::endl;
ss << "qkv_bias: " << qkv_bias << std::endl;
ss << "fmha_custom_mask: " << fmha_custom_mask << std::endl;
ss << "rotary_inv_freq: " << rotary_inv_freq << std::endl;
ss << "rotary_cos_sin: " << rotary_cos_sin << std::endl;
ss << "input_seq_length: " << input_seq_length << std::endl;
ss << "max_past_kv_len: " << max_past_kv_len << std::endl;
ss << "max_attention_window: " << max_attention_window << std::endl;
ss << "cyclic_attention_window_size: " << cyclic_attention_window_size << std::endl;
ss << "sink_token_length: " << sink_token_length << std::endl;
ss << "q_seq_lengths: "
<< *(runtime::ITensor::wrap(
(void*) q_seq_lengths, nvinfer1::DataType::kINT32, runtime::ITensor::makeShape({batch_size})))
<< std::endl;
ss << "kv_seq_lengths: "
<< *(runtime::ITensor::wrap(
(void*) kv_seq_lengths, nvinfer1::DataType::kINT32, runtime::ITensor::makeShape({batch_size})))
<< std::endl;
ss << "kv_scale_orig_quant: " << kv_scale_orig_quant << std::endl;
ss << "kv_scale_quant_orig: " << kv_scale_quant_orig << std::endl;
ss << "attention_output_orig_quant: " << attention_output_orig_quant << std::endl;
ss << "alibi_slopes: " << alibi_slopes << std::endl;
ss << "context_buf: " << context_buf << std::endl;
ss << "key_value_cache: " << (half*) key_value_cache << std::endl;
ss << "block_offsets: " << block_offsets << std::endl;
ss << "host_block_offsets: " << host_block_offsets << std::endl;
ss << "host_primary_pool_pointer: " << host_primary_pool_pointer << std::endl;
ss << "host_secondary_pool_pointer: " << host_secondary_pool_pointer << std::endl;
ss << "batch_size: " << batch_size << std::endl;
ss << "num_tokens: " << num_tokens << std::endl;
ss << "max_blocks_per_sequence: " << max_blocks_per_sequence << std::endl;
ss << "workspace: " << workspace << std::endl;
ss << "relative_attention_bias: " << relative_attention_bias << std::endl;
ss << "relative_attention_bias_stride: " << relative_attention_bias_stride << std::endl;
ss << "cross_qkv: " << cross_qkv << std::endl;
ss << "cross_qkv_length: " << cross_qkv_length << std::endl;
ss << "encoder_input_lengths: " << encoder_input_lengths << std::endl;
ss << "num_encoder_tokens: " << num_encoder_tokens << std::endl;
return ss.str();
}
};
template <typename T, typename KVCacheBuffer>
int enqueueContext(EnqueueContextParams<T, KVCacheBuffer> const& params, cudaStream_t stream);
template <typename T, typename KVCacheBuffer>
struct EnqueueGenerationParams
{
T const* attention_input;
T const* qkv_bias;
// Rotary inv_freq cache buffer to avoid re-computing.
float const* rotary_inv_freq;
// NOTE: input_seq_length might be larger than one in the medusa mode.
int32_t input_seq_length;
int32_t const* sequence_lengths;
int32_t max_past_kv_length;
int32_t beam_width;
int32_t const* context_lengths;
float const* kv_scale_orig_quant;
float const* kv_scale_quant_orig;
float const* attention_output_orig_quant;
T const* alibi_slopes;
void* context_buf;
void* key_value_cache;
kernels::KVBlockArray::DataType* block_offsets;
void* host_primary_pool_pointer;
void* host_secondary_pool_pointer;
// By default, max_attention_window == cyclic_attention_window_size
// unless each layer has different cyclic kv cache length.
// Max cache capacity (used to allocate KV cache)
int32_t max_attention_window;
// Cyclic kv cache capacity (used to get the cyclic kv cache position for new tokens)
int32_t cyclic_attention_window_size;
int32_t sink_token_length;
int32_t num_requests;
int32_t max_blocks_per_sequence;
int32_t const* cache_indir;
int32_t* semaphores;
void* workspace;
int32_t const* host_past_key_value_lengths;
// optional when relative position
T const* relative_attention_bias = nullptr;
int relative_attention_bias_stride = 0;
// optional when cross attention
int32_t const* encoder_input_lengths = nullptr;
int32_t const* host_context_lengths = nullptr;
// optional when speculative decoding is used.
bool const* spec_decoding_mask = nullptr;
int32_t const* spec_decoding_packed_mask = nullptr;
int32_t const* spec_decoding_position_offsets = nullptr;
int32_t const* spec_decoding_generation_lengths = nullptr;
bool spec_decoding_is_generation_length_variable = false;
int32_t spec_decoding_max_generation_length = 1;
int32_t total_num_input_tokens;
int64_t const* runtime_perf_knobs = nullptr;
};
template <typename T, typename KVCacheBuffer>
int enqueueGeneration(EnqueueGenerationParams<T, KVCacheBuffer> const& params, cudaStream_t stream);
// Called in configurePlugin().
template <typename T, typename KVCacheBuffer>
void prepareEnqueueGeneration(EnqueueGenerationParams<T, KVCacheBuffer> const& params);
template <typename T, typename KVCacheBuffer>
bool convertMMHAParamsToXQAParams(tensorrt_llm::kernels::XQAParams& xqaParams,
EnqueueGenerationParams<T, KVCacheBuffer> const& generationsParams, bool forConfigurePlugin);
bool isRelativePosition() const
{
return mPositionEmbeddingType == tensorrt_llm::kernels::PositionEmbeddingType::kRELATIVE;
}
bool isALiBi() const
{
return mPositionEmbeddingType == tensorrt_llm::kernels::PositionEmbeddingType::kALIBI
|| mPositionEmbeddingType == tensorrt_llm::kernels::PositionEmbeddingType::kALIBI_WITH_SCALE;
}
bool isAliBiWithScale() const
{
return mPositionEmbeddingType == tensorrt_llm::kernels::PositionEmbeddingType::kALIBI_WITH_SCALE;
}
bool isRoPE() const
{
return mPositionEmbeddingType == tensorrt_llm::kernels::PositionEmbeddingType::kROPE_GPTJ
|| mPositionEmbeddingType == tensorrt_llm::kernels::PositionEmbeddingType::kROPE_GPT_NEOX
|| mPositionEmbeddingType == tensorrt_llm::kernels::PositionEmbeddingType::kLONG_ROPE;
}
bool isLongRoPE() const
{
return mPositionEmbeddingType == tensorrt_llm::kernels::PositionEmbeddingType::kLONG_ROPE;
}
bool isCrossAttention() const
{
return mCrossAttention;
}
bool useKVCache() const
{
return mUseKVCache;
}
bool useCustomMask() const
{
return mMaskType == tensorrt_llm::kernels::AttentionMaskType::CUSTOM_MASK;
}
void reserveSemaphoreArray(int32_t size);
void debugCheckSemaphores(cudaStream_t stream);
protected:
static constexpr int kReservedMaxSeqLenTilePerSeq = 64;
const std::string mLayerName;
int mLayerIdx;
int mNumHeads;
int mVisionStart;
int mVisionLength;
int mNumKVHeads;
int mHeadSize;
int mUnidirectional;
float mQScaling;
float mQKTanhScale;
int mRotaryEmbeddingDim;
float mRotaryEmbeddingBase;
tensorrt_llm::kernels::RotaryScalingType mRotaryEmbeddingScaleType;
float mRotaryEmbeddingScale;
float mRotaryEmbeddingShortMscale;
float mRotaryEmbeddingLongMscale;
int mRotaryEmbeddingMaxPositions;
int mRotaryEmbeddingOriginalMaxPositions;
tensorrt_llm::kernels::PositionEmbeddingType mPositionEmbeddingType;
bool mRemovePadding = false;
tensorrt_llm::kernels::AttentionMaskType mMaskType;
tensorrt_llm::kernels::BlockSparseParams mBlockSparseParams;
// NOTE: default values for paged kv cache.
bool mPagedKVCache = false;
int mTokensPerBlock = 0;
tensorrt_llm::common::QuantMode mKVCacheQuantMode;
int mTpSize = 1;
int mTpRank = 0;
bool mUnfuseQkvGemm = false;
nvinfer1::DataType mType;
int32_t mMaxContextLength;
bool mQKVBiasEnabled;
bool mCrossAttention = false;
int mMaxDistance = 0;
bool mPosShiftEnabled = false;
bool mPagedContextFMHA = false;
bool mFP8ContextFMHA = false;
bool mDenseContextFMHA = false;
bool mIsSpecDecodingEnabled = false;
bool mSpecDecodingIsGenerationLengthVariable = false;
int32_t mSpecDecodingMaxGenerationLength = 1;
// Speculative decoding packed mask.
uint4* mSpecDecodingPackedMask;
uint4* mSpecDecodingPackedHostMask;
// fmha runner (disable by default)
// flag: disabled = 0, enabled = 1, enabled with fp32 accumulation = 2
bool mEnableContextFMHA = false;
bool mFMHAForceFP32Acc = false;
int mSM = tensorrt_llm::common::getSMVersion();
int mMultiProcessorCount = tensorrt_llm::common::getMultiProcessorCount();
int mMaxSharedMemoryPerBlockOptin = tensorrt_llm::common::getMaxSharedMemoryPerBlockOptin();
// The default copy constructor will leave it as nullptr. clone() shall initialize it.
std::shared_ptr<CUDADriverWrapper> mDriver;
UniqPtrWNullCopy<tensorrt_llm::kernels::FusedMHARunnerV2> mFMHARunner;
UniqPtrWNullCopy<tensorrt_llm::kernels::DecoderXQARunner> mDecoderXQARunner;
bool mMultiBlockMode;
bool mEnableXQA;
int mDeviceId = -1;
static bool mForceMultiBlockWarned;
// The default copy constructor will leave it as nullptr. clone() shall initialize it.
UniqPtrWNullCopy<tensorrt_llm::common::CublasMMWrapper> mCublasWrapper;
bool mUseKVCache = true;
// This is implementation details which we want to save when serializing, but not expose as
// a plugin field or a constructor parameter
int32_t mNbMultiBlockSemaphores = 0;
struct Deleter
{
void operator()(void* ptr)
{
cudaFree(ptr);
}
};
UniqPtrWNullCopy<int32_t[], Deleter> mMultiBlockSemaphores = {};
std::string toString() const
{
// member variables
std::stringstream ss;
ss << "gptAttentionCommon members ====================" << std::endl;
ss << "mNumHeads: " << mNumHeads << std::endl;
ss << "mNumKVHeads: " << mNumKVHeads << std::endl;
ss << "mHeadSize: " << mHeadSize << std::endl;
ss << "mUnidirectional: " << mUnidirectional << std::endl;
ss << "mQScaling: " << mQScaling << std::endl;
ss << "mRotaryEmbeddingDim: " << mRotaryEmbeddingDim << std::endl;
ss << "mRotaryEmbeddingBase: " << mRotaryEmbeddingBase << std::endl;
ss << "mRotaryEmbeddingScaleType: " << static_cast<int>(mRotaryEmbeddingScaleType) << std::endl;
ss << "mRotaryEmbeddingScale: " << mRotaryEmbeddingScale << std::endl;
ss << "mRotaryEmbeddingMaxPositions: " << mRotaryEmbeddingMaxPositions << std::endl;
ss << "mPositionEmbeddingType: " << static_cast<int>(mPositionEmbeddingType) << std::endl;
ss << "mRemovePadding: " << std::boolalpha << mRemovePadding << std::endl;
ss << "mMaskType: " << static_cast<int>(mMaskType) << std::endl;
ss << "mPagedKVCache: " << std::boolalpha << mPagedKVCache << std::endl;
ss << "mTokensPerBlock: " << mTokensPerBlock << std::endl;
ss << "mKVCacheQuantMode: " << static_cast<int>(mKVCacheQuantMode.value()) << std::endl;
ss << "mTpSize: " << mTpSize << std::endl;
ss << "mTpRank: " << mTpRank << std::endl;
ss << "mUnfuseQkvGemm: " << std::boolalpha << mUnfuseQkvGemm << std::endl;
ss << "mType: " << static_cast<int>(mType) << std::endl;
ss << "mMaxContextLength: " << mMaxContextLength << std::endl;
ss << "mQKVBiasEnabled: " << std::boolalpha << mQKVBiasEnabled << std::endl;
ss << "mCrossAttention: " << std::boolalpha << mCrossAttention << std::endl;
ss << "mMaxDistance: " << mMaxDistance << std::endl;
ss << "mPosShiftEnabled: " << std::boolalpha << mPosShiftEnabled << std::endl;
ss << "mPagedContextFMHA: " << std::boolalpha << mPagedContextFMHA << std::endl;
ss << "mFP8ContextFMHA: " << std::boolalpha << mFP8ContextFMHA << std::endl;
ss << "mDenseContextFMHA: " << std::boolalpha << mDenseContextFMHA << std::endl;
ss << "mEnableContextFMHA: " << std::boolalpha << mEnableContextFMHA << std::endl;
ss << "mFMHAForceFP32Acc: " << std::boolalpha << mFMHAForceFP32Acc << std::endl;
ss << "mSM: " << mSM << std::endl;
ss << "mMultiProcessorCount: " << mMultiProcessorCount << std::endl;
ss << "mMaxSharedMemoryPerBlockOptin: " << mMaxSharedMemoryPerBlockOptin << std::endl;
ss << "mMultiBlockMode: " << std::boolalpha << mMultiBlockMode << std::endl;
ss << "mEnableXQA: " << std::boolalpha << mEnableXQA << std::endl;
ss << "mDeviceId: " << mDeviceId << std::endl;
ss << "mUseKVCache: " << std::boolalpha << mUseKVCache << std::endl;
ss << "mForceMultiBlockWarned: " << mForceMultiBlockWarned << std::endl;
return ss.str();
}
};
class GPTAttentionPluginCreatorCommon : public BaseCreator
{
public:
GPTAttentionPluginCreatorCommon();
nvinfer1::PluginFieldCollection const* getFieldNames() noexcept override;
template <typename T>
T* deserializePluginImpl(char const* name, void const* serialData, size_t serialLength) noexcept;
protected:
std::vector<nvinfer1::PluginField> mPluginAttributes;
nvinfer1::PluginFieldCollection mFC{};
};
} // namespace tensorrt_llm::plugins
Reference in New Issue View Git Blame Copy Permalink