Skip to content

Commit 8b66876

Browse files
committed
[None][feat] DSv4 prep: compressor and mHC primitives
Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.com>
1 parent 20b6068 commit 8b66876

23 files changed

Lines changed: 12630 additions & 1 deletion

File tree

cpp/tensorrt_llm/CMakeLists.txt

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -200,6 +200,8 @@ set(TRTLLM_LINK_LIBS
200200
layers_src
201201
runtime_src
202202
testing_src
203+
mhcKernels_src
204+
compressorKernels_src
203205
userbuffers_src
204206
${DECODER_SHARED_TARGET_0}
205207
${DECODER_SHARED_TARGET_1})

cpp/tensorrt_llm/kernels/CMakeLists.txt

Lines changed: 6 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -30,6 +30,8 @@ add_subdirectory(dsv3MinLatencyKernels)
3030
add_subdirectory(causalConv1d)
3131
add_subdirectory(fusedGatedRMSNormQuant)
3232
add_subdirectory(mamba2MTPSSMCache)
33+
add_subdirectory(mhcKernels)
34+
add_subdirectory(compressorKernels)
3335

3436
file(GLOB_RECURSE SRC_CPP *.cpp)
3537
file(GLOB_RECURSE SRC_CU *.cu)
@@ -55,6 +57,10 @@ list(FILTER SRC_CU EXCLUDE REGEX "userbuffers/.*")
5557
list(FILTER SRC_CU EXCLUDE REGEX "fusedLayernormKernels/.*")
5658
list(FILTER SRC_CU EXCLUDE REGEX "fusedGatedRMSNormQuant/.*")
5759
list(FILTER SRC_CU EXCLUDE REGEX "mamba2MTPSSMCache/.*")
60+
list(FILTER SRC_CPP EXCLUDE REGEX "mhcKernels/.*")
61+
list(FILTER SRC_CU EXCLUDE REGEX "mhcKernels/.*")
62+
list(FILTER SRC_CPP EXCLUDE REGEX "compressorKernels/.*")
63+
list(FILTER SRC_CU EXCLUDE REGEX "compressorKernels/.*")
5864

5965
if(NOT ENABLE_MULTI_DEVICE)
6066
list(FILTER SRC_CU EXCLUDE REGEX "customAllReduceKernels*.*cu$")
Lines changed: 25 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,25 @@
1+
#
2+
# SPDX-FileCopyrightText: Copyright (c) 1993-2026 NVIDIA CORPORATION &
3+
# AFFILIATES. All rights reserved. SPDX-License-Identifier: Apache-2.0
4+
#
5+
# Licensed under the Apache License, Version 2.0 (the "License"); you may not
6+
# use this file except in compliance with the License. You may obtain a copy of
7+
# the License at
8+
#
9+
# http://www.apache.org/licenses/LICENSE-2.0
10+
#
11+
# Unless required by applicable law or agreed to in writing, software
12+
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
13+
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
14+
# License for the specific language governing permissions and limitations under
15+
# the License.
16+
#
17+
18+
set(SRC_CU compressorKernels.cu)
19+
20+
add_library(compressorKernels_src OBJECT ${SRC_CU})
21+
set_property(TARGET compressorKernels_src PROPERTY POSITION_INDEPENDENT_CODE ON)
22+
set_property(TARGET compressorKernels_src PROPERTY CUDA_RESOLVE_DEVICE_SYMBOLS
23+
ON)
24+
target_compile_options(compressorKernels_src
25+
PRIVATE $<$<COMPILE_LANGUAGE:CUDA>:--use_fast_math>)

0 commit comments

Comments
 (0)