*Your operating system distribution may already have packages for NVIDIA CUDA. Distro packages are often preferable, but instructions are distro-specific. Please consult distro-specific docs for dependencies if available!*
_Your operating system distribution may already have packages for NVIDIA CUDA. Distro packages are often preferable, but instructions are distro-specific. Please consult distro-specific docs for dependencies if available!_
Install `cmake` and `golang` as well as [NVIDIA CUDA](https://developer.nvidia.com/cuda-downloads)
Install `cmake` and `golang` as well as [NVIDIA CUDA](https://developer.nvidia.com/cuda-downloads)
development and runtime packages.
development and runtime packages.
Typically the build scripts will auto-detect CUDA, however, if your Linux distro
Typically the build scripts will auto-detect CUDA, however, if your Linux distro
or installation approach uses unusual paths, you can specify the location by
or installation approach uses unusual paths, you can specify the location by
specifying an environment variable `CUDA_LIB_DIR` to the location of the shared
specifying an environment variable `CUDA_LIB_DIR` to the location of the shared
libraries, and `CUDACXX` to the location of the nvcc compiler. You can customize
libraries, and `CUDACXX` to the location of the nvcc compiler. You can customize
set set of target CUDA architectues by setting `CMAKE_CUDA_ARCHITECTURES` (e.g. "50;60;70")
set set of target CUDA architectues by setting `CMAKE_CUDA_ARCHITECTURES` (e.g. "50;60;70")
Then generate dependencies:
Then generate dependencies:
...
@@ -67,7 +67,7 @@ go build .
...
@@ -67,7 +67,7 @@ go build .
#### Linux ROCm (AMD)
#### Linux ROCm (AMD)
*Your operating system distribution may already have packages for AMD ROCm and CLBlast. Distro packages are often preferable, but instructions are distro-specific. Please consult distro-specific docs for dependencies if available!*
_Your operating system distribution may already have packages for AMD ROCm and CLBlast. Distro packages are often preferable, but instructions are distro-specific. Please consult distro-specific docs for dependencies if available!_
Install [CLBlast](https://github.com/CNugteren/CLBlast/blob/master/doc/installation.md) and [ROCm](https://rocm.docs.amd.com/en/latest/deploy/linux/quick_start.html) development packages first, as well as `cmake` and `golang`.
Install [CLBlast](https://github.com/CNugteren/CLBlast/blob/master/doc/installation.md) and [ROCm](https://rocm.docs.amd.com/en/latest/deploy/linux/quick_start.html) development packages first, as well as `cmake` and `golang`.
...
@@ -75,7 +75,7 @@ Typically the build scripts will auto-detect ROCm, however, if your Linux distro
...
@@ -75,7 +75,7 @@ Typically the build scripts will auto-detect ROCm, however, if your Linux distro
or installation approach uses unusual paths, you can specify the location by
or installation approach uses unusual paths, you can specify the location by
specifying an environment variable `ROCM_PATH` to the location of the ROCm
specifying an environment variable `ROCM_PATH` to the location of the ROCm
install (typically `/opt/rocm`), and `CLBlast_DIR` to the location of the
install (typically `/opt/rocm`), and `CLBlast_DIR` to the location of the
CLBlast install (typically `/usr/lib/cmake/CLBlast`). You can also customize
CLBlast install (typically `/usr/lib/cmake/CLBlast`). You can also customize
the AMD GPU targets by setting AMDGPU_TARGETS (e.g. `AMDGPU_TARGETS="gfx1101;gfx1102"`)
the AMD GPU targets by setting AMDGPU_TARGETS (e.g. `AMDGPU_TARGETS="gfx1101;gfx1102"`)
```
```
...
@@ -88,17 +88,17 @@ Then build the binary:
...
@@ -88,17 +88,17 @@ Then build the binary:
go build .
go build .
```
```
ROCm requires elevated privileges to access the GPU at runtime. On most distros you can add your user account to the `render` group, or run as root.
ROCm requires elevated privileges to access the GPU at runtime. On most distros you can add your user account to the `render` group, or run as root.
#### Advanced CPU Settings
#### Advanced CPU Settings
By default, running `go generate ./...` will compile a few different variations
By default, running `go generate ./...` will compile a few different variations
of the LLM library based on common CPU families and vector math capabilities,
of the LLM library based on common CPU families and vector math capabilities,
including a lowest-common-denominator which should run on almost any 64 bit CPU
including a lowest-common-denominator which should run on almost any 64 bit CPU
somewhat slowly. At runtime, Ollama will auto-detect the optimal variation to
somewhat slowly. At runtime, Ollama will auto-detect the optimal variation to
load. If you would like to build a CPU-based build customized for your
load. If you would like to build a CPU-based build customized for your
processor, you can set `OLLAMA_CUSTOM_CPU_DEFS` to the llama.cpp flags you would
processor, you can set `OLLAMA_CUSTOM_CPU_DEFS` to the llama.cpp flags you would
like to use. For example, to compile an optimized binary for an Intel i9-9880H,
like to use. For example, to compile an optimized binary for an Intel i9-9880H,
you might use:
you might use:
```
```
...
@@ -108,8 +108,7 @@ go build .
...
@@ -108,8 +108,7 @@ go build .
#### Containerized Linux Build
#### Containerized Linux Build
If you have Docker available, you can build linux binaries with `./scripts/build_linux.sh` which has the CUDA and ROCm dependencies included. The resulting binary is placed in `./dist`
If you have Docker available, you can build linux binaries with `./scripts/build_linux.sh` which has the CUDA and ROCm dependencies included. The resulting binary is placed in `./dist`
### Windows
### Windows
...
@@ -118,7 +117,7 @@ Note: The windows build for Ollama is still under development.
...
@@ -118,7 +117,7 @@ Note: The windows build for Ollama is still under development.
Install required tools:
Install required tools:
- MSVC toolchain - C/C++ and cmake as minimal requirements
- MSVC toolchain - C/C++ and cmake as minimal requirements
