Skip to content

PR: Refine ggml-hexagon backend(Qualcomm Hexagon NPU backend) for latest ggml,whisper.cpp,llama.cpp #12326

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.

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

Open
wants to merge 138 commits into
base: master
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
138 commits
Select commit Hold shift + click to select a range
e43e1f7
ggml-qnn: add Qualcomm QNN backend for GGML
zhouwg Feb 14, 2025
97c7f0c
ggml-qnn: santiy check
zhouwg Feb 15, 2025
5992fb3
ggml-qnn: update script build-run-android.sh to compare peformance of…
zhouwg Feb 16, 2025
ce53f15
ggml-qnn: fix minor issue in test-backend-ops.cpp
zhouwg Feb 17, 2025
223d4ab
ggml-qnn: merge QNN RPC feature from https://github.com/zhouwg/kantv/…
zhouwg Feb 18, 2025
67bd15c
ggml-qnn: sync from branch kantvai-ggmlqnn-npurpc
zhouwg Feb 18, 2025
f107318
ggml-qnn: a concise approach to offload mulmat to QNN backend(sync fr…
zhouwg Feb 19, 2025
aa31e75
ggml-qnn: remove redundant codes
zhouwg Feb 20, 2025
48cf633
ggml-qnn: sync from branch kantvai-ggmlqnn-npurpc
zhouwg Feb 20, 2025
0dea928
ggml-qnn: sync from branch kantvai-ggmlqnn-npurpc
zhouwg Feb 20, 2025
c188b89
ggml-qnn: sync from branch kantvai-ggmlqnn-npurpc
zhouwg Feb 21, 2025
a2874fe
ggml-qnn: add Qualcomm QNN backend for GGML
zhouwg Feb 14, 2025
cb631db
ggml-qnn: merge QNN RPC feature from https://github.com/zhouwg/kantv/…
zhouwg Feb 18, 2025
4b551ef
ggml-qnn: sync from branch kantvai-ggmlqnn-npurpc
zhouwg Feb 18, 2025
3bdf29e
ggml-qnn: a concise approach to offload mulmat to QNN backend(sync fr…
zhouwg Feb 19, 2025
ee62f54
ggml-qnn: remove redundant codes
zhouwg Feb 20, 2025
8bfb5e5
ggml-qnn: sync from branch kantvai-ggmlqnn-npurpc
zhouwg Feb 20, 2025
5ff6afc
ggml-qnn: sync from branch kantvai-ggmlqnn-npurpc
zhouwg Feb 20, 2025
d09a337
ggml-qnn: sync from branch kantvai-ggmlqnn-npurpc
zhouwg Feb 21, 2025
b9db451
ggml-qnn: fix a minior typo in internal doc
zhouwg Feb 23, 2025
fe58211
ggml-qnn: refine function ggml_qnn_create_general_tensor() to avoid c…
zhouwg Feb 23, 2025
08be259
ggml-qnn: fix a minor typo in source code
zhouwg Feb 24, 2025
44c9c2e
build: avoid ggml-qnn backend breaking other backend's builds
zhouwg Feb 24, 2025
e983504
ggml-qnn: remove redundant codes to make PR reviewers happy
zhouwg Feb 25, 2025
1f21af8
ggml-qnn: refine code format
zhouwg Feb 25, 2025
6351541
ggml-qnn: offload quantized type mulmat to QNN backend
zhouwg Feb 26, 2025
a0f93b4
ggml-qnn: refine source code structure to make code more clearly
zhouwg Feb 27, 2025
77ecf7f
ggml-qnn: enable release build with necessary logs to make reviewers …
zhouwg Feb 27, 2025
99ac2e1
ggml-qnn: enable all quantize type with 2d mulmat
zhouwg Feb 27, 2025
4c88a67
ggml-qnn: enable log output of GGMLQNN_LOG_INFO in command line mode …
zhouwg Feb 28, 2025
ea5098c
ggml-qnn: Windows port --- step2
zhouwg Feb 28, 2025
bf24d68
ggml-qnn: merge UT code and corresponding script from local dev branc…
zhouwg Mar 2, 2025
21ee81f
ggml-qnn: merge ggml_qnn_mul_mat_4d from local dev branch to make wor…
zhouwg Mar 2, 2025
03c6124
ggml-qnn: submit AI-assisted ggml_qnn_mul_mat_4d(not worked currently…
zhouwg Mar 2, 2025
01ef583
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- step2
zhouwg Mar 2, 2025
b322e29
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- step3
zhouwg Mar 2, 2025
c9fdbd9
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- step4
zhouwg Mar 2, 2025
52fc786
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- step5
zhouwg Mar 2, 2025
05c6a0e
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- step6
zhouwg Mar 2, 2025
978ca21
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- step7
zhouwg Mar 2, 2025
de8b126
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- step8
zhouwg Mar 2, 2025
a9a8514
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- good in step9
zhouwg Mar 2, 2025
78b8f56
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- narrow down t…
zhouwg Mar 2, 2025
2228e55
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- step10
zhouwg Mar 2, 2025
9cb57c4
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- narrow down t…
zhouwg Mar 2, 2025
0ec4332
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- step11
zhouwg Mar 2, 2025
960eedf
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 --- both ok in st…
zhouwg Mar 2, 2025
71ab59c
ggml-qnn: AI-assisted ggml_qnn_mul_mat_4d by Grok 3 ---finalizing ver…
zhouwg Mar 2, 2025
0f1364c
ggml-qnn: refine ggml_qnn_mul_mat and ggml_qnn_general_node according…
zhouwg Mar 2, 2025
b6b56ac
ggml-qnn: remove no-needed comments
zhouwg Mar 2, 2025
d226ca3
ggml-qnn: Windows port --- step3
zhouwg Mar 3, 2025
857c675
ggml-qnn: remove un-needed function
zhouwg Mar 4, 2025
50ae6b7
ggml-qnn:rebase to upstream
zhouwg Mar 4, 2025
014e63c
ggml-qnn: fix a minior issue during rebase to upstream
zhouwg Mar 4, 2025
938d872
ggml-qnn: update script according to https://github.com/ggml-org/llam…
zhouwg Mar 4, 2025
136f2c2
ggml-qnn: fix a minior issue in ggmlqnn_create_general_tensor()
zhouwg Mar 4, 2025
17c2c17
ggml-qnn: active member variable _device_id in class qnn_instance
zhouwg Mar 4, 2025
823ce8c
ggml-qnn: refine ggml_qnn_general_node and ggml_qnn_mul_mat to make c…
zhouwg Mar 4, 2025
2ede6e8
ggml-qnn: Windows port --- step4
zhouwg Mar 6, 2025
7fe0b7c
ggml-qnn: Windows port -- step5
zhouwg Mar 7, 2025
d3cbb11
ggml-qnn: WoA(Windows on ARM) -- step6
zhouwg Mar 8, 2025
90fe6f9
ggml-qnn: rebase to upstream
zhouwg Mar 9, 2025
d72c9b7
ggml-qnn: pr to upstream
zhouwg Mar 11, 2025
deb3c37
ggml-qnn: rebase to upstream
zhouwg Mar 18, 2025
276931d
ggml-qnn: self code-review
zhouwg Mar 18, 2025
507ea29
ggml-qnn: rebase upstream
zhouwg Mar 19, 2025
33a0905
ggml-qnn: add approach through Hexagon cDSP
zhouwg Mar 22, 2025
3cc58df
ggml-qnn: refine general approach through Hexagon cDSP
zhouwg Mar 23, 2025
245c18b
ggml-qnn: refine the entire ggml-qnn.cpp to make code more clear
zhouwg Mar 24, 2025
13a6a26
ggml-qnn: refine the entire ggml-qnn.cpp to make code more clear
zhouwg Mar 24, 2025
24f6f09
ggml-qnn: add build script for libggmlop_skel.so
zhouwg Mar 24, 2025
db4ccae
ggml-qnn: remove redundant functions in this PR and make codes more c…
zhouwg Mar 25, 2025
3ccbade
ggml-qnn: original ggml_compute_forward_add and ggml_compute_forward_…
zhouwg Mar 25, 2025
e2e1fc0
ggml-qnn: modify build-run-android.sh to verify mulmat and validate m…
zhouwg Mar 25, 2025
85f8a79
ggml-qnn: make host code(ggml-qnn.cpp) more clear and more stable
zhouwg Mar 26, 2025
81bbe73
ggml-qnn: refine code according to self code-review and make code mor…
zhouwg Mar 26, 2025
afb91be
ggml-qnn: offload more ggml op to Hexagon cDSP
zhouwg Mar 27, 2025
a6b4886
ggml-hexagon: code on AP(arm-cpu) side is stable now
zhouwg Mar 28, 2025
7a29b85
ggml-hexagon: optimize GGML_OP_ADD on cDSP side
zhouwg Mar 28, 2025
f42cce7
ggml-hexagon: simplify hexagon-kernel build logic in CMakeLists.txt
zhouwg Mar 29, 2025
c70bfa6
ggml-hexagon: release ggml-hexagon v0.98
zhouwg Mar 29, 2025
fff8309
ggml-hexagon: release ggml-hexagon v0.99
zhouwg Mar 29, 2025
e70da3d
ggml-hexagon: try to offload q6_k mulmat to cDSP
zhouwg Mar 29, 2025
2886895
ggml-hexagon: fix minior issue in ggml-hexagon.cpp after self code-re…
zhouwg Mar 29, 2025
09ae717
ggml-hexagon: check validation of ggml-hexagon.cfg before create appr…
zhouwg Mar 30, 2025
7ca0db5
ggml-hexagon: fix all compiler warnings in ggml-hexagon.cpp
zhouwg Mar 30, 2025
5d1892f
ggml-hexagon: enable only one backend device for HWACCEL_CDSP and ena…
zhouwg Mar 31, 2025
cd8ea99
ggml-hexagon: rpc ion memory pool and test-backend-ops works fine in …
zhouwg Mar 31, 2025
f50f4d6
ggml-hexagon: make comprision of mulmat performance between HWACCEL_Q…
zhouwg Mar 31, 2025
90d47b1
ggml-hexagon: release ggml-hexagon v1.00
zhouwg Mar 31, 2025
b317fb6
ggml-hexagon: rebase to upstream
zhouwg Apr 1, 2025
827c7e2
ggml-hexagon: check configuration of enable_rpc_dma_mempool in functi…
zhouwg Apr 1, 2025
6ae407e
ggml-hexagon: uniform rpc_ion_memsize and rpc_ion_usage between HWACC…
zhouwg Apr 1, 2025
cbef30a
ggml-hexagon: make buffer mechanism more clear in HWACCEL_CDSP approach
zhouwg Apr 1, 2025
2a5b64b
ggml-hexagon: add perf function in hexagon kernerls on cDSP side
zhouwg Apr 2, 2025
b784d60
ggml-hexagon: fix a stupid issue of why set rpc latency failure and i…
zhouwg Apr 2, 2025
4d72e5b
ggml-hexagon: make helper function ggmlhexagon_get_timestring() threa…
zhouwg Apr 2, 2025
91a62d0
ggml-hexagon: fix a typo in ggml-hexagon.cpp
zhouwg Apr 2, 2025
b113df8
ggml-hexagon: list all known todo and fixme tasks in ggml-hexagon.cpp
zhouwg Apr 2, 2025
683c68a
ggml-hexagon: fix units MB -> MiB
zhouwg Apr 2, 2025
eba13be
ggml-hexagon: try to make ggml-hexagon backend works fine in a standa…
zhouwg Apr 3, 2025
643ac87
ggml-hexagon: remove reduament code and make debug log more clear
zhouwg Apr 3, 2025
781b108
ggml-hexagon: add gemma-3-4b-it-Q8_0.gguf to verify q8_0 mulmat on cDSP
zhouwg Apr 3, 2025
2022f1a
ggml-hexagon:add skeleton code of offload GGML_OP_SOFT_MAX/GGML_OP_RM…
zhouwg Apr 3, 2025
34fe12c
ggml-hexagon: release ggml-dsp v0.60 on cDSP side
zhouwg Apr 4, 2025
06f5a32
ggml-hexagon: merge build logic in kernels/Makefile to ggml-hexagon/C…
zhouwg Apr 5, 2025
d049c3f
ggml-hexagon: fix a typo in ggml-hexagon.cpp
zhouwg Apr 5, 2025
2b5ab94
ggml-hexagon: uniform NDEBUG usage in ggml-hexagon.cpp and ggml-dsp.c
zhouwg Apr 6, 2025
c89dbf2
ggml-hexagon: add profiler feature for purpose of visualize NPU perfo…
zhouwg Apr 7, 2025
89e9b27
ggml-hexagon: remove so-called dma memory pool to avoid confusion and…
zhouwg Apr 8, 2025
79109ad
ggml-hexagon: make function ggmlhexagon_init_rpcmempool in ggml-hexag…
zhouwg Apr 8, 2025
4e0c383
ggml-hexagon: fix potential resource leak in class hexagon_profiler
zhouwg Apr 8, 2025
660a735
ggml-hexagon: enable multi-threading feature on cDSP side
zhouwg Apr 8, 2025
418243f
ggml-hexagon: upgrade QNN SDK to v2.33.0.250327
zhouwg Apr 9, 2025
4c1bda2
ggml-hexagon: fix typo in ggml-hexagon.cpp
zhouwg Apr 9, 2025
d0c4798
ggml-dsp: probe QuRT RTOS information in function ggmlop_dsp_open
zhouwg Apr 9, 2025
2bf2a9b
ggml-hexagon: setting enable_rpc_ion_mempool to 1 and make test-backe…
zhouwg Apr 10, 2025
0c03fec
ggml-hexagon: check whether user's specified htp arch is valid in CMa…
zhouwg Apr 10, 2025
4349b99
ggml-hexagon: sync with upstream
zhouwg Apr 11, 2025
c9f71c1
ggml-hexagon: refine pinned-memory feature
zhouwg Apr 11, 2025
50e0ea8
ggml-hexagon: refine build system in ggml-hexagon
zhouwg Apr 11, 2025
04b2011
ggml-hexagon: remove redundant code in struct ggml_backend_hexagon_bu…
zhouwg Apr 11, 2025
dfcb1c8
ggml-hexagon: upgrade Android NDK to android-ndk-r28
zhouwg Apr 11, 2025
7002bcd
ggml-dsp: split ggml-dsp.c into multiple files and cleanup
zhouwg Apr 11, 2025
7eb9f7f
ggml-dsp: refine ggml-dsp and make ggml-dsp more clear
zhouwg Apr 12, 2025
d5c8c64
ggml-hexagon: fix a minior issue in dev ops
zhouwg Apr 12, 2025
282e3f5
ggml-hexagon: fix a build issue in CI
zhouwg Apr 12, 2025
f74426c
ggml-dsp: cleanup code
zhouwg Apr 15, 2025
d2370f3
ggml-hexagon: sync with upstream
zhouwg Apr 15, 2025
3d1f3d6
ggml-dsp: cleanup code
zhouwg Apr 16, 2025
0f46871
ggml-dsp:refine ggmlhexagon_dsp_add_f32
zhouwg Apr 16, 2025
bca11dd
ggml-dsp: refine logic of thread_counts
zhouwg Apr 17, 2025
a67e529
ggml-hexagon: release v1.06 and ready for code review
zhouwg Apr 17, 2025
c48446e
ggml-dsp: make GGML_OP_ADD more faster on cDSP side
zhouwg Apr 19, 2025
bb70a8a
ggml-hexagon: sync from project kantv(make ggml-hexagon backend can w…
zhouwg Apr 24, 2025
a6bc78f
sync with upstream llama.cpp and sync ggml-hexagon.cpp from project k…
zhouwg Apr 29, 2025
3070b2f
ggml-hexagon: sync with upstream llama.cpp
zhouwg May 1, 2025
f82a2ca
ggml-hexagon: sync with upstream llama.cpp
zhouwg May 6, 2025
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
16 changes: 16 additions & 0 deletions CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,21 @@ set(CMAKE_WARN_UNUSED_CLI YES)

set(CMAKE_EXPORT_COMPILE_COMMANDS ON)

if(CMAKE_SYSTEM_NAME STREQUAL "Android")
if(DEFINED HTP_ARCH_VERSION)
if (${HTP_ARCH_VERSION} STREQUAL "v75" OR ${HTP_ARCH_VERSION} STREQUAL "v79")
#works fine on Snapdragon 8Gen3&8Elite with 1.5x - 3x performance gains with the default ggml backend
#refer to:https://github.com/kantv-ai/kantv/issues/287
set(OPT_FLAG " -O3 -march=armv8.7-a -flto -D_GNU_SOURCE -fvectorize -ffp-model=fast -fno-finite-math-only")
message("OPT_FLAG:${OPT_FLAG}")
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -DGGML_USE_HEXAGON ${DEBUG_FLAG} ${OPT_FLAG}")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DGGML_USE_HEXAGON ${DEBUG_FLAG} ${OPT_FLAG}")
set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} -DGGML_USE_HEXAGON ${DEBUG_FLAG} ${OPT_FLAG}")
set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -DGGML_USE_HEXAGON ${DEBUG_FLAG} ${OPT_FLAG}")
endif()
endif()
endif()

if (NOT XCODE AND NOT MSVC AND NOT CMAKE_BUILD_TYPE)
set(CMAKE_BUILD_TYPE Release CACHE STRING "Build type" FORCE)
set_property(CACHE CMAKE_BUILD_TYPE PROPERTY STRINGS "Debug" "Release" "MinSizeRel" "RelWithDebInfo")
Expand Down Expand Up @@ -120,6 +135,7 @@ llama_option_depr(WARNING LLAMA_RPC GGML_RPC)
llama_option_depr(WARNING LLAMA_SYCL GGML_SYCL)
llama_option_depr(WARNING LLAMA_SYCL_F16 GGML_SYCL_F16)
llama_option_depr(WARNING LLAMA_CANN GGML_CANN)
llama_option_depr(WARNING LLAMA_HEXAGON GGML_HEXAGON)

if (NOT MSVC)
if (LLAMA_SANITIZE_THREAD)
Expand Down
2 changes: 2 additions & 0 deletions ggml/CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -204,6 +204,7 @@ option(GGML_OPENCL_EMBED_KERNELS "ggml: embed kernels"
option(GGML_OPENCL_USE_ADRENO_KERNELS "ggml: use optimized kernels for Adreno" ON)
set (GGML_OPENCL_TARGET_VERSION "300" CACHE STRING
"gmml: OpenCL API version to target")
option(GGML_HEXAGON "ggml: use HEXAGON" OFF)

# toolchain for vulkan-shaders-gen
set (GGML_VULKAN_SHADERS_GEN_TOOLCHAIN "" CACHE FILEPATH "ggml: toolchain file for vulkan-shaders-gen")
Expand Down Expand Up @@ -269,6 +270,7 @@ set(GGML_PUBLIC_HEADERS
include/ggml-rpc.h
include/ggml-sycl.h
include/ggml-vulkan.h
include/ggml-hexagon.h
include/gguf.h)

set_target_properties(ggml PROPERTIES PUBLIC_HEADER "${GGML_PUBLIC_HEADERS}")
Expand Down
54 changes: 54 additions & 0 deletions ggml/include/ggml-hexagon.h
Original file line number Diff line number Diff line change
@@ -0,0 +1,54 @@
/*
* Copyright (c) 2024-2025 The ggml authors
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal in the Software without restriction, including without limitation the
* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
* sell copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
#pragma once

#include "ggml.h"
#include "ggml-backend.h"

#ifdef __cplusplus
extern "C" {
#endif

#define GGML_HEXAGON_MAX_DEVICES 4
#define GGML_HEXAGON_BACKEND_NAME "hexagon"

enum HEXAGONBackend {
HEXAGON_BACKEND_QNNCPU = 0,
HEXAGON_BACKEND_QNNGPU = 1,
HEXAGON_BACKEND_QNNNPU = 2,
HEXAGON_BACKEND_CDSP = 3,
HEXAGON_BACKEND_GGML = 4, //"fake" HEXAGON backend for compare performance between HEXAGON backend and ggml backend
};

GGML_BACKEND_API ggml_backend_t ggml_backend_hexagon_init(size_t dev_num, const char * qnn_lib_path);

GGML_BACKEND_API bool ggml_backend_is_hexagon(ggml_backend_t backend);

GGML_BACKEND_API int ggml_backend_hexagon_get_device_count(void);

GGML_BACKEND_API ggml_backend_reg_t ggml_backend_hexagon_reg(void);

const char * ggml_backend_hexagon_get_devname(size_t dev_num);

#ifdef __cplusplus
}
#endif
1 change: 1 addition & 0 deletions ggml/src/CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -313,6 +313,7 @@ ggml_add_backend(RPC)
ggml_add_backend(SYCL)
ggml_add_backend(Vulkan)
ggml_add_backend(OpenCL)
ggml_add_backend(HEXAGON)

foreach (target ggml-base ggml)
target_include_directories(${target} PUBLIC $<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}/../include> $<INSTALL_INTERFACE:include>)
Expand Down
8 changes: 8 additions & 0 deletions ggml/src/ggml-backend-reg.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -65,6 +65,10 @@
#include "ggml-kompute.h"
#endif

#ifdef GGML_USE_HEXAGON
#include "ggml-hexagon.h"
#endif

// disable C++17 deprecation warning for std::codecvt_utf8
#if defined(__clang__)
# pragma clang diagnostic push
Expand Down Expand Up @@ -187,6 +191,9 @@ struct ggml_backend_registry {
#ifdef GGML_USE_KOMPUTE
register_backend(ggml_backend_kompute_reg());
#endif
#ifdef GGML_USE_HEXAGON
register_backend(ggml_backend_hexagon_reg());
#endif
#ifdef GGML_USE_CPU
register_backend(ggml_backend_cpu_reg());
#endif
Expand Down Expand Up @@ -577,6 +584,7 @@ void ggml_backend_load_all_from_path(const char * dir_path) {
ggml_backend_load_best("vulkan", silent, dir_path);
ggml_backend_load_best("opencl", silent, dir_path);
ggml_backend_load_best("musa", silent, dir_path);
ggml_backend_load_best("hexagon", silent, dir_path);
ggml_backend_load_best("cpu", silent, dir_path);
// check the environment variable GGML_BACKEND_PATH to load an out-of-tree backend
const char * backend_path = std::getenv("GGML_BACKEND_PATH");
Expand Down
133 changes: 133 additions & 0 deletions ggml/src/ggml-hexagon/CMakeLists.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,133 @@
project(ggml-hexagon)
message(STATUS "Using HEXAGON backend")
message("CMAKE_SYSTEM_NAME : ${CMAKE_SYSTEM_NAME}")

set(CMAKE_CXX_STANDARD 20)
set(CMAKE_CXX_STANDARD_REQUIRED ON)

if(NOT DEFINED QNN_SDK_PATH)
message(FATAL_ERROR "QNN_SDK_PATH not defined")
endif()

if(NOT DEFINED HEXAGON_SDK_PATH)
message(FATAL_ERROR "HEXAGON_SDK_PATH not defined")
endif()

message("QNN_SDK_PATH : ${QNN_SDK_PATH}")
message("HEXAGON_SDK_PATH: ${HEXAGON_SDK_PATH}")
message("HTP_ARCH_VERSION: ${HTP_ARCH_VERSION}")

if (CMAKE_BUILD_TYPE STREQUAL "Debug")
set(DEBUG_FLAG "-Wall")
message("Debug mode:${DEBUG_FLAG}")
else()
set(DEBUG_FLAG "-DNDEBUG -Wall")
#manually disable all verbose logs in ggml-hexagon/CMakeLists.txt to
#make compare NPU performance through llama-bench more clear
#set(DEBUG_FLAG "-DNDEBUG -Wall -DDISABLE_ALL_LOG")
message("Release mode:${DEBUG_FLAG}")
endif()

#v68 --- Snapdragon 888
#v69 --- Snapdragon 8 Gen1
#v73 --- Snapdragon 8 Gen2
#v75 --- Snapdragon 8 Gen3
#v79 --- Snapdragon 8 Elite(aka Gen4)
if(NOT DEFINED HTP_ARCH_VERSION)
message(FATAL_ERROR "HTP_ARCH_VERSION not defined, valid htp arch: v68,v69,v73,v75,v79")
endif()

#check whether user's specified htp arch is valid
set(CHECK_HTP_ARCH "WRONG")
foreach (feat v68 v69 v73 v75 v79)
if (${feat} STREQUAL ${HTP_ARCH_VERSION})
set(CHECK_HTP_ARCH "GOOD")
endif()
endforeach()
if (${CHECK_HTP_ARCH} STREQUAL "WRONG")
message(FATAL_ERROR "ggml-hexagon backend only support htp arch v68,v69,v73,v75,v79")
endif()

#check optimization flags
set(OPT_FLAG " ")
if (${HTP_ARCH_VERSION} STREQUAL "v75" OR ${HTP_ARCH_VERSION} STREQUAL "v79")
#works fine on Snapdragon 8Gen3&8Elite with 1.5x - 3x performance gains with the default ggml backend
set(OPT_FLAG " -O3 -march=armv8.7-a -mcpu=cortex-x1 -mtune=cortex-x1 -flto -D_GNU_SOURCE -fvectorize -ffp-model=fast -fno-finite-math-only")
endif()
message("OPT_FLAG:${OPT_FLAG}")

if(CMAKE_SYSTEM_NAME STREQUAL "Android")
find_library(LOG_LIB log)

add_library(cdsprpc
SHARED
IMPORTED)
set_target_properties(cdsprpc
PROPERTIES
IMPORTED_LOCATION
${HEXAGON_SDK_PATH}/ipc/fastrpc/remote/ship/android_aarch64/libcdsprpc.so)

set(QNN_LINK_LIBRARIES ${LOG_LIB} cdsprpc)
set(QNN_DEFAULT_LIB_SEARCH_PATH "/data/local/tmp/" CACHE STRING "customized library search path for QNN backend")

include_directories(${HEXAGON_SDK_PATH}/incs)
include_directories(${HEXAGON_SDK_PATH}/incs/stddef)
include_directories(${HEXAGON_SDK_PATH}/ipc/fastrpc/incs)
include_directories(${HEXAGON_SDK_PATH}/ipc/fastrpc/rpcmem/inc)
include_directories(${HEXAGON_SDK_PATH}/ipc/fastrpc/remote/ship/android_Debug_aarch64)
include_directories(${HEXAGON_SDK_PATH}/utils/examples)
include_directories(${HEXAGON_SDK_PATH}/ipc/fastrpc/rtld/ship/android_aarch64)
include_directories(${HEXAGON_SDK_PATH}/libs/atomic/inc)
include_directories(${HEXAGON_SDK_PATH}/libs/atomic/android_Debug_aarch64/ship)
include_directories(${CMAKE_SOURCE_DIR}/ggml/src/ggml-hexagon/)
include_directories(${CMAKE_SOURCE_DIR}/ggml/src/ggml-hexagon/kernels/)
elseif(CMAKE_SYSTEM_NAME STREQUAL "Windows")
set(QNN_DEFAULT_LIB_SEARCH_PATH "C:\\" CACHE STRING "customized library search path for QNN backend")
else()
message(FATAL_ERROR "ggml-hexagon now only available on Android and Windows(Windows on ARM)")
endif()

set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -DGGML_USE_HEXAGON ${DEBUG_FLAG} ${OPT_FLAG}")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DGGML_USE_HEXAGON ${DEBUG_FLAG} ${OPT_FLAG}")
set(CMAKE_C_FLAGS_RELEASE "${CMAKE_C_FLAGS_RELEASE} -DGGML_USE_HEXAGON ${DEBUG_FLAG} ${OPT_FLAG}")
set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -DGGML_USE_HEXAGON ${DEBUG_FLAG} ${OPT_FLAG}")

file(GLOB HEXAGON_SOURCES "${CMAKE_CURRENT_LIST_DIR}/*.cpp" "${CMAKE_CURRENT_LIST_DIR}/kernels/stub.c")
ggml_add_backend_library(ggml-hexagon ${HEXAGON_SOURCES})

target_include_directories(ggml-hexagon PRIVATE ${QNN_SDK_PATH}/include/QNN ${HEXAGON_SDK_PATH} ${CMAKE_CURRENT_LIST_DIR})
target_link_libraries(ggml-hexagon PRIVATE ${QNN_LINK_LIBRARIES})

string(REGEX REPLACE "/$" "" QNN_DEFAULT_LIB_SEARCH_PATH "${QNN_DEFAULT_LIB_SEARCH_PATH}")
target_compile_definitions(ggml-hexagon PRIVATE QNN_DEFAULT_LIB_SEARCH_PATH="${QNN_DEFAULT_LIB_SEARCH_PATH}/")

#cross compiling source codes of hexagon kernels which running on cDSP side
function(ggml_hexagon_build_kernel KNAME)
message(STATUS "ggml_hexagon: build hexagon-kernel ${KNAME}")

add_custom_command(
TARGET ${PROJECT_NAME}
POST_BUILD
COMMAND echo "current working path:`pwd`\n"
COMMAND echo "${CMAKE_CURRENT_LIST_DIR}/kernels"
COMMAND make -C ${CMAKE_CURRENT_LIST_DIR}/kernels/ clean
COMMAND make -C ${CMAKE_CURRENT_LIST_DIR}/kernels/ HEXAGON_SDK_PATH=${HEXAGON_SDK_PATH} HTP_ARCH_VERSION=${HTP_ARCH_VERSION} DEBUG_FLAG=${DEBUG_FLAG}
COMMAND echo "current working path:`pwd`\n"
COMMAND ls -l ../../../bin/libggmlop-skel.so
COMMENT "build hexagon-kernel"
)
endfunction()

function(ggml_hexagon_setup_cfg KNAME)
message(STATUS "ggml_hexagon: setup runtime configuration file ${KNAME}")
add_custom_command(
TARGET ${PROJECT_NAME}
POST_BUILD
COMMAND echo "current working path:`pwd`\n"
COMMAND /bin/cp -fv ../../../../../scripts/${KNAME} ../../../bin/
COMMENT "setup runtime configuration file"
)
endfunction()

ggml_hexagon_build_kernel("cdsp")
ggml_hexagon_setup_cfg("ggml-hexagon.cfg")
Loading
Loading