Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Update TensorRT-LLM #1122

Merged
merged 2 commits into from
Feb 21, 2024
Merged

Update TensorRT-LLM #1122

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
980dc28
Update TensorRT-LLM
kaiyux Feb 21, 2024
d5a651d
update
Shixiaowei02 Feb 21, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions benchmarks/cpp/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -154,3 +154,5 @@ Take GPT-350M as an example for single GPU with static batching
--static_emulated_timeout 100 \
--dataset ../../benchmarks/cpp/preprocessed_dataset.json
```

`gptManagerBenchmark` can also be used with the high-level C++ API defined by the `executor::Executor` class (see `cpp/include/tensorrt_llm/executor/executor.h`). This can be done by passing the argument `--api executor`. Note that the Executor class is still under development and currently does not support models with tp or pp > 1.
393 changes: 327 additions & 66 deletions benchmarks/cpp/gptManagerBenchmark.cpp
View file
Open in desktop

Large diffs are not rendered by default.

14 changes: 14 additions & 0 deletions cpp/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -60,6 +60,20 @@ else()
message(STATUS "Importing batch manager")
endif()

if(EXISTS "${CMAKE_CURRENT_SOURCE_DIR}/tensorrt_llm/executor/CMakeLists.txt")
set(BUILD_EXECUTOR_DEFAULT ON)
else()
set(BUILD_EXECUTOR_DEFAULT OFF)
endif()

option(BUILD_EXECUTOR "Build executor from source" ${BUILD_EXECUTOR_DEFAULT})

if(BUILD_EXECUTOR)
message(STATUS "Building executor")
else()
message(STATUS "Importing executor")
endif()

if(BUILD_PYT)
message(STATUS "Building PyTorch")
else()
Expand Down
8 changes: 8 additions & 0 deletions cpp/include/tensorrt_llm/batch_manager/kvCacheConfig.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,7 @@

#pragma once

#include "tensorrt_llm/executor/executor.h"
#include "tensorrt_llm/runtime/common.h"

#include <optional>
Expand All @@ -42,6 +43,13 @@ class KvCacheConfig
{
}

explicit KvCacheConfig(executor::KvCacheConfig const& kvCacheConfig)
: KvCacheConfig(kvCacheConfig.getMaxTokens(), kvCacheConfig.getMaxAttentionWindow(),
kvCacheConfig.getSinkTokenLength(), kvCacheConfig.getFreeGpuMemoryFraction(),
kvCacheConfig.getEnableBlockReuse(), kvCacheConfig.getUseUvm())
{
}

std::optional<SizeType> maxTokens;
std::optional<SizeType> maxAttentionWindow;
std::optional<SizeType> sinkTokenLength;
Expand Down
Loading