Mirror_Repos/ollama_ollama
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
de1ce45913871c5e3fa58c8da7640f911bf23c24
ollama_ollama/llama/server/CMakePresets.json
T
Daniel Hiltgen de1ce45913 cuda: add CC 10.0 for linux in CUDA v12 (#17025) 
Add compute capability 10.0 to the Linux CUDA v12 preset so B200-class devices can use the cuda_v12 backend with drivers that do not meet the CUDA v13 minimum.

Fixes #12583
2026-07-20 13:09:36 -07:00

322 lines
9.1 KiB
JSON
Raw Blame History

{
"version": 3,
"configurePresets": [
{
"name": "default",
"binaryDir": "${sourceDir}/../../build/llama-server",
"installDir": "${sourceDir}/../../dist",
"cacheVariables": {
"CMAKE_BUILD_TYPE": "Release",
"BUILD_SHARED_LIBS": "ON",
"GGML_BACKEND_DL": "ON",
"GGML_NATIVE": "OFF",
"GGML_OPENMP": "OFF"
}
},
{
"name": "cpu",
"inherits": ["default"],
"binaryDir": "${sourceDir}/../../build/llama-server-cpu",
"cacheVariables": {
"GGML_CPU_ALL_VARIANTS": "ON",
"OLLAMA_RUNNER_DIR": ""
}
},
{
"name": "cpu_windows",
"inherits": ["cpu"],
"cacheVariables": {
"GGML_OPENMP": "ON"
}
},
{
"name": "cpu_arm64",
"inherits": ["default"],
"binaryDir": "${sourceDir}/../../build/llama-server-cpu_arm64",
"toolchainFile": "${sourceDir}/../../cmake/windows-arm64-llvm-mingw.cmake",
"cacheVariables": {
"GGML_CPU_ALL_VARIANTS": "OFF",
"OLLAMA_RUNNER_DIR": "",
"OLLAMA_WINDOWS_RUNTIME_ARCH": "arm64"
}
},
{
"name": "darwin",
"inherits": ["default"],
"binaryDir": "${sourceDir}/../../build/llama-server-darwin",
"cacheVariables": {
"BUILD_SHARED_LIBS": "OFF",
"GGML_BACKEND_DL": "OFF",
"GGML_METAL": "ON",
"GGML_METAL_EMBED_LIBRARY": "ON",
"OLLAMA_RUNNER_DIR": ""
}
},
{
"name": "llama_cuda_v12_base",
"hidden": true,
"inherits": ["default"],
"cacheVariables": {
"GGML_CUDA": "ON",
"CMAKE_CUDA_FLAGS": "-Wno-deprecated-gpu-targets -t 2",
"OLLAMA_RUNNER_DIR": "cuda_v12",
"OLLAMA_GPU_BACKEND": "cuda"
}
},
{
"name": "llama_cuda_v12_linux",
"inherits": ["llama_cuda_v12_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-cuda_v12",
"cacheVariables": {
"CMAKE_CUDA_ARCHITECTURES": "50-virtual;52-virtual;60;61;70;75;80;86;89;90;90a;100;120"
}
},
{
"name": "llama_cuda_v12_windows",
"inherits": ["llama_cuda_v12_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-cuda_v12",
"cacheVariables": {
"CMAKE_CUDA_ARCHITECTURES": "50-virtual;52-virtual;60-virtual;61-virtual;70;75;80;86;89;90;90a;120"
}
},
{
"name": "llama_cuda_v12_user_arch",
"inherits": ["llama_cuda_v12_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-cuda_v12"
},
{
"name": "llama_cuda_v13_base",
"hidden": true,
"inherits": ["default"],
"cacheVariables": {
"GGML_CUDA": "ON",
"CMAKE_CUDA_FLAGS": "-t 4",
"OLLAMA_RUNNER_DIR": "cuda_v13",
"OLLAMA_GPU_BACKEND": "cuda"
}
},
{
"name": "llama_cuda_v13_linux",
"inherits": ["llama_cuda_v13_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-cuda_v13",
"cacheVariables": {
"CMAKE_CUDA_ARCHITECTURES": "75-virtual;80-virtual;86-virtual;87-virtual;89-virtual;90-virtual;90a-virtual;100-virtual;103-virtual;110-virtual;120-virtual;121-virtual"
}
},
{
"name": "llama_cuda_v13_windows",
"inherits": ["llama_cuda_v13_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-cuda_v13",
"cacheVariables": {
"CMAKE_CUDA_ARCHITECTURES": "75-virtual;80-virtual;86-virtual;89-virtual;100-virtual;120-virtual"
}
},
{
"name": "llama_cuda_v13_user_arch",
"inherits": ["llama_cuda_v13_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-cuda_v13"
},
{
"name": "llama_cuda_jetpack5_base",
"hidden": true,
"inherits": ["default"],
"cacheVariables": {
"GGML_CUDA": "ON",
"OLLAMA_RUNNER_DIR": "cuda_jetpack5",
"OLLAMA_GPU_BACKEND": "cuda"
}
},
{
"name": "llama_cuda_jetpack5",
"inherits": ["llama_cuda_jetpack5_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-cuda_jetpack5",
"cacheVariables": {
"CMAKE_CUDA_ARCHITECTURES": "72;87"
}
},
{
"name": "llama_cuda_jetpack5_user_arch",
"inherits": ["llama_cuda_jetpack5_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-cuda_jetpack5"
},
{
"name": "llama_cuda_jetpack6_base",
"hidden": true,
"inherits": ["default"],
"cacheVariables": {
"GGML_CUDA": "ON",
"OLLAMA_RUNNER_DIR": "cuda_jetpack6",
"OLLAMA_GPU_BACKEND": "cuda"
}
},
{
"name": "llama_cuda_jetpack6",
"inherits": ["llama_cuda_jetpack6_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-cuda_jetpack6",
"cacheVariables": {
"CMAKE_CUDA_ARCHITECTURES": "87"
}
},
{
"name": "llama_cuda_jetpack6_user_arch",
"inherits": ["llama_cuda_jetpack6_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-cuda_jetpack6"
},
{
"name": "rocm_v7_1_base",
"hidden": true,
"inherits": ["default"],
"cacheVariables": {
"GGML_HIP": "ON",
"CMAKE_HIP_PLATFORM": "amd",
"OLLAMA_RUNNER_DIR": "rocm_v7_1",
"OLLAMA_GPU_BACKEND": "hip"
}
},
{
"name": "rocm_v7_1_windows",
"inherits": ["rocm_v7_1_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-rocm_v7_1",
"cacheVariables": {
"AMDGPU_TARGETS": "gfx1030;gfx1100;gfx1101;gfx1102;gfx1150;gfx1151;gfx1200;gfx1201"
}
},
{
"name": "rocm_v7_1_user_arch",
"inherits": ["rocm_v7_1_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-rocm_v7_1"
},
{
"name": "rocm_v7_2_base",
"hidden": true,
"inherits": ["default"],
"cacheVariables": {
"GGML_HIP": "ON",
"CMAKE_HIP_PLATFORM": "amd",
"OLLAMA_RUNNER_DIR": "rocm_v7_2",
"OLLAMA_GPU_BACKEND": "hip"
}
},
{
"name": "rocm_v7_2_linux",
"inherits": ["rocm_v7_2_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-rocm_v7_2",
"cacheVariables": {
"AMDGPU_TARGETS": "gfx908:xnack-;gfx90a:xnack+;gfx90a:xnack-;gfx942;gfx950;gfx1030;gfx1100;gfx1101;gfx1102;gfx1150;gfx1151;gfx1200;gfx1201"
}
},
{
"name": "rocm_v7_2_user_arch",
"inherits": ["rocm_v7_2_base"],
"binaryDir": "${sourceDir}/../../build/llama-server-rocm_v7_2"
},
{
"name": "vulkan",
"inherits": ["default"],
"binaryDir": "${sourceDir}/../../build/llama-server-vulkan",
"cacheVariables": {
"GGML_VULKAN": "ON",
"OLLAMA_RUNNER_DIR": "vulkan",
"OLLAMA_GPU_BACKEND": "vulkan"
}
}
],
"buildPresets": [
{
"name": "cpu",
"configurePreset": "cpu",
"targets": ["llama-server", "llama-quantize"]
},
{
"name": "cpu_windows",
"configurePreset": "cpu_windows",
"targets": ["llama-server", "llama-quantize"]
},
{
"name": "cpu_arm64",
"configurePreset": "cpu_arm64",
"targets": ["llama-server", "llama-quantize"]
},
{
"name": "darwin",
"configurePreset": "darwin",
"targets": ["llama-server", "llama-quantize"]
},
{
"name": "llama_cuda_v12_linux",
"configurePreset": "llama_cuda_v12_linux",
"targets": ["ggml-cuda"]
},
{
"name": "llama_cuda_v12_windows",
"configurePreset": "llama_cuda_v12_windows",
"targets": ["ggml-cuda"]
},
{
"name": "llama_cuda_v12_user_arch",
"configurePreset": "llama_cuda_v12_user_arch",
"targets": ["ggml-cuda"]
},
{
"name": "llama_cuda_v13_linux",
"configurePreset": "llama_cuda_v13_linux",
"targets": ["ggml-cuda"]
},
{
"name": "llama_cuda_v13_windows",
"configurePreset": "llama_cuda_v13_windows",
"targets": ["ggml-cuda"]
},
{
"name": "llama_cuda_v13_user_arch",
"configurePreset": "llama_cuda_v13_user_arch",
"targets": ["ggml-cuda"]
},
{
"name": "llama_cuda_jetpack5",
"configurePreset": "llama_cuda_jetpack5",
"targets": ["ggml-cuda"]
},
{
"name": "llama_cuda_jetpack5_user_arch",
"configurePreset": "llama_cuda_jetpack5_user_arch",
"targets": ["ggml-cuda"]
},
{
"name": "llama_cuda_jetpack6",
"configurePreset": "llama_cuda_jetpack6",
"targets": ["ggml-cuda"]
},
{
"name": "llama_cuda_jetpack6_user_arch",
"configurePreset": "llama_cuda_jetpack6_user_arch",
"targets": ["ggml-cuda"]
},
{
"name": "rocm_v7_1_windows",
"configurePreset": "rocm_v7_1_windows",
"targets": ["ggml-hip"]
},
{
"name": "rocm_v7_1_user_arch",
"configurePreset": "rocm_v7_1_user_arch",
"targets": ["ggml-hip"]
},
{
"name": "rocm_v7_2_linux",
"configurePreset": "rocm_v7_2_linux",
"targets": ["ggml-hip"]
},
{
"name": "rocm_v7_2_user_arch",
"configurePreset": "rocm_v7_2_user_arch",
"targets": ["ggml-hip"]
},
{
"name": "vulkan",
"configurePreset": "vulkan",
"targets": ["ggml-vulkan"]
}
]
}
Reference in New Issue View Git Blame Copy Permalink