Skip to content

GitLab

"...git@developer.sourcefind.cn:modelzoo/qwen_lmdeploy.git" did not exist on "f4422fabfa3cb661b205d20bb6bb8875dc488b7c"
Unverified Commit c0f818a0 authored by Daniel Hiltgen's avatar Daniel Hiltgen Committed by GitHub
Browse files

Merge pull request #3948 from dhiltgen/win_generate 

Refactor windows generate for more modular usage
parents 2619850f 8671fded
Hide whitespace changes
Inline Side-by-side
Showing with 169 additions and 129 deletions
llm/generate/gen_windows.ps1
View file @ c0f818a0
...@@ -26,14 +26,20 @@ function amdGPUs { ...@@ -26,14 +26,20 @@ function amdGPUs {
$GPU_LIST -join ';' $GPU_LIST -join ';'
} }
$script:cmakeTargets = @("ollama_llama_server")
function init_vars { function init_vars {
$script:SRC_DIR = $(resolve-path "..\..\") if (!$script:SRC_DIR) {
$script:llamacppDir = "../llama.cpp" $script:SRC_DIR = $(resolve-path "..\..\")
}
if (!$script:llamacppDir) {
$script:llamacppDir = "../llama.cpp"
}
$script:cmakeDefs = @( $script:cmakeDefs = @(
"-DBUILD_SHARED_LIBS=on", "-DBUILD_SHARED_LIBS=on",
"-DLLAMA_NATIVE=off" "-DLLAMA_NATIVE=off"
) )
$script:cmakeTargets = @("ollama_llama_server") $script:commonCpuDefs = @("-DCMAKE_POSITION_INDEPENDENT_CODE=on")
$script:ARCH = "amd64" # arm not yet supported. $script:ARCH = "amd64" # arm not yet supported.
$script:DIST_BASE = "${script:SRC_DIR}\dist\windows-${script:ARCH}\ollama_runners" $script:DIST_BASE = "${script:SRC_DIR}\dist\windows-${script:ARCH}\ollama_runners"
if ($env:CGO_CFLAGS -contains "-g") { if ($env:CGO_CFLAGS -contains "-g") {
...@@ -166,150 +172,184 @@ function cleanup { ...@@ -166,150 +172,184 @@ function cleanup {
} }
} }
init_vars
git_module_setup
apply_patches
# -DLLAMA_AVX -- 2011 Intel Sandy Bridge & AMD Bulldozer # -DLLAMA_AVX -- 2011 Intel Sandy Bridge & AMD Bulldozer
# -DLLAMA_AVX2 -- 2013 Intel Haswell & 2015 AMD Excavator / 2017 AMD Zen # -DLLAMA_AVX2 -- 2013 Intel Haswell & 2015 AMD Excavator / 2017 AMD Zen
# -DLLAMA_FMA (FMA3) -- 2013 Intel Haswell & 2012 AMD Piledriver # -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 = @(
"-G", "MinGW Makefiles"
"-DCMAKE_C_COMPILER=gcc.exe",
"-DCMAKE_CXX_COMPILER=g++.exe",
"-DBUILD_SHARED_LIBS=off",
"-DLLAMA_NATIVE=off",
"-DLLAMA_AVX=off",
"-DLLAMA_AVX2=off",
"-DLLAMA_AVX512=off",
"-DLLAMA_F16C=off",
"-DLLAMA_FMA=off")
$script:buildDir="../build/windows/${script:ARCH}_static"
write-host "Building static library"
build
# 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"
$script:distDir="$script:DIST_BASE\cpu"
write-host "Building LCD CPU"
build
sign
install
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"
$script:distDir="$script:DIST_BASE\cpu_avx"
write-host "Building AVX CPU"
build
sign
install
init_vars function build_static() {
$script:cmakeDefs = $script:commonCpuDefs + @("-A", "x64", "-DLLAMA_AVX=on", "-DLLAMA_AVX2=on", "-DLLAMA_AVX512=off", "-DLLAMA_FMA=on", "-DLLAMA_F16C=on") + $script:cmakeDefs if ($null -eq ${env:OLLAMA_SKIP_CPU_GENERATE}) {
$script:buildDir="../build/windows/${script:ARCH}/cpu_avx2" # GCC build for direct linking into the Go binary
$script:distDir="$script:DIST_BASE\cpu_avx2" init_vars
write-host "Building AVX2 CPU" # cmake will silently fallback to msvc compilers if mingw isn't in the path, so detect and fail fast
build # as we need this to be compiled by gcc for golang to be able to link with itx
sign write-host "Checking for MinGW..."
install # error action ensures we exit on failure
} else { get-command gcc
write-host "Skipping CPU generation step as requested" get-command mingw32-make
$script:cmakeTargets = @("llama", "ggml")
$script:cmakeDefs = @(
"-G", "MinGW Makefiles"
"-DCMAKE_C_COMPILER=gcc.exe",
"-DCMAKE_CXX_COMPILER=g++.exe",
"-DBUILD_SHARED_LIBS=off",
"-DLLAMA_NATIVE=off",
"-DLLAMA_AVX=off",
"-DLLAMA_AVX2=off",
"-DLLAMA_AVX512=off",
"-DLLAMA_F16C=off",
"-DLLAMA_FMA=off")
$script:buildDir="../build/windows/${script:ARCH}_static"
write-host "Building static library"
build
} else {
write-host "Skipping CPU generation step as requested"
}
} }
if ($null -ne $script:CUDA_LIB_DIR) { function build_cpu() {
# Then build cuda as a dynamically loaded library if ($null -eq ${env:OLLAMA_SKIP_CPU_GENERATE}) {
$nvcc = "$script:CUDA_LIB_DIR\nvcc.exe" # remaining llama.cpp builds use MSVC
$script:CUDA_VERSION=(get-item ($nvcc | split-path | split-path)).Basename init_vars
if ($null -ne $script:CUDA_VERSION) { $script:cmakeDefs = $script:commonCpuDefs + @("-A", "x64", "-DLLAMA_AVX=off", "-DLLAMA_AVX2=off", "-DLLAMA_AVX512=off", "-DLLAMA_FMA=off", "-DLLAMA_F16C=off") + $script:cmakeDefs
$script:CUDA_VARIANT="_"+$script:CUDA_VERSION $script:buildDir="../build/windows/${script:ARCH}/cpu"
$script:distDir="$script:DIST_BASE\cpu"
write-host "Building LCD CPU"
build
sign
install
} else {
write-host "Skipping CPU generation step as requested"
} }
init_vars }
$script:buildDir="../build/windows/${script:ARCH}/cuda$script:CUDA_VARIANT"
$script:distDir="$script:DIST_BASE\cuda$script:CUDA_VARIANT" function build_cpu_avx() {
$script:cmakeDefs += @("-A", "x64", "-DLLAMA_CUDA=ON", "-DLLAMA_AVX=on", "-DLLAMA_AVX2=off", "-DCUDAToolkit_INCLUDE_DIR=$script:CUDA_INCLUDE_DIR", "-DCMAKE_CUDA_ARCHITECTURES=${script:CMAKE_CUDA_ARCHITECTURES}") if ($null -eq ${env:OLLAMA_SKIP_CPU_GENERATE}) {
if ($null -ne $env:OLLAMA_CUSTOM_CUDA_DEFS) { init_vars
write-host "OLLAMA_CUSTOM_CUDA_DEFS=`"${env:OLLAMA_CUSTOM_CUDA_DEFS}`"" $script:cmakeDefs = $script:commonCpuDefs + @("-A", "x64", "-DLLAMA_AVX=on", "-DLLAMA_AVX2=off", "-DLLAMA_AVX512=off", "-DLLAMA_FMA=off", "-DLLAMA_F16C=off") + $script:cmakeDefs
$script:cmakeDefs +=@("${env:OLLAMA_CUSTOM_CUDA_DEFS}") $script:buildDir="../build/windows/${script:ARCH}/cpu_avx"
write-host "building custom CUDA GPU" $script:distDir="$script:DIST_BASE\cpu_avx"
write-host "Building AVX CPU"
build
sign
install
} else {
write-host "Skipping CPU generation step as requested"
} }
build }
sign
install
write-host "copying CUDA dependencies to ${script:SRC_DIR}\dist\windows-${script:ARCH}\" function build_cpu_avx2() {
cp "${script:CUDA_LIB_DIR}\cudart64_*.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\" if ($null -eq ${env:OLLAMA_SKIP_CPU_GENERATE}) {
cp "${script:CUDA_LIB_DIR}\cublas64_*.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\" init_vars
cp "${script:CUDA_LIB_DIR}\cublasLt64_*.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\" $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"
$script:distDir="$script:DIST_BASE\cpu_avx2"
write-host "Building AVX2 CPU"
build
sign
install
} else {
write-host "Skipping CPU generation step as requested"
}
} }
if ($null -ne $env:HIP_PATH) { function build_cuda() {
$script:ROCM_VERSION=(get-item $env:HIP_PATH).Basename if ($null -ne $script:CUDA_LIB_DIR) {
if ($null -ne $script:ROCM_VERSION) { # Then build cuda as a dynamically loaded library
$script:ROCM_VARIANT="_v"+$script:ROCM_VERSION $nvcc = "$script:CUDA_LIB_DIR\nvcc.exe"
$script:CUDA_VERSION=(get-item ($nvcc | split-path | split-path)).Basename
if ($null -ne $script:CUDA_VERSION) {
$script:CUDA_VARIANT="_"+$script:CUDA_VERSION
}
init_vars
$script:buildDir="../build/windows/${script:ARCH}/cuda$script:CUDA_VARIANT"
$script:distDir="$script:DIST_BASE\cuda$script:CUDA_VARIANT"
$script:cmakeDefs += @("-A", "x64", "-DLLAMA_CUDA=ON", "-DLLAMA_AVX=on", "-DLLAMA_AVX2=off", "-DCUDAToolkit_INCLUDE_DIR=$script:CUDA_INCLUDE_DIR", "-DCMAKE_CUDA_ARCHITECTURES=${script:CMAKE_CUDA_ARCHITECTURES}")
if ($null -ne $env:OLLAMA_CUSTOM_CUDA_DEFS) {
write-host "OLLAMA_CUSTOM_CUDA_DEFS=`"${env:OLLAMA_CUSTOM_CUDA_DEFS}`""
$script:cmakeDefs +=@("${env:OLLAMA_CUSTOM_CUDA_DEFS}")
write-host "building custom CUDA GPU"
}
build
sign
install
write-host "copying CUDA dependencies to ${script:SRC_DIR}\dist\windows-${script:ARCH}\"
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}\"
} }
}
init_vars function build_rocm() {
$script:buildDir="../build/windows/${script:ARCH}/rocm$script:ROCM_VARIANT" if ($null -ne $env:HIP_PATH) {
$script:distDir="$script:DIST_BASE\rocm$script:ROCM_VARIANT" $script:ROCM_VERSION=(get-item $env:HIP_PATH).Basename
$script:cmakeDefs += @( if ($null -ne $script:ROCM_VERSION) {
"-G", "Ninja", $script:ROCM_VARIANT="_v"+$script:ROCM_VERSION
"-DCMAKE_C_COMPILER=clang.exe", }
"-DCMAKE_CXX_COMPILER=clang++.exe",
"-DLLAMA_HIPBLAS=on",
"-DHIP_PLATFORM=amd",
"-DLLAMA_AVX=on",
"-DLLAMA_AVX2=off",
"-DCMAKE_POSITION_INDEPENDENT_CODE=on",
"-DAMDGPU_TARGETS=$(amdGPUs)",
"-DGPU_TARGETS=$(amdGPUs)"
)
# Make sure the ROCm binary dir is first in the path init_vars
$env:PATH="$env:HIP_PATH\bin;$env:PATH" $script:buildDir="../build/windows/${script:ARCH}/rocm$script:ROCM_VARIANT"
$script:distDir="$script:DIST_BASE\rocm$script:ROCM_VARIANT"
$script:cmakeDefs += @(
"-G", "Ninja",
"-DCMAKE_C_COMPILER=clang.exe",
"-DCMAKE_CXX_COMPILER=clang++.exe",
"-DLLAMA_HIPBLAS=on",
"-DHIP_PLATFORM=amd",
"-DLLAMA_AVX=on",
"-DLLAMA_AVX2=off",
"-DCMAKE_POSITION_INDEPENDENT_CODE=on",
"-DAMDGPU_TARGETS=$(amdGPUs)",
"-DGPU_TARGETS=$(amdGPUs)"
)
# We have to clobber the LIB var from the developer shell for clang to work properly # Make sure the ROCm binary dir is first in the path
$env:LIB="" $env:PATH="$env:HIP_PATH\bin;$env:PATH"
if ($null -ne $env:OLLAMA_CUSTOM_ROCM_DEFS) {
write-host "OLLAMA_CUSTOM_ROCM_DEFS=`"${env:OLLAMA_CUSTOM_ROCM_DEFS}`"" # We have to clobber the LIB var from the developer shell for clang to work properly
$script:cmakeDefs += @("${env:OLLAMA_CUSTOM_ROCM_DEFS}") $env:LIB=""
write-host "building custom ROCM GPU" if ($null -ne $env:OLLAMA_CUSTOM_ROCM_DEFS) {
} write-host "OLLAMA_CUSTOM_ROCM_DEFS=`"${env:OLLAMA_CUSTOM_ROCM_DEFS}`""
write-host "Building ROCm" $script:cmakeDefs += @("${env:OLLAMA_CUSTOM_ROCM_DEFS}")
build write-host "building custom ROCM GPU"
# Ninja doesn't prefix with config name }
${script:config}="" write-host "Building ROCm"
if ($null -ne $script:DUMPBIN) { build
& "$script:DUMPBIN" /dependents "${script:buildDir}/bin/ollama_llama_server.exe" | select-string ".dll" # Ninja doesn't prefix with config name
} ${script:config}=""
sign if ($null -ne $script:DUMPBIN) {
install & "$script:DUMPBIN" /dependents "${script:buildDir}/bin/ollama_llama_server.exe" | select-string ".dll"
}
sign
install
# Assumes v5.7, may need adjustments for v6 # Assumes v5.7, may need adjustments for v6
rm -ea 0 -recurse -force -path "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\" rm -ea 0 -recurse -force -path "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\"
md "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\rocblas\library\" -ea 0 > $null md "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\rocblas\library\" -ea 0 > $null
cp "${env:HIP_PATH}\bin\hipblas.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\" cp "${env:HIP_PATH}\bin\hipblas.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\"
cp "${env:HIP_PATH}\bin\rocblas.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\" cp "${env:HIP_PATH}\bin\rocblas.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\"
# amdhip64.dll dependency comes from the driver and must be installed on the host to use AMD GPUs # amdhip64.dll dependency comes from the driver and must be installed on the host to use AMD GPUs
cp "${env:HIP_PATH}\bin\rocblas\library\*" "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\rocblas\library\" 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 cleanup
write-host "`ngo generate completed. LLM runners: $(get-childitem -path $script:DIST_BASE)" 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])
}
}
\ No newline at end of file
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment