EngineX-Mthreads/enginex-mthreads-llama.cpp
4
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

sync from b7516

Browse Source
This commit is contained in:
xiezhongtao
2026-01-16 11:16:14 +08:00
parent f4ae4cc7da
commit 6ee41dd9e3
380 changed files with 18435 additions and 38806 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
scripts/debug-test.sh
View File

@@ -109,7 +109,8 @@ rm -rf "$build_dir" && mkdir "$build_dir" || abort "Failed to make $build_dir"
# Step 2: Setup Build Environment and Compile Test Binaries
###########################################################
cmake -B "./$build_dir" -DCMAKE_BUILD_TYPE=Debug -DGGML_CUDA=1 || abort "Failed to build environment"
# Note: test-eval-callback requires -DLLAMA_CURL
cmake -B "./$build_dir" -DCMAKE_BUILD_TYPE=Debug -DGGML_CUDA=1 -DLLAMA_CURL=1 || abort "Failed to build environment"
pushd "$build_dir"
make -j || abort "Failed to compile"
popd > /dev/null || exit 1

79
scripts/pr2wt.sh
View File

@@ -1,79 +0,0 @@
#!/usr/bin/env bash
# intialize a new worktree from a PR number:
#
# - creates a new remote using the fork's clone URL
# - creates a local branch tracking the remote branch
# - creates a new worktree in a parent folder, suffixed with "-pr-$PR"
#
# sample usage:
# ./scripts/pr2wt.sh 12345
# ./scripts/pr2wt.sh 12345 opencode
# ./scripts/pr2wt.sh 12345 "cmake -B build && cmake --build build"
# ./scripts/pr2wt.sh 12345 "bash -l"
function usage() {
echo "usage: $0 <pr_number> [cmd]"
exit 1
}
# check we are in the right directory
if [[ ! -f "scripts/pr2wt.sh" ]]; then
echo "error: this script must be run from the root of the repository"
exit 1
fi
if [[ $# -lt 1 || $# -gt 2 ]]; then
usage
fi
PR=$1
[[ "$PR" =~ ^[0-9]+$ ]] || { echo "error: PR number must be numeric"; exit 1; }
url_origin=$(git config --get remote.origin.url) || {
echo "error: no remote named 'origin' in this repository"
exit 1
}
org_repo=$(echo $url_origin | cut -d/ -f4-)
org_repo=${org_repo%.git}
echo "org/repo: $org_repo"
meta=$(curl -sSLf -H "Accept: application/vnd.github+json" "https://api.github.com/repos/$org_repo/pulls/$PR")
url_remote=$(echo "$meta" | jq -r '.head.repo.clone_url')
head_ref=$(echo "$meta" | jq -r '.head.ref')
echo "url: $url_remote"
echo "head_ref: $head_ref"
url_remote_cur=$(git config --get "remote.pr/$PR.url" 2>/dev/null || true)
if [[ "$url_remote_cur" != "$url_remote" ]]; then
git remote rm pr/$PR 2> /dev/null
git remote add pr/$PR "$url_remote"
fi
git fetch "pr/$PR" "$head_ref"
dir=$(basename $(pwd))
git branch -D pr/$PR 2> /dev/null
git worktree add -b pr/$PR ../$dir-pr-$PR pr/$PR/$head_ref 2> /dev/null
wt_path=$(cd ../$dir-pr-$PR && pwd)
echo "git worktree created in $wt_path"
cd $wt_path
git branch --set-upstream-to=pr/$PR/$head_ref
git pull --ff-only || {
echo "error: failed to pull pr/$PR"
exit 1
}
if [[ $# -eq 2 ]]; then
echo "executing: $2"
eval "$2"
fi

2
scripts/serve-static.js
View File

@@ -4,7 +4,7 @@ const path = require('path');
// This file is used for testing wasm build from emscripten
// Example build command:
// emcmake cmake -B build-wasm -DGGML_WEBGPU=ON -DLLAMA_OPENSSL=OFF
// emcmake cmake -B build-wasm -DGGML_WEBGPU=ON -DLLAMA_CURL=OFF
// cmake --build build-wasm --target test-backend-ops -j
const PORT = 8080;

22
scripts/snapdragon/adb/run-bench.sh
View File

@@ -10,23 +10,14 @@ branch=.
adbserial=
[ "$S" != "" ] && adbserial="-s $S"
adbhost=
[ "$H" != "" ] && adbhost="-H $H"
model="Llama-3.2-3B-Instruct-Q4_0.gguf"
[ "$M" != "" ] && model="$M"
device="HTP0"
[ "$D" != "" ] && device="$D"
verbose=
[ "$V" != "" ] && verbose="GGML_HEXAGON_VERBOSE=$V" cli_opts="$cli_opts -v"
experimental=
[ "$E" != "" ] && experimental="GGML_HEXAGON_EXPERIMENTAL=$E"
profile=
[ "$PROF" != "" ] && profile="GGML_HEXAGON_PROFILE=$PROF GGML_HEXAGON_OPSYNC=1" cli_opts="$cli_opts -v"
verbose=""
[ "$V" != "" ] && verbose="$V"
opmask=
[ "$OPMASK" != "" ] && opmask="GGML_HEXAGON_OPMASK=$OPMASK"
@@ -37,16 +28,13 @@ nhvx=
ndev=
[ "$NDEV" != "" ] && ndev="GGML_HEXAGON_NDEV=$NDEV"
hb=
[ "$HB" != "" ] && hb="GGML_HEXAGON_HOSTBUF=$HB"
set -x
adb $adbserial $adbhost shell " \
adb $adbserial shell " \
cd $basedir; \
LD_LIBRARY_PATH=$basedir/$branch/lib \
ADSP_LIBRARY_PATH=$basedir/$branch/lib \
$ndev $nhvx $opmask $verbose $experimental $profile $hb ./$branch/bin/llama-bench --device $device --mmap 0 -m $basedir/../gguf/$model \
$ndev $nhvx $opmask ./$branch/bin/llama-bench --device $device --mmap 0 -m $basedir/../gguf/$model \
--poll 1000 -t 6 --cpu-mask 0xfc --cpu-strict 1 \
--batch-size 128 -ngl 99 $cli_opts $@ \
--batch-size 128 -ngl 99 $@ \
"

26
scripts/snapdragon/adb/run-cli.sh
View File

@@ -12,26 +12,23 @@ branch=.
adbserial=
[ "$S" != "" ] && adbserial="-s $S"
adbhost=
[ "$H" != "" ] && adbhost="-H $H"
model="Llama-3.2-3B-Instruct-Q4_0.gguf"
[ "$M" != "" ] && model="$M"
device="HTP0"
[ "$D" != "" ] && device="$D"
verbose=
[ "$V" != "" ] && verbose="GGML_HEXAGON_VERBOSE=$V"
experimental=
[ "$E" != "" ] && experimental="GGML_HEXAGON_EXPERIMENTAL=$E"
verbose=
[ "$V" != "" ] && verbose="GGML_HEXAGON_VERBOSE=$V" cli_opts="$cli_opts -v"
sched=
[ "$SCHED" != "" ] && sched="GGML_SCHED_DEBUG=2" cli_opts="$cli_opts -v"
profile=
[ "$PROF" != "" ] && profile="GGML_HEXAGON_PROFILE=$PROF GGML_HEXAGON_OPSYNC=1" cli_opts="$cli_opts -v"
[ "$PROF" != "" ] && profile="GGML_HEXAGON_PROFILE=$PROF GGML_HEXAGON_OPSYNC=1"
opmask=
[ "$OPMASK" != "" ] && opmask="GGML_HEXAGON_OPMASK=$OPMASK"
@@ -42,18 +39,15 @@ nhvx=
ndev=
[ "$NDEV" != "" ] && ndev="GGML_HEXAGON_NDEV=$NDEV"
hb=
[ "$HB" != "" ] && hb="GGML_HEXAGON_HOSTBUF=$HB"
set -x
adb $adbserial $adbhost shell " \
adb $adbserial shell " \
cd $basedir; ulimit -c unlimited; \
LD_LIBRARY_PATH=$basedir/$branch/lib \
ADSP_LIBRARY_PATH=$basedir/$branch/lib \
$verbose $experimental $sched $opmask $profile $nhvx $ndev $hb \
./$branch/bin/llama-cli --no-mmap -m $basedir/../gguf/$model \
--poll 1000 -t 6 --cpu-mask 0xfc --cpu-strict 1 \
--ctx-size 8192 --batch-size 128 -fa on \
-ngl 99 --device $device $cli_opts $@ \
$verbose $experimental $sched $opmask $profile $nhvx $ndev \
./$branch/bin/llama-completion --no-mmap -m $basedir/../gguf/$model \
--poll 1000 -t 6 --cpu-mask 0xfc --cpu-strict 1 \
--ctx-size 8192 --batch-size 128 -ctk q8_0 -ctv q8_0 -fa on \
-ngl 99 --device $device $cli_opts $@ \
"

59
scripts/snapdragon/adb/run-completion.sh
View File

@@ -1,59 +0,0 @@
#!/bin/sh
#
# Basedir on device
basedir=/data/local/tmp/llama.cpp
cli_opts=
branch=.
[ "$B" != "" ] && branch=$B
adbserial=
[ "$S" != "" ] && adbserial="-s $S"
adbhost=
[ "$H" != "" ] && adbhost="-H $H"
model="Llama-3.2-3B-Instruct-Q4_0.gguf"
[ "$M" != "" ] && model="$M"
device="HTP0"
[ "$D" != "" ] && device="$D"
experimental=
[ "$E" != "" ] && experimental="GGML_HEXAGON_EXPERIMENTAL=$E"
verbose=
[ "$V" != "" ] && verbose="GGML_HEXAGON_VERBOSE=$V" cli_opts="$cli_opts -v"
sched=
[ "$SCHED" != "" ] && sched="GGML_SCHED_DEBUG=2" cli_opts="$cli_opts -v"
profile=
[ "$PROF" != "" ] && profile="GGML_HEXAGON_PROFILE=$PROF GGML_HEXAGON_OPSYNC=1" cli_opts="$cli_opts -v"
opmask=
[ "$OPMASK" != "" ] && opmask="GGML_HEXAGON_OPMASK=$OPMASK"
nhvx=
[ "$NHVX" != "" ] && nhvx="GGML_HEXAGON_NHVX=$NHVX"
ndev=
[ "$NDEV" != "" ] && ndev="GGML_HEXAGON_NDEV=$NDEV"
hb=
[ "$HB" != "" ] && hb="GGML_HEXAGON_HOSTBUF=$HB"
set -x
adb $adbserial $adbhost shell " \
cd $basedir; ulimit -c unlimited; \
LD_LIBRARY_PATH=$basedir/$branch/lib \
ADSP_LIBRARY_PATH=$basedir/$branch/lib \
$verbose $experimental $sched $opmask $profile $nhvx $ndev $hb \
./$branch/bin/llama-completion --no-mmap -m $basedir/../gguf/$model \
--poll 1000 -t 6 --cpu-mask 0xfc --cpu-strict 1 \
--ctx-size 8192 --batch-size 128 -fa on \
-ngl 99 -no-cnv --device $device $cli_opts $@ \
"

5
scripts/snapdragon/adb/run-mtmd.sh
View File

@@ -12,9 +12,6 @@ branch=.
adbserial=
[ "$S" != "" ] && adbserial="-s $S"
adbhost=
[ "$H" != "" ] && adbhost="-H $H"
model="gemma-3-4b-it-Q4_0.gguf"
[ "$M" != "" ] && model="$M"
@@ -54,7 +51,7 @@ mtmd_backend=
set -x
adb $adbserial $adbhost shell " \
adb $adbserial shell " \
cd $basedir; ulimit -c unlimited; \
LD_LIBRARY_PATH=$basedir/$branch/lib \
ADSP_LIBRARY_PATH=$basedir/$branch/lib \

7
scripts/snapdragon/adb/run-tool.sh
View File

@@ -12,9 +12,6 @@ branch=.
adbserial=
[ "$S" != "" ] && adbserial="-s $S"
adbhost=
[ "$H" != "" ] && adbhost="-H $H"
device="HTP0"
[ "$D" != "" ] && device="$D"
@@ -22,7 +19,7 @@ verbose=
[ "$V" != "" ] && verbose="GGML_HEXAGON_VERBOSE=$V"
experimental=
[ "$E" != "" ] && experimental="GGML_HEXAGON_EXPERIMENTAL=$E"
[ "$E" != "" ] && experimental="GGML_HEXAGON_EXPERIMENTAL=$V"
sched=
[ "$SCHED" != "" ] && sched="GGML_SCHED_DEBUG=2" cli_opts="$cli_opts -v"
@@ -46,7 +43,7 @@ set -x
tool=$1; shift
adb $adbserial $adbhost shell " \
adb $adbserial shell " \
cd $basedir; ulimit -c unlimited; \
LD_LIBRARY_PATH=$basedir/$branch/lib \
ADSP_LIBRARY_PATH=$basedir/$branch/lib \

2
scripts/sync-ggml.last
View File

@@ -1 +1 @@
ebc3a0f4a56be1c9424a89fbec09962ac34fde85
130bc125a88bb57664b88932c48c38a1cb316fac

3
scripts/sync_vendor.py
View File

@@ -16,8 +16,7 @@ vendor = {
# "https://github.com/mackron/miniaudio/raw/refs/tags/0.11.23/miniaudio.h": "vendor/miniaudio/miniaudio.h",
"https://github.com/mackron/miniaudio/raw/669ed3e844524fcd883231b13095baee9f6de304/miniaudio.h": "vendor/miniaudio/miniaudio.h",
"https://raw.githubusercontent.com/yhirose/cpp-httplib/refs/tags/v0.30.1/httplib.h": "vendor/cpp-httplib/httplib.h",
"https://raw.githubusercontent.com/yhirose/cpp-httplib/refs/tags/v0.30.1/LICENSE": "vendor/cpp-httplib/LICENSE",
"https://raw.githubusercontent.com/yhirose/cpp-httplib/refs/tags/v0.28.0/httplib.h": "vendor/cpp-httplib/httplib.h",
"https://raw.githubusercontent.com/sheredom/subprocess.h/b49c56e9fe214488493021017bf3954b91c7c1f5/subprocess.h": "vendor/sheredom/subprocess.h",
}

2
scripts/tool_bench.py
View File

@@ -7,7 +7,7 @@
Simple usage example:
cmake -B build && cmake --build build --config Release -j -t llama-server
cmake -B build -DLLAMA_CURL=1 && cmake --build build --config Release -j -t llama-server
export LLAMA_SERVER_BIN_PATH=$PWD/build/bin/llama-server
export LLAMA_CACHE=${LLAMA_CACHE:-$HOME/Library/Caches/llama.cpp}