github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #2120] How to install libnvidia-ml.so? #1211

New Issue
Closed
opened 2026-04-12 10:59:11 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-04-12 10:59:11 -05:00
Owner
Copy Link

Originally created by @silverwind63 on GitHub (Jan 21, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2120

Hi guys! I have been using ollama with ollama webui this month.However,it output

WARNING:

You should always run with libnvidia-ml.so that is installed with your
NVIDIA Display Driver. By default it's installed in /usr/lib and /usr/lib64.
libnvidia-ml.so in GDK package is a stub library that is attached only for
build purposes (e.g. machine that you build your application doesn't have
to have Display Driver installed).

And whenever I want to run any model(which is capable to load it and the speed is about 5 tokens/s) it will always run into cuda memory error.
My system:
RAM:16GB
GPU:3060ti 8GB
SYSTEM:archlinux
Kernel:6.7.0-arch3-1
Nvidia GPU Driver:nvidia-dkms 545.29.06-1
I have also installed following package which is related to nvidia:

lib32-nvidia-utils 545.29.06-1
libnvidia-container 1.14.3-1
libnvidia-container-tools 1.14.3-1
libva-nvidia-driver-git 0.0.11.r1.gea6d862-1
nvidia-container-toolkit 1.14.3-9
nvidia-docker-compose 0.1.6-1
nvidia-utils 545.29.06-1
Originally created by @silverwind63 on GitHub (Jan 21, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2120 Hi guys! I have been using ollama with ollama webui this month.However,it output ``` WARNING: You should always run with libnvidia-ml.so that is installed with your NVIDIA Display Driver. By default it's installed in /usr/lib and /usr/lib64. libnvidia-ml.so in GDK package is a stub library that is attached only for build purposes (e.g. machine that you build your application doesn't have to have Display Driver installed). ``` And whenever I want to run any model(which is capable to load it and the speed is about 5 tokens/s) it will always run into cuda memory error. My system: RAM:16GB GPU:3060ti 8GB SYSTEM:archlinux Kernel:6.7.0-arch3-1 Nvidia GPU Driver:nvidia-dkms 545.29.06-1 I have also installed following package which is related to nvidia: ``` lib32-nvidia-utils 545.29.06-1 libnvidia-container 1.14.3-1 libnvidia-container-tools 1.14.3-1 libva-nvidia-driver-git 0.0.11.r1.gea6d862-1 nvidia-container-toolkit 1.14.3-9 nvidia-docker-compose 0.1.6-1 nvidia-utils 545.29.06-1 ```
GiteaMirror commented 2026-04-12 10:59:12 -05:00
Author
Owner
Copy Link

@Nowai commented on GitHub (Jan 21, 2024):

I have the same Issue. Basically the same arch setup.
What is your output for Discovered GPU libraries?

INFO Discovered GPU libraries: [/opt/cuda/targets/x86_64-linux/lib/stubs/libnvidia-ml.so /usr/lib/libnvidia-ml.so.545.29.06 /usr/lib32/libnvidia-ml.so.545.29.06 /usr/lib64/libnvidia-ml.so.545.29.06]

It appears that it tries to load the wrong file, even though the correct one is available.

llama.cpp runs perfectly with GPU support.

EDIT:

With the help from Socialnetwooky from discord I could fix it. It seems that the ollama package is currently broken. The steps to fix it are as follows:

Make a new tmp folder.
Copy the PKGBUILD
Copy the remaining files from: https://gitlab.archlinux.org/archlinux/packaging/packages/ollama-cuda
execute makepkg -sri

PKGBUILD:

pkgname=ollama-cuda
pkgdesc='Create, run and share large language models (LLMs) with CUDA'
pkgver=0.1.20
pkgrel=3
arch=(x86_64)
url='https://github.com/jmorganca/ollama'
license=(MIT)
_ollamacommit=ab6be852c77064d7abeffb0b03c096aab90e95fe # tag: v0.1.20
# The llama.cpp git submodule commit hash can be found here:
# https://github.com/jmorganca/ollama/tree/v0.1.20/llm
_llama_cpp_commit=328b83de23b33240e28f4e74900d1d06726f5eb1
makedepends=(cmake cuda git go)
provides=(ollama)
conflicts=(ollama)
source=(git+$url#commit=$_ollamacommit
        llama.cpp::git+https://github.com/ggerganov/llama.cpp#commit=$_llama_cpp_commit
        sysusers.conf
        tmpfiles.d
        ollama.service)
b2sums=('SKIP'
        'SKIP'
	'SKIP'
	'SKIP'
	'SKIP')
#        '65d39053cd1dd09562473c2e58f66a447ce0225b32607685f60350596b3288d6568c1cb897393b20236260e632427de1a952e72fe358407020f6cc7820fd4f60'
#        '6f0b6886108e8d5f385bf7f9bebc60218797c53d4b88a69cc98564ad02c558cb86633b4f713ddd919d618146c04ac0a9215aa6c32ae192701af9d7850264dd56' 
#	'be72a39e823d6631095ce407c92af6aee8650302eeaaa55a970a43592daaa141369d3a5a5eb6a992f6c2f5461370228ac7a84f3318422419fd868575454487d6') 

prepare() {
  cd ${pkgname/-cuda}

  rm -frv llm/llama.cpp
  
  export GIN_MODE=release

  # Copy git submodule files instead of symlinking because the build process is sensitive to symlinks.
  cp -r "$srcdir/llama.cpp" llm/llama.cpp

  # Turn LTO on and set the build type to Release
  sed -i 's,T_CODE=on,T_CODE=on -D LLAMA_LTO=off -D CMAKE_BUILD_TYPE=Release,g' llm/generate/gen_linux.sh

  sed -i 's,var mode string = gin.DebugMode,var mode string = gin.ReleaseMode,g' server/routes.go
  # Let gen_linux.sh find libcudart.so
  sed -i 's,/usr/local/cuda/lib64,/opt/cuda/targets/x86_64-linux/lib,g' llm/generate/gen_linux.sh

  # Let gpu.go find libnvidia-ml.so from the cuda package
  #sed -i 's,/opt/cuda/lib64/libnvidia-ml.so*,/opt/cuda/targets/x86_64-linux/lib/stubs/libnvidia-ml.so*,g' gpu/gpu.go
  sed -i 's,/opt/cuda/lib64/libnvidia-ml.so*,/lib/libnvidia-ml.so*,g' gpu/gpu.go
}

build() {
  cd ${pkgname/-cuda}
  export CGO_CFLAGS="$CFLAGS" CGO_CPPFLAGS="$CPPFLAGS" CGO_CXXFLAGS="$CXXFLAGS" CGO_LDFLAGS="$LDFLAGS"
  go generate ./...
  go build -buildmode=pie -ldflags=-fno-lto -trimpath -mod=readonly -modcacherw -ldflags=-linkmode=external \
    -ldflags=-buildid='' -ldflags="-X=github.com/jmorganca/ollama/version.Version=$pkgver"
}

check() {
  cd ${pkgname/-cuda}
  go test ./api ./format
  ./ollama --version > /dev/null
}

package() {
  install -Dm755 ${pkgname/-cuda}/${pkgname/-cuda} "$pkgdir/usr/bin/${pkgname/-cuda}"
  install -dm700 "$pkgdir/var/lib/ollama"
  install -Dm644 ollama.service "$pkgdir/usr/lib/systemd/system/ollama.service"
  install -Dm644 sysusers.conf "$pkgdir/usr/lib/sysusers.d/ollama.conf"
  install -Dm644 tmpfiles.d "$pkgdir/usr/lib/tmpfiles.d/ollama.conf"
  install -Dm644 ${pkgname/-cuda}/LICENSE "$pkgdir/usr/share/licenses/$pkgname/LICENSE"
}
<!-- gh-comment-id:1902612730 --> @Nowai commented on GitHub (Jan 21, 2024): I have the same Issue. Basically the same arch setup. What is your output for Discovered GPU libraries? `INFO Discovered GPU libraries: [/opt/cuda/targets/x86_64-linux/lib/stubs/libnvidia-ml.so /usr/lib/libnvidia-ml.so.545.29.06 /usr/lib32/libnvidia-ml.so.545.29.06 /usr/lib64/libnvidia-ml.so.545.29.06]` It appears that it tries to load the wrong file, even though the correct one is available. llama.cpp runs perfectly with GPU support. EDIT: With the help from Socialnetwooky from discord I could fix it. It seems that the ollama package is currently broken. The steps to fix it are as follows: Make a new tmp folder. Copy the PKGBUILD Copy the remaining files from: https://gitlab.archlinux.org/archlinux/packaging/packages/ollama-cuda execute makepkg -sri PKGBUILD: ``` pkgname=ollama-cuda pkgdesc='Create, run and share large language models (LLMs) with CUDA' pkgver=0.1.20 pkgrel=3 arch=(x86_64) url='https://github.com/jmorganca/ollama' license=(MIT) _ollamacommit=ab6be852c77064d7abeffb0b03c096aab90e95fe # tag: v0.1.20 # The llama.cpp git submodule commit hash can be found here: # https://github.com/jmorganca/ollama/tree/v0.1.20/llm _llama_cpp_commit=328b83de23b33240e28f4e74900d1d06726f5eb1 makedepends=(cmake cuda git go) provides=(ollama) conflicts=(ollama) source=(git+$url#commit=$_ollamacommit llama.cpp::git+https://github.com/ggerganov/llama.cpp#commit=$_llama_cpp_commit sysusers.conf tmpfiles.d ollama.service) b2sums=('SKIP' 'SKIP' 'SKIP' 'SKIP' 'SKIP') # '65d39053cd1dd09562473c2e58f66a447ce0225b32607685f60350596b3288d6568c1cb897393b20236260e632427de1a952e72fe358407020f6cc7820fd4f60' # '6f0b6886108e8d5f385bf7f9bebc60218797c53d4b88a69cc98564ad02c558cb86633b4f713ddd919d618146c04ac0a9215aa6c32ae192701af9d7850264dd56' # 'be72a39e823d6631095ce407c92af6aee8650302eeaaa55a970a43592daaa141369d3a5a5eb6a992f6c2f5461370228ac7a84f3318422419fd868575454487d6') prepare() { cd ${pkgname/-cuda} rm -frv llm/llama.cpp export GIN_MODE=release # Copy git submodule files instead of symlinking because the build process is sensitive to symlinks. cp -r "$srcdir/llama.cpp" llm/llama.cpp # Turn LTO on and set the build type to Release sed -i 's,T_CODE=on,T_CODE=on -D LLAMA_LTO=off -D CMAKE_BUILD_TYPE=Release,g' llm/generate/gen_linux.sh sed -i 's,var mode string = gin.DebugMode,var mode string = gin.ReleaseMode,g' server/routes.go # Let gen_linux.sh find libcudart.so sed -i 's,/usr/local/cuda/lib64,/opt/cuda/targets/x86_64-linux/lib,g' llm/generate/gen_linux.sh # Let gpu.go find libnvidia-ml.so from the cuda package #sed -i 's,/opt/cuda/lib64/libnvidia-ml.so*,/opt/cuda/targets/x86_64-linux/lib/stubs/libnvidia-ml.so*,g' gpu/gpu.go sed -i 's,/opt/cuda/lib64/libnvidia-ml.so*,/lib/libnvidia-ml.so*,g' gpu/gpu.go } build() { cd ${pkgname/-cuda} export CGO_CFLAGS="$CFLAGS" CGO_CPPFLAGS="$CPPFLAGS" CGO_CXXFLAGS="$CXXFLAGS" CGO_LDFLAGS="$LDFLAGS" go generate ./... go build -buildmode=pie -ldflags=-fno-lto -trimpath -mod=readonly -modcacherw -ldflags=-linkmode=external \ -ldflags=-buildid='' -ldflags="-X=github.com/jmorganca/ollama/version.Version=$pkgver" } check() { cd ${pkgname/-cuda} go test ./api ./format ./ollama --version > /dev/null } package() { install -Dm755 ${pkgname/-cuda}/${pkgname/-cuda} "$pkgdir/usr/bin/${pkgname/-cuda}" install -dm700 "$pkgdir/var/lib/ollama" install -Dm644 ollama.service "$pkgdir/usr/lib/systemd/system/ollama.service" install -Dm644 sysusers.conf "$pkgdir/usr/lib/sysusers.d/ollama.conf" install -Dm644 tmpfiles.d "$pkgdir/usr/lib/tmpfiles.d/ollama.conf" install -Dm644 ${pkgname/-cuda}/LICENSE "$pkgdir/usr/share/licenses/$pkgname/LICENSE" } ```
GiteaMirror commented 2026-04-12 10:59:13 -05:00
Author
Owner
Copy Link

@silverwind63 commented on GitHub (Jan 24, 2024):

I have the same Issue. Basically the same arch setup. What is your output for Discovered GPU libraries?

INFO Discovered GPU libraries: [/opt/cuda/targets/x86_64-linux/lib/stubs/libnvidia-ml.so /usr/lib/libnvidia-ml.so.545.29.06 /usr/lib32/libnvidia-ml.so.545.29.06 /usr/lib64/libnvidia-ml.so.545.29.06]

It appears that it tries to load the wrong file, even though the correct one is available.

llama.cpp runs perfectly with GPU support.

EDIT:

With the help from Socialnetwooky from discord I could fix it. It seems that the ollama package is currently broken. The steps to fix it are as follows:

Make a new tmp folder. Copy the PKGBUILD Copy the remaining files from: https://gitlab.archlinux.org/archlinux/packaging/packages/ollama-cuda execute makepkg -sri

PKGBUILD:

pkgname=ollama-cuda
pkgdesc='Create, run and share large language models (LLMs) with CUDA'
pkgver=0.1.20
pkgrel=3
arch=(x86_64)
url='https://github.com/jmorganca/ollama'
license=(MIT)
_ollamacommit=ab6be852c77064d7abeffb0b03c096aab90e95fe # tag: v0.1.20
# The llama.cpp git submodule commit hash can be found here:
# https://github.com/jmorganca/ollama/tree/v0.1.20/llm
_llama_cpp_commit=328b83de23b33240e28f4e74900d1d06726f5eb1
makedepends=(cmake cuda git go)
provides=(ollama)
conflicts=(ollama)
source=(git+$url#commit=$_ollamacommit
        llama.cpp::git+https://github.com/ggerganov/llama.cpp#commit=$_llama_cpp_commit
        sysusers.conf
        tmpfiles.d
        ollama.service)
b2sums=('SKIP'
        'SKIP'
	'SKIP'
	'SKIP'
	'SKIP')
#        '65d39053cd1dd09562473c2e58f66a447ce0225b32607685f60350596b3288d6568c1cb897393b20236260e632427de1a952e72fe358407020f6cc7820fd4f60'
#        '6f0b6886108e8d5f385bf7f9bebc60218797c53d4b88a69cc98564ad02c558cb86633b4f713ddd919d618146c04ac0a9215aa6c32ae192701af9d7850264dd56' 
#	'be72a39e823d6631095ce407c92af6aee8650302eeaaa55a970a43592daaa141369d3a5a5eb6a992f6c2f5461370228ac7a84f3318422419fd868575454487d6') 

prepare() {
  cd ${pkgname/-cuda}

  rm -frv llm/llama.cpp
  
  export GIN_MODE=release

  # Copy git submodule files instead of symlinking because the build process is sensitive to symlinks.
  cp -r "$srcdir/llama.cpp" llm/llama.cpp

  # Turn LTO on and set the build type to Release
  sed -i 's,T_CODE=on,T_CODE=on -D LLAMA_LTO=off -D CMAKE_BUILD_TYPE=Release,g' llm/generate/gen_linux.sh

  sed -i 's,var mode string = gin.DebugMode,var mode string = gin.ReleaseMode,g' server/routes.go
  # Let gen_linux.sh find libcudart.so
  sed -i 's,/usr/local/cuda/lib64,/opt/cuda/targets/x86_64-linux/lib,g' llm/generate/gen_linux.sh

  # Let gpu.go find libnvidia-ml.so from the cuda package
  #sed -i 's,/opt/cuda/lib64/libnvidia-ml.so*,/opt/cuda/targets/x86_64-linux/lib/stubs/libnvidia-ml.so*,g' gpu/gpu.go
  sed -i 's,/opt/cuda/lib64/libnvidia-ml.so*,/lib/libnvidia-ml.so*,g' gpu/gpu.go
}

build() {
  cd ${pkgname/-cuda}
  export CGO_CFLAGS="$CFLAGS" CGO_CPPFLAGS="$CPPFLAGS" CGO_CXXFLAGS="$CXXFLAGS" CGO_LDFLAGS="$LDFLAGS"
  go generate ./...
  go build -buildmode=pie -ldflags=-fno-lto -trimpath -mod=readonly -modcacherw -ldflags=-linkmode=external \
    -ldflags=-buildid='' -ldflags="-X=github.com/jmorganca/ollama/version.Version=$pkgver"
}

check() {
  cd ${pkgname/-cuda}
  go test ./api ./format
  ./ollama --version > /dev/null
}

package() {
  install -Dm755 ${pkgname/-cuda}/${pkgname/-cuda} "$pkgdir/usr/bin/${pkgname/-cuda}"
  install -dm700 "$pkgdir/var/lib/ollama"
  install -Dm644 ollama.service "$pkgdir/usr/lib/systemd/system/ollama.service"
  install -Dm644 sysusers.conf "$pkgdir/usr/lib/sysusers.d/ollama.conf"
  install -Dm644 tmpfiles.d "$pkgdir/usr/lib/tmpfiles.d/ollama.conf"
  install -Dm644 ${pkgname/-cuda}/LICENSE "$pkgdir/usr/share/licenses/$pkgname/LICENSE"
}

When I run makepkg -sri to install it,it show me these errors:

-- The C compiler identification is GNU 13.2.1
-- The CXX compiler identification is GNU 13.2.1
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found Git: /usr/bin/git (found version "2.43.0") 
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE  
-- Could not find nvcc, please set CUDAToolkit_ROOT.
CMake Warning at CMakeLists.txt:356 (message):
  cuBLAS not found


-- CUDA host compiler is GNU 
CMake Error at CMakeLists.txt:532 (get_flags):
  get_flags Function invoked with incorrect arguments for function named:
  get_flags


-- CMAKE_SYSTEM_PROCESSOR: x86_64
-- x86 detected
-- Configuring incomplete, errors occurred!
llm/generate/generate_linux.go:3: running "bash": exit status 1
==> ERROR: A failure occurred in build().
    Aborting...
<!-- gh-comment-id:1907483858 --> @silverwind63 commented on GitHub (Jan 24, 2024): > I have the same Issue. Basically the same arch setup. What is your output for Discovered GPU libraries? > > `INFO Discovered GPU libraries: [/opt/cuda/targets/x86_64-linux/lib/stubs/libnvidia-ml.so /usr/lib/libnvidia-ml.so.545.29.06 /usr/lib32/libnvidia-ml.so.545.29.06 /usr/lib64/libnvidia-ml.so.545.29.06]` > > It appears that it tries to load the wrong file, even though the correct one is available. > > llama.cpp runs perfectly with GPU support. > > EDIT: > > With the help from Socialnetwooky from discord I could fix it. It seems that the ollama package is currently broken. The steps to fix it are as follows: > > Make a new tmp folder. Copy the PKGBUILD Copy the remaining files from: https://gitlab.archlinux.org/archlinux/packaging/packages/ollama-cuda execute makepkg -sri > > PKGBUILD: > > ``` > pkgname=ollama-cuda > pkgdesc='Create, run and share large language models (LLMs) with CUDA' > pkgver=0.1.20 > pkgrel=3 > arch=(x86_64) > url='https://github.com/jmorganca/ollama' > license=(MIT) > _ollamacommit=ab6be852c77064d7abeffb0b03c096aab90e95fe # tag: v0.1.20 > # The llama.cpp git submodule commit hash can be found here: > # https://github.com/jmorganca/ollama/tree/v0.1.20/llm > _llama_cpp_commit=328b83de23b33240e28f4e74900d1d06726f5eb1 > makedepends=(cmake cuda git go) > provides=(ollama) > conflicts=(ollama) > source=(git+$url#commit=$_ollamacommit > llama.cpp::git+https://github.com/ggerganov/llama.cpp#commit=$_llama_cpp_commit > sysusers.conf > tmpfiles.d > ollama.service) > b2sums=('SKIP' > 'SKIP' > 'SKIP' > 'SKIP' > 'SKIP') > # '65d39053cd1dd09562473c2e58f66a447ce0225b32607685f60350596b3288d6568c1cb897393b20236260e632427de1a952e72fe358407020f6cc7820fd4f60' > # '6f0b6886108e8d5f385bf7f9bebc60218797c53d4b88a69cc98564ad02c558cb86633b4f713ddd919d618146c04ac0a9215aa6c32ae192701af9d7850264dd56' > # 'be72a39e823d6631095ce407c92af6aee8650302eeaaa55a970a43592daaa141369d3a5a5eb6a992f6c2f5461370228ac7a84f3318422419fd868575454487d6') > > prepare() { > cd ${pkgname/-cuda} > > rm -frv llm/llama.cpp > > export GIN_MODE=release > > # Copy git submodule files instead of symlinking because the build process is sensitive to symlinks. > cp -r "$srcdir/llama.cpp" llm/llama.cpp > > # Turn LTO on and set the build type to Release > sed -i 's,T_CODE=on,T_CODE=on -D LLAMA_LTO=off -D CMAKE_BUILD_TYPE=Release,g' llm/generate/gen_linux.sh > > sed -i 's,var mode string = gin.DebugMode,var mode string = gin.ReleaseMode,g' server/routes.go > # Let gen_linux.sh find libcudart.so > sed -i 's,/usr/local/cuda/lib64,/opt/cuda/targets/x86_64-linux/lib,g' llm/generate/gen_linux.sh > > # Let gpu.go find libnvidia-ml.so from the cuda package > #sed -i 's,/opt/cuda/lib64/libnvidia-ml.so*,/opt/cuda/targets/x86_64-linux/lib/stubs/libnvidia-ml.so*,g' gpu/gpu.go > sed -i 's,/opt/cuda/lib64/libnvidia-ml.so*,/lib/libnvidia-ml.so*,g' gpu/gpu.go > } > > build() { > cd ${pkgname/-cuda} > export CGO_CFLAGS="$CFLAGS" CGO_CPPFLAGS="$CPPFLAGS" CGO_CXXFLAGS="$CXXFLAGS" CGO_LDFLAGS="$LDFLAGS" > go generate ./... > go build -buildmode=pie -ldflags=-fno-lto -trimpath -mod=readonly -modcacherw -ldflags=-linkmode=external \ > -ldflags=-buildid='' -ldflags="-X=github.com/jmorganca/ollama/version.Version=$pkgver" > } > > check() { > cd ${pkgname/-cuda} > go test ./api ./format > ./ollama --version > /dev/null > } > > package() { > install -Dm755 ${pkgname/-cuda}/${pkgname/-cuda} "$pkgdir/usr/bin/${pkgname/-cuda}" > install -dm700 "$pkgdir/var/lib/ollama" > install -Dm644 ollama.service "$pkgdir/usr/lib/systemd/system/ollama.service" > install -Dm644 sysusers.conf "$pkgdir/usr/lib/sysusers.d/ollama.conf" > install -Dm644 tmpfiles.d "$pkgdir/usr/lib/tmpfiles.d/ollama.conf" > install -Dm644 ${pkgname/-cuda}/LICENSE "$pkgdir/usr/share/licenses/$pkgname/LICENSE" > } > ``` When I run makepkg -sri to install it,it show me these errors: ``` -- The C compiler identification is GNU 13.2.1 -- The CXX compiler identification is GNU 13.2.1 -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Check for working C compiler: /usr/bin/cc - skipped -- Detecting C compile features -- Detecting C compile features - done -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Check for working CXX compiler: /usr/bin/c++ - skipped -- Detecting CXX compile features -- Detecting CXX compile features - done -- Found Git: /usr/bin/git (found version "2.43.0") -- Performing Test CMAKE_HAVE_LIBC_PTHREAD -- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success -- Found Threads: TRUE -- Could not find nvcc, please set CUDAToolkit_ROOT. CMake Warning at CMakeLists.txt:356 (message): cuBLAS not found -- CUDA host compiler is GNU CMake Error at CMakeLists.txt:532 (get_flags): get_flags Function invoked with incorrect arguments for function named: get_flags -- CMAKE_SYSTEM_PROCESSOR: x86_64 -- x86 detected -- Configuring incomplete, errors occurred! llm/generate/generate_linux.go:3: running "bash": exit status 1 ==> ERROR: A failure occurred in build(). Aborting... ```
GiteaMirror commented 2026-04-12 10:59:13 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Jan 26, 2024):

We've moved the stub library to the bottom of the list we try and this fix is in 0.1.22. I believe this should be resolved. Please re-open if you're still seeing the problem on 0.1.22.

<!-- gh-comment-id:1912696789 --> @dhiltgen commented on GitHub (Jan 26, 2024): We've moved the stub library to the [bottom of the list](https://github.com/ollama/ollama/blob/main/gpu/gpu.go#L50) we try and this fix is in 0.1.22. I believe this should be resolved. Please re-open if you're still seeing the problem on 0.1.22.
GiteaMirror commented 2026-04-12 10:59:13 -05:00
Author
Owner
Copy Link

@Rabcor commented on GitHub (Jan 27, 2024):

When I run makepkg -sri to install it,it show me these errors:

-- The C compiler identification is GNU 13.2.1
-- The CXX compiler identification is GNU 13.2.1
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found Git: /usr/bin/git (found version "2.43.0") 
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE  
-- Could not find nvcc, please set CUDAToolkit_ROOT.
CMake Warning at CMakeLists.txt:356 (message):
  cuBLAS not found


-- CUDA host compiler is GNU 
CMake Error at CMakeLists.txt:532 (get_flags):
  get_flags Function invoked with incorrect arguments for function named:
  get_flags


-- CMAKE_SYSTEM_PROCESSOR: x86_64
-- x86 detected
-- Configuring incomplete, errors occurred!
llm/generate/generate_linux.go:3: running "bash": exit status 1
==> ERROR: A failure occurred in build().
    Aborting...

Same here,, i tried adding nvidia root in the build function then it threw out a different error about not finding default cuda architectures, so i threw in a variable that told it where the nvcc compiler is, and the result from that was just even more errors :(

<!-- gh-comment-id:1913123791 --> @Rabcor commented on GitHub (Jan 27, 2024): > When I run makepkg -sri to install it,it show me these errors: > > ``` > -- The C compiler identification is GNU 13.2.1 > -- The CXX compiler identification is GNU 13.2.1 > -- Detecting C compiler ABI info > -- Detecting C compiler ABI info - done > -- Check for working C compiler: /usr/bin/cc - skipped > -- Detecting C compile features > -- Detecting C compile features - done > -- Detecting CXX compiler ABI info > -- Detecting CXX compiler ABI info - done > -- Check for working CXX compiler: /usr/bin/c++ - skipped > -- Detecting CXX compile features > -- Detecting CXX compile features - done > -- Found Git: /usr/bin/git (found version "2.43.0") > -- Performing Test CMAKE_HAVE_LIBC_PTHREAD > -- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success > -- Found Threads: TRUE > -- Could not find nvcc, please set CUDAToolkit_ROOT. > CMake Warning at CMakeLists.txt:356 (message): > cuBLAS not found > > > -- CUDA host compiler is GNU > CMake Error at CMakeLists.txt:532 (get_flags): > get_flags Function invoked with incorrect arguments for function named: > get_flags > > > -- CMAKE_SYSTEM_PROCESSOR: x86_64 > -- x86 detected > -- Configuring incomplete, errors occurred! > llm/generate/generate_linux.go:3: running "bash": exit status 1 > ==> ERROR: A failure occurred in build(). > Aborting... > ``` Same here,, i tried adding nvidia root in the build function then it threw out a different error about not finding default cuda architectures, so i threw in a variable that told it where the nvcc compiler is, and the result from that was just even more errors :(
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#1211
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?