gered/ollama
1
0
Fork 0
Code Issues Pull requests Actions Packages Projects Releases Wiki Activity

Merge pull request #3948 from dhiltgen/win_generate

Browse source 
Refactor windows generate for more modular usage
This commit is contained in:
Daniel Hiltgen 2024-04-26 09:17:20 -07:00 committed by GitHub
parent 2619850fb4 8671fdeda6
commit c0f818a07a
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 180 additions and 140 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

92
llm/generate/gen_windows.ps1
View file

@ -26,14 +26,20 @@ function amdGPUs {
$GPU_LIST -join ';'
}
$script:cmakeTargets = @("ollama_llama_server")
function init_vars {
if (!$script:SRC_DIR) {
$script:SRC_DIR = $(resolve-path "..\..\")
}
if (!$script:llamacppDir) {
$script:llamacppDir = "../llama.cpp"
}
$script:cmakeDefs = @(
"-DBUILD_SHARED_LIBS=on",
"-DLLAMA_NATIVE=off"
)
$script:cmakeTargets = @("ollama_llama_server")
$script:commonCpuDefs = @("-DCMAKE_POSITION_INDEPENDENT_CODE=on")
$script:ARCH = "amd64" # arm not yet supported.
$script:DIST_BASE = "${script:SRC_DIR}\dist\windows-${script:ARCH}\ollama_runners"
if ($env:CGO_CFLAGS -contains "-g") {
@ -166,28 +172,24 @@ function cleanup {
}
}
init_vars
git_module_setup
apply_patches
# -DLLAMA_AVX -- 2011 Intel Sandy Bridge & AMD Bulldozer
# -DLLAMA_AVX2 -- 2013 Intel Haswell & 2015 AMD Excavator / 2017 AMD Zen
# -DLLAMA_FMA (FMA3) -- 2013 Intel Haswell & 2012 AMD Piledriver
$script:commonCpuDefs = @("-DCMAKE_POSITION_INDEPENDENT_CODE=on")
if ($null -eq ${env:OLLAMA_SKIP_CPU_GENERATE}) {
# GCC build for direct linking into the Go binary
init_vars
# cmake will silently fallback to msvc compilers if mingw isn't in the path, so detect and fail fast
# as we need this to be compiled by gcc for golang to be able to link with itx
write-host "Checking for MinGW..."
# error action ensures we exit on failure
get-command gcc
get-command mingw32-make
$script:cmakeTargets = @("llama", "ggml")
$script:cmakeDefs = @(
function build_static() {
if ($null -eq ${env:OLLAMA_SKIP_CPU_GENERATE}) {
# GCC build for direct linking into the Go binary
init_vars
# cmake will silently fallback to msvc compilers if mingw isn't in the path, so detect and fail fast
# as we need this to be compiled by gcc for golang to be able to link with itx
write-host "Checking for MinGW..."
# error action ensures we exit on failure
get-command gcc
get-command mingw32-make
$script:cmakeTargets = @("llama", "ggml")
$script:cmakeDefs = @(
"-G", "MinGW Makefiles"
"-DCMAKE_C_COMPILER=gcc.exe",
"-DCMAKE_CXX_COMPILER=g++.exe",
@ -198,11 +200,17 @@ $script:cmakeDefs = @(
"-DLLAMA_AVX512=off",
"-DLLAMA_F16C=off",
"-DLLAMA_FMA=off")
$script:buildDir="../build/windows/${script:ARCH}_static"
write-host "Building static library"
build
$script:buildDir="../build/windows/${script:ARCH}_static"
write-host "Building static library"
build
} else {
write-host "Skipping CPU generation step as requested"
}
}
# remaining llama.cpp builds use MSVC
function build_cpu() {
if ($null -eq ${env:OLLAMA_SKIP_CPU_GENERATE}) {
# remaining llama.cpp builds use MSVC
init_vars
$script:cmakeDefs = $script:commonCpuDefs + @("-A", "x64", "-DLLAMA_AVX=off", "-DLLAMA_AVX2=off", "-DLLAMA_AVX512=off", "-DLLAMA_FMA=off", "-DLLAMA_F16C=off") + $script:cmakeDefs
$script:buildDir="../build/windows/${script:ARCH}/cpu"
@ -211,7 +219,13 @@ build
build
sign
install
} else {
write-host "Skipping CPU generation step as requested"
}
}
function build_cpu_avx() {
if ($null -eq ${env:OLLAMA_SKIP_CPU_GENERATE}) {
init_vars
$script:cmakeDefs = $script:commonCpuDefs + @("-A", "x64", "-DLLAMA_AVX=on", "-DLLAMA_AVX2=off", "-DLLAMA_AVX512=off", "-DLLAMA_FMA=off", "-DLLAMA_F16C=off") + $script:cmakeDefs
$script:buildDir="../build/windows/${script:ARCH}/cpu_avx"
@ -220,7 +234,13 @@ build
build
sign
install
} else {
write-host "Skipping CPU generation step as requested"
}
}
function build_cpu_avx2() {
if ($null -eq ${env:OLLAMA_SKIP_CPU_GENERATE}) {
init_vars
$script:cmakeDefs = $script:commonCpuDefs + @("-A", "x64", "-DLLAMA_AVX=on", "-DLLAMA_AVX2=on", "-DLLAMA_AVX512=off", "-DLLAMA_FMA=on", "-DLLAMA_F16C=on") + $script:cmakeDefs
$script:buildDir="../build/windows/${script:ARCH}/cpu_avx2"
@ -229,11 +249,13 @@ build
build
sign
install
} else {
} else {
write-host "Skipping CPU generation step as requested"
}
}
if ($null -ne $script:CUDA_LIB_DIR) {
function build_cuda() {
if ($null -ne $script:CUDA_LIB_DIR) {
# Then build cuda as a dynamically loaded library
$nvcc = "$script:CUDA_LIB_DIR\nvcc.exe"
$script:CUDA_VERSION=(get-item ($nvcc | split-path | split-path)).Basename
@ -257,9 +279,11 @@ if ($null -ne $script:CUDA_LIB_DIR) {
cp "${script:CUDA_LIB_DIR}\cudart64_*.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\"
cp "${script:CUDA_LIB_DIR}\cublas64_*.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\"
cp "${script:CUDA_LIB_DIR}\cublasLt64_*.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\"
}
}
if ($null -ne $env:HIP_PATH) {
function build_rocm() {
if ($null -ne $env:HIP_PATH) {
$script:ROCM_VERSION=(get-item $env:HIP_PATH).Basename
if ($null -ne $script:ROCM_VERSION) {
$script:ROCM_VARIANT="_v"+$script:ROCM_VERSION
@ -308,8 +332,24 @@ if ($null -ne $env:HIP_PATH) {
cp "${env:HIP_PATH}\bin\rocblas.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\"
# amdhip64.dll dependency comes from the driver and must be installed on the host to use AMD GPUs
cp "${env:HIP_PATH}\bin\rocblas\library\*" "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\rocblas\library\"
}
}
init_vars
if ($($args.count) -eq 0) {
git_module_setup
apply_patches
build_static
build_cpu_avx
build_cpu_avx2
build_cuda
build_rocm
cleanup
write-host "`ngo generate completed. LLM runners: $(get-childitem -path $script:DIST_BASE)"
cleanup
write-host "`ngo generate completed. LLM runners: $(get-childitem -path $script:DIST_BASE)"
} else {
for ( $i = 0; $i -lt $args.count; $i++ ) {
write-host "performing $($args[$i])"
& $($args[$i])
}
}