Can someone please write a definitive guide on how to avoid "Ceres was compiled without CUDA/cuDSS support" error on Windows 11

[okj-okj]

Hello,

I very much would like to be able to run the bundle adjustment on my GPU. I followed several guides, none of them worked out for me. I'm on Windows 11 and I have a 30 series RTX Nvidia GPU. Please if you know or happen to know someone who knows how to do this, don't hesitate and reply to this issue.

Thanks in advance,

okj

[okj-okj]

OK, I figured it out, here's my guide. I switched to Ubuntu 25.04 as I got frustrated with Win11. Happy Glomapping on the GPU!

sudo apt update && sudo apt upgrade -y
sudo apt install -y build-essential curl zip unzip tar git pkg-config

================================================================================================

sudo apt-get install cmake
sudo apt-get install libgoogle-glog-dev libgflags-dev
sudo apt-get install libatlas-base-dev
sudo apt-get install libeigen3-dev
sudo apt-get install libsuitesparse-dev

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt-get update
sudo apt-get install cuda-toolkit

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/x86_64/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt-get update
sudo apt-get -y install cudss

sudo apt-get install libabsl-dev

git clone https://github.com/ceres-solver/ceres-solver.git
cd ceres-solver
mkdir build
cd build
cmake .. -DCMAKE_CUDA_ARCHITECTURES="native" -DCMAKE_BUILD_TYPE=Release -DBUILD_TESTING=OFF -DBUILD_EXAMPLES=OFF -DBUILD_BENCHMARKS=OFF
make -j4

#in ceres-solver/cmakelists.txt after cmake before project
set(CMAKE_CUDA_COMPILER /usr/local/cuda/bin/nvcc)

sudo mkdir -p /usr/include/absl/log
sudo ln -s /usr/include/glog/vlog_is_on.h /usr/include/absl/log/vlog_is_on.h

#include "glog/logging.h"

find internal/ceres/ -name ".cc" -exec grep -l "VLOG" {} ; |
xargs grep -L "glog/logging.h" |
while read file; do
sed -i '/^#include.$/a #include "glog/logging.h"' "$file"
done

make -j4
sudo make install

================================================================================================

sudo apt-get install
git
cmake
ninja-build
build-essential
libboost-program-options-dev
libboost-graph-dev
libboost-system-dev
libeigen3-dev
libfreeimage-dev
libmetis-dev
libgoogle-glog-dev
libgtest-dev
libgmock-dev
libsqlite3-dev
libglew-dev
qtbase5-dev
libqt5opengl5-dev
libcgal-dev
libceres-dev
libcurl4-openssl-dev
libmkl-full-dev

git clone https://github.com/colmap/colmap.git
cd colmap
mkdir build
cd build
cmake .. -GNinja -DBLA_VENDOR=Intel10_64lp -DCMAKE_CUDA_ARCHITECTURES="86" -DCMAKE_BUILD_TYPE=Release

#put this in cmakelists before project after cmake
set(CMAKE_CUDA_COMPILER /usr/local/cuda/bin/nvcc)

sudo ninja
sudo ninja install

================================================================================================

sudo apt install libflann-dev

#remove -werror in glomap/glomap/cmakelists.txt
-werror in glomap/build/_deps/poselib-src/cmakelists.txt

Clone GLOMAP

git clone https://github.com/colmap/glomap.git
cd glomap
mkdir build
cd build
sudo cmake .. -GNinja -DCMAKE_CUDA_ARCHITECTURES=native

#put this in cmakelists before project after cmake
set(CMAKE_CUDA_COMPILER /usr/local/cuda/bin/nvcc)

sudo ninja && sudo ninja install

[okj-okj]

UPDATED GUIDE

sudo apt update && sudo apt upgrade -y
sudo apt install -y build-essential curl zip unzip tar git pkg-config

================================================================================================

sudo apt-get install cmake
sudo apt-get install libgoogle-glog-dev libgflags-dev
sudo apt-get install libatlas-base-dev
sudo apt-get install libeigen3-dev
sudo apt-get install libsuitesparse-dev

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt-get update
sudo apt-get install cuda-toolkit

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/x86_64/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt-get update
sudo apt-get -y install cudss

sudo apt-get install libabsl-dev

git clone https://github.com/ceres-solver/ceres-solver.git
cd ceres-solver
mkdir build
cd build
cmake .. -DCMAKE_CUDA_ARCHITECTURES="native" -DCMAKE_BUILD_TYPE=Release -DBUILD_TESTING=OFF -DBUILD_EXAMPLES=OFF -DBUILD_BENCHMARKS=OFF
make -j4

#in ceres-solver/cmakelists.txt after cmake before project
set(CMAKE_CUDA_COMPILER /usr/local/cuda/bin/nvcc)

cd ..

find internal/ceres/ -name ".cc" -exec grep -l "VLOG" {} +
| xargs grep -L 'glog/logging.h'
| while read -r file; do
sed -i '/^#include[[:space:]][<"].*[>"]/a #include "glog/logging.h"' "$file"
done

sudo mkdir -p /usr/include/absl/log
sudo ln -s /usr/include/glog/vlog_is_on.h /usr/include/absl/log/vlog_is_on.h

make -j4
sudo make install

================================================================================================

sudo apt-get install git cmake ninja-build build-essential libboost-program-options-dev libboost-graph-dev libboost-system-dev libeigen3-dev libfreeimage-dev libmetis-dev libgoogle-glog-dev libgtest-dev libgmock-dev libsqlite3-dev libglew-dev qtbase5-dev libqt5opengl5-dev libcgal-dev libceres-dev libcurl4-openssl-dev libmkl-full-dev

git clone https://github.com/colmap/colmap.git
cd colmap
mkdir build
cd build
cmake .. -GNinja -DBLA_VENDOR=Intel10_64lp -DCMAKE_CUDA_ARCHITECTURES="86" -DCMAKE_BUILD_TYPE=Release -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda-12.6 -DCMAKE_CUDA_COMPILER=/usr/local/cuda-12.6/bin/nvcc

#put this in cmakelists before project after cmake
set(CMAKE_CUDA_COMPILER /usr/local/cuda/bin/nvcc)

sudo apt install libcublas12

sudo ninja
sudo ninja install

================================================================================================

sudo apt install libflann-dev

#remove -werror in glomap/glomap/cmakelists.txt
-werror in glomap/build/_deps/poselib-src/cmakelists.txt

git clone https://github.com/colmap/glomap.git
cd glomap
mkdir build
cd build
sudo cmake .. -GNinja -DCMAKE_CUDA_ARCHITECTURES=native

#put this in cmakelists before project after cmake
set(CMAKE_CUDA_COMPILER /usr/local/cuda/bin/nvcc)

sudo ninja
sudo ninja install

[bossowl33]

im not great at linux, could you add a bit of additional info? mainly how do i run all of these commands, do i copy and paste the entire command list to a single .sh file? Everytime i have run it, even on a clean ubuntu 25.04 install it fails for one or two different reasons, i don't expect you to debug it for me i just want to know if im doing it right in the first place.

[Kation]

Why windows cuda release not work?

[rlesac]

Same issue here, downloaded the 1.2.0 from Oct 31.

I20251226 15:54:05.091052 77596 global_positioning.cc:52] Setting up the global positioner problem
W20251226 15:54:08.770196 77596 global_positioning.cc:512] Requested to use GPU for bundle adjustment, but Ceres was compiled without CUDA support. Falling back to CPU-based dense solvers.
W20251226 15:54:08.770573 77596 global_positioning.cc:529] Requested to use GPU for bundle adjustment, but Ceres was compiled without cuDSS support. Falling back to CPU-based sparse solvers.

Obviously I downloaded glomap-x64-windows-cuda.zip