Skip to content
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
30 commits
Select commit Hold shift + click to select a range
374101f
cmake : enable building llama.cpp using system libggml (#12321)
ckastner Mar 17, 2025
2f21123
vulkan: Adjust coopmat2 tile sizes and selection heuristic (#12258)
jeffbolznv Mar 17, 2025
891c639
vulkan: Pad N dimension of B matrix for coopmat2 perf, to avoid bound…
jeffbolznv Mar 17, 2025
f07690c
vulkan: use fp32 in coopmat2 q4_k dequant function (#12309)
jeffbolznv Mar 17, 2025
cf2270e
vulkan: subgroup size tuning (#12087)
daniandtheweb Mar 17, 2025
484a8ab
vulkan: Add N/2 and N/4 optimized paths in coopmat2 shader (#12312)
jeffbolznv Mar 17, 2025
01e8f21
ggml-vulkan: remove unused find_program(glslc) (#12416)
guusw Mar 17, 2025
b1b132e
cuda : enable CUDA Graph on CUDA Toolkit < 12.x (#12394)
gaugarg-nv Mar 17, 2025
60c9029
docs : bring llama-cli conversation/template docs up-to-date (#12426)
CISC Mar 17, 2025
7dfad38
llama: Add support for RWKV v7 architecture (#12412)
MollySophia Mar 17, 2025
a53f7f7
fixed compilation warnings in ggml-sycl (#12424)
lslusarczyk Mar 18, 2025
fd123cf
Vulkan: Default to 1GB allocations instead of 4GB to avoid fragmentat…
0cc4m Mar 18, 2025
d9a1452
ggml : add SVE support for q6_K_q8_K (#12361)
fj-y-saito Mar 18, 2025
eba92d6
cmake : fix PowerPC build (#12241)
mehendarkarprajwal Mar 18, 2025
810e0af
server : fix warmup draft cache type (#12446)
ggerganov Mar 18, 2025
35cae5b
SYCL: using graphs is configurable by environment variable and compil…
lslusarczyk Mar 18, 2025
8551c44
context : always use non-causal attention for encoder graphs (#12447)
ggerganov Mar 18, 2025
99aa304
llama : add support for EXAONE tied word embeddings (#12451)
ngxson Mar 18, 2025
c6af216
speculative : fix seg fault in certain cases (#12454)
ggerganov Mar 18, 2025
29fff30
llama : support converting Mistral Small text-only (#12450)
ngxson Mar 18, 2025
bb115d2
musa: override warp_size of musa device to 32 (#12445)
yeahdongcn Mar 18, 2025
75422e8
graph : normalize Q, K, V shapes + sync cross attention (#12449)
ggerganov Mar 18, 2025
d84635b
opencl: improve profiling (#12442)
lhez Mar 18, 2025
c446b2e
vulkan: Submit once enough matmul work has been recorded (#12406)
jeffbolznv Mar 19, 2025
a686171
convert : Support chat_template.json (#12460)
CISC Mar 19, 2025
108e53c
llama : add support for GPT2, Bloom and CodeShell tied word embedding…
CISC Mar 19, 2025
0fd8487
Fix visionOS build and add CI (#12415)
guusw Mar 19, 2025
a9b5928
vulkan: optimize iq1 coopmat2 dequant functions (#12427)
jeffbolznv Mar 19, 2025
517b5dd
CUDA: Improve flash decoding kernel GPU occupancy for BS=1 case (#12183)
gaugarg-nv Mar 19, 2025
568013d
context : clear sets containing encoder output sequence ids before st…
fairydreaming Mar 19, 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
29 changes: 29 additions & 0 deletions .github/workflows/build.yml
Original file line number Diff line number Diff line change
Expand Up @@ -676,6 +676,35 @@ jobs:
-DCMAKE_XCODE_ATTRIBUTE_DEVELOPMENT_TEAM=ggml
cmake --build build --config Release -j $(sysctl -n hw.logicalcpu) -- CODE_SIGNING_ALLOWED=NO

macOS-latest-cmake-visionos:
runs-on: macos-latest

steps:
- name: Clone
id: checkout
uses: actions/checkout@v4

- name: Dependencies
id: depends
continue-on-error: true
run: |
brew update

- name: Build
id: cmake_build
run: |
sysctl -a
cmake -B build -G Xcode \
-DGGML_METAL_USE_BF16=ON \
-DGGML_METAL_EMBED_LIBRARY=ON \
-DLLAMA_BUILD_EXAMPLES=OFF \
-DLLAMA_BUILD_TESTS=OFF \
-DLLAMA_BUILD_SERVER=OFF \
-DCMAKE_SYSTEM_NAME=visionOS \
-DCMAKE_OSX_DEPLOYMENT_TARGET=1.0 \
-DCMAKE_XCODE_ATTRIBUTE_DEVELOPMENT_TEAM=ggml
cmake --build build --config Release -j $(sysctl -n hw.logicalcpu) -- CODE_SIGNING_ALLOWED=NO

macOS-latest-swift:
runs-on: macos-latest

Expand Down
10 changes: 9 additions & 1 deletion CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -29,6 +29,8 @@ else()
set(LLAMA_STANDALONE OFF)
endif()

option(LLAMA_USE_SYSTEM_GGML "Use system libggml" OFF)

if (EMSCRIPTEN)
set(BUILD_SHARED_LIBS_DEFAULT OFF)

Expand Down Expand Up @@ -145,7 +147,13 @@ endif()
# 3rd-party
#

if (NOT TARGET ggml)
if (LLAMA_USE_SYSTEM_GGML)
message(STATUS "Using system-provided libggml, skipping ggml build")
find_package(ggml REQUIRED)
add_library(ggml ALIAS ggml::ggml)
endif()

if (NOT TARGET ggml AND NOT LLAMA_USE_SYSTEM_GGML)
add_subdirectory(ggml)
# ... otherwise assume ggml is added by a parent CMakeLists.txt
endif()
Expand Down
8 changes: 4 additions & 4 deletions build-xcframework.sh
Original file line number Diff line number Diff line change
Expand Up @@ -432,8 +432,8 @@ cmake -B build-visionos -G Xcode \
-DCMAKE_SYSTEM_NAME=visionOS \
-DCMAKE_OSX_SYSROOT=xros \
-DCMAKE_XCODE_ATTRIBUTE_SUPPORTED_PLATFORMS=xros \
-DCMAKE_C_FLAGS="-D_XOPEN_SOURCE=700 -Du_int=unsigned\ int -Du_char=unsigned\ char -Du_short=unsigned\ short ${COMMON_C_FLAGS}" \
-DCMAKE_CXX_FLAGS="-D_XOPEN_SOURCE=700 -Du_int=unsigned\ int -Du_char=unsigned\ char -Du_short=unsigned\ short ${COMMON_CXX_FLAGS}" \
-DCMAKE_C_FLAGS="-D_XOPEN_SOURCE=700 ${COMMON_C_FLAGS}" \
-DCMAKE_CXX_FLAGS="-D_XOPEN_SOURCE=700 ${COMMON_CXX_FLAGS}" \
-S .
cmake --build build-visionos --config Release -- -quiet

Expand All @@ -445,8 +445,8 @@ cmake -B build-visionos-sim -G Xcode \
-DCMAKE_SYSTEM_NAME=visionOS \
-DCMAKE_OSX_SYSROOT=xrsimulator \
-DCMAKE_XCODE_ATTRIBUTE_SUPPORTED_PLATFORMS=xrsimulator \
-DCMAKE_C_FLAGS="-D_XOPEN_SOURCE=700 -Du_int=unsigned\ int -Du_char=unsigned\ char -Du_short=unsigned\ short ${COMMON_C_FLAGS}" \
-DCMAKE_CXX_FLAGS="-D_XOPEN_SOURCE=700 -Du_int=unsigned\ int -Du_char=unsigned\ char -Du_short=unsigned\ short ${COMMON_CXX_FLAGS}" \
-DCMAKE_C_FLAGS="-D_XOPEN_SOURCE=700 ${COMMON_C_FLAGS}" \
-DCMAKE_CXX_FLAGS="-D_XOPEN_SOURCE=700 ${COMMON_CXX_FLAGS}" \
-S .
cmake --build build-visionos-sim --config Release -- -quiet

Expand Down
2 changes: 2 additions & 0 deletions cmake/common.cmake
Original file line number Diff line number Diff line change
@@ -1,3 +1,5 @@
include("ggml/cmake/common.cmake")

function(llama_add_compile_flags)
if (LLAMA_FATAL_WARNINGS)
if (CMAKE_CXX_COMPILER_ID MATCHES "GNU" OR CMAKE_CXX_COMPILER_ID MATCHES "Clang")
Expand Down
Loading