Skip to content

Sync master with upstream release b5606 #119

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

Merged
merged 2 commits into from
Jun 9, 2025
Merged

Sync master with upstream release b5606 #119

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
5787b5d
ci: add LoongArch cross-compile build (#13944)
wojiushixiaobai Jun 7, 2025
247e5c6
cuda : fix buffer type check with integrated GPUs (#14069)
slaren Jun 8, 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
113 changes: 113 additions & 0 deletions .github/workflows/build-linux-cross.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -231,3 +231,116 @@ jobs:
-DCMAKE_FIND_ROOT_PATH_MODE_INCLUDE=BOTH
cmake --build build --config Release -j $(nproc)
debian-13-loongarch64-cpu-cross:
runs-on: ubuntu-24.04
container: debian@sha256:653dfb9f86c3782e8369d5f7d29bb8faba1f4bff9025db46e807fa4c22903671

steps:
- uses: actions/checkout@v4
- name: Setup LoongArch
run: |
rm -f /etc/apt/sources.list.d/*
cat << EOF | tee /etc/apt/sources.list.d/debian-ports.list
deb http://snapshot.debian.org/archive/debian/20250515T202920Z/ trixie main
EOF
( echo 'quiet "true";'; \
echo 'APT::Get::Assume-Yes "true";'; \
echo 'APT::Install-Recommends "false";'; \
echo 'Acquire::Check-Valid-Until "false";'; \
echo 'Acquire::Retries "5";'; \
) > /etc/apt/apt.conf.d/99snapshot-repos
apt-get update
apt-get install -y ca-certificates debian-ports-archive-keyring cmake git zip
dpkg --add-architecture loong64
# Add arch-specific repositories for non-amd64 architectures
cat << EOF | tee /etc/apt/sources.list.d/loong64-ports.list
deb [arch=loong64] http://snapshot.debian.org/archive/debian-ports/20250515T194251Z/ sid main
EOF
apt-get update || true ;# Prevent failure due to missing URLs.
apt-get install -y --no-install-recommends \
build-essential \
gcc-14-loongarch64-linux-gnu \
g++-14-loongarch64-linux-gnu
- name: Build
run: |
cmake -B build -DLLAMA_CURL=OFF \
-DCMAKE_BUILD_TYPE=Release \
-DGGML_OPENMP=OFF \
-DLLAMA_BUILD_EXAMPLES=ON \
-DLLAMA_BUILD_TOOLS=ON \
-DLLAMA_BUILD_TESTS=OFF \
-DCMAKE_SYSTEM_NAME=Linux \
-DCMAKE_SYSTEM_PROCESSOR=loongarch64 \
-DCMAKE_C_COMPILER=loongarch64-linux-gnu-gcc-14 \
-DCMAKE_CXX_COMPILER=loongarch64-linux-gnu-g++-14 \
-DCMAKE_POSITION_INDEPENDENT_CODE=ON \
-DCMAKE_FIND_ROOT_PATH=/usr/lib/loongarch64-linux-gnu \
-DCMAKE_FIND_ROOT_PATH_MODE_PROGRAM=NEVER \
-DCMAKE_FIND_ROOT_PATH_MODE_LIBRARY=ONLY \
-DCMAKE_FIND_ROOT_PATH_MODE_INCLUDE=BOTH
cmake --build build --config Release -j $(nproc)
debian-13-loongarch64-vulkan-cross:
runs-on: ubuntu-24.04
container: debian@sha256:653dfb9f86c3782e8369d5f7d29bb8faba1f4bff9025db46e807fa4c22903671

steps:
- uses: actions/checkout@v4
- name: Setup LoongArch
run: |
rm -f /etc/apt/sources.list.d/*
cat << EOF | tee /etc/apt/sources.list.d/debian-ports.list
deb http://snapshot.debian.org/archive/debian/20250515T202920Z/ trixie main
EOF
( echo 'quiet "true";'; \
echo 'APT::Get::Assume-Yes "true";'; \
echo 'APT::Install-Recommends "false";'; \
echo 'Acquire::Check-Valid-Until "false";'; \
echo 'Acquire::Retries "5";'; \
) > /etc/apt/apt.conf.d/99snapshot-repos
apt-get update
apt-get install -y ca-certificates debian-ports-archive-keyring cmake git zip
dpkg --add-architecture loong64
# Add arch-specific repositories for non-amd64 architectures
cat << EOF | tee /etc/apt/sources.list.d/loong64-ports.list
deb [arch=loong64] http://snapshot.debian.org/archive/debian-ports/20250515T194251Z/ sid main
EOF
apt-get update || true ;# Prevent failure due to missing URLs.
apt-get install -y --no-install-recommends \
build-essential \
glslc \
gcc-14-loongarch64-linux-gnu \
g++-14-loongarch64-linux-gnu \
libvulkan-dev:loong64
- name: Build
run: |
cmake -B build -DLLAMA_CURL=OFF \
-DCMAKE_BUILD_TYPE=Release \
-DGGML_VULKAN=ON \
-DGGML_OPENMP=OFF \
-DLLAMA_BUILD_EXAMPLES=ON \
-DLLAMA_BUILD_TOOLS=ON \
-DLLAMA_BUILD_TESTS=OFF \
-DCMAKE_SYSTEM_NAME=Linux \
-DCMAKE_SYSTEM_PROCESSOR=loongarch64 \
-DCMAKE_C_COMPILER=loongarch64-linux-gnu-gcc-14 \
-DCMAKE_CXX_COMPILER=loongarch64-linux-gnu-g++-14 \
-DCMAKE_POSITION_INDEPENDENT_CODE=ON \
-DCMAKE_FIND_ROOT_PATH=/usr/lib/loongarch64-linux-gnu \
-DCMAKE_FIND_ROOT_PATH_MODE_PROGRAM=NEVER \
-DCMAKE_FIND_ROOT_PATH_MODE_LIBRARY=ONLY \
-DCMAKE_FIND_ROOT_PATH_MODE_INCLUDE=BOTH
cmake --build build --config Release -j $(nproc)
5 changes: 1 addition & 4 deletions ggml/src/ggml-cuda/ggml-cuda.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1144,7 +1144,6 @@ typedef void (*ggml_cuda_op_mul_mat_t)(
static cudaError_t ggml_cuda_cpy_tensor_2d(
void * dst, const struct ggml_tensor * src, int64_t i3, int64_t i2, int64_t i1_low, int64_t i1_high, cudaStream_t stream) {

GGML_ASSERT(ggml_backend_buffer_is_cuda(src->buffer));
const char * src_ptr = (const char *) src->data;
char * dst_ptr = (char *) dst;

Expand Down Expand Up @@ -1427,8 +1426,6 @@ static void ggml_cuda_op_mul_mat(
const int64_t nb2 = dst->nb[2];
const int64_t nb3 = dst->nb[3];

GGML_ASSERT(ggml_backend_buffer_is_cuda(dst->buffer));
GGML_ASSERT(ggml_backend_buffer_is_cuda(src1->buffer));
ggml_backend_cuda_buffer_context * src1_ctx = (ggml_backend_cuda_buffer_context *) src1->buffer->context;
ggml_backend_cuda_buffer_context * dst_ctx = (ggml_backend_cuda_buffer_context *) dst->buffer->context;

Expand Down Expand Up @@ -1750,7 +1747,7 @@ static void ggml_cuda_mul_mat_batched_cublas(ggml_backend_cuda_context & ctx, co
GGML_ASSERT(!ggml_is_transposed(src0));
GGML_ASSERT(!ggml_is_transposed(src1));

GGML_ASSERT(ggml_backend_buffer_is_cuda(src0->buffer));
GGML_ASSERT(!ggml_backend_buft_is_cuda_split(src0->buffer->buft));
GGML_ASSERT(src0->type == GGML_TYPE_F16);

// Byte offsets and tensor dimensions are currently used in an inconsistent way for dst.
Expand Down