Merge mainline llama.cpp (#3)

* Merging mainline - WIP

* Merging mainline - WIP

AVX2 and CUDA appear to work.
CUDA performance seems slightly (~1-2%) lower as it is so often
the case with llama.cpp/ggml after some "improvements" have been made.

* Merging mainline - fix Metal

* Remove check

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>

15 files changed, 164 insertions, 308 deletions

diff --git a/scripts/LlamaConfig.cmake.in b/scripts/LlamaConfig.cmake.in
deleted file mode 100644
index 9311055d..00000000
--- a/scripts/LlamaConfig.cmake.in
+++ /dev/null
@@ -1,61 +0,0 @@
-set(LLAMA_VERSION @LLAMA_INSTALL_VERSION@)
-set(LLAMA_BUILD_COMMIT @LLAMA_BUILD_COMMIT@)
-set(LLAMA_BUILD_NUMBER @LLAMA_BUILD_NUMBER@)
-set(LLAMA_SHARED_LIB @BUILD_SHARED_LIBS@)
-set(LLAMA_BLAS @LLAMA_BLAS@)
-set(LLAMA_CUDA @LLAMA_CUDA@)
-set(LLAMA_METAL @LLAMA_METAL@)
-set(LLAMA_HIPBLAS @LLAMA_HIPBLAS@)
-set(LLAMA_ACCELERATE @LLAMA_ACCELERATE@)
-
-@PACKAGE_INIT@
-
-set_and_check(LLAMA_INCLUDE_DIR "@PACKAGE_LLAMA_INCLUDE_INSTALL_DIR@")
-set_and_check(LLAMA_LIB_DIR "@PACKAGE_LLAMA_LIB_INSTALL_DIR@")
-set_and_check(LLAMA_BIN_DIR "@PACKAGE_LLAMA_BIN_INSTALL_DIR@")
-
-# Ensure transient dependencies satisfied
-
-find_package(Threads REQUIRED)
-if (APPLE AND LLAMA_ACCELERATE)
-    find_library(ACCELERATE_FRAMEWORK Accelerate REQUIRED)
-endif()
-
-if (LLAMA_BLAS)
-    find_package(BLAS REQUIRED)
-endif()
-
-if (LLAMA_CUDA)
-    find_package(CUDAToolkit REQUIRED)
-endif()
-
-if (LLAMA_METAL)
-    find_library(FOUNDATION_LIBRARY Foundation REQUIRED)
-    find_library(METAL_FRAMEWORK Metal REQUIRED)
-    find_library(METALKIT_FRAMEWORK MetalKit REQUIRED)
-endif()
-
-if (LLAMA_HIPBLAS)
-    find_package(hip REQUIRED)
-    find_package(hipblas REQUIRED)
-    find_package(rocblas REQUIRED)
-endif()
-
-find_library(llama_LIBRARY llama
-    REQUIRED
-    HINTS ${LLAMA_LIB_DIR})
-
-set(_llama_link_deps "Threads::Threads" "@LLAMA_EXTRA_LIBS@")
-set(_llama_transient_defines "@LLAMA_TRANSIENT_DEFINES@")
-add_library(llama UNKNOWN IMPORTED)
-set_target_properties(llama
-    PROPERTIES
-        INTERFACE_INCLUDE_DIRECTORIES "${LLAMA_INCLUDE_DIR}"
-        INTERFACE_LINK_LIBRARIES "${_llama_link_deps}"
-        INTERFACE_COMPILE_DEFINITIONS "${_llama_transient_defines}"
-        IMPORTED_LINK_INTERFACE_LANGUAGES "CXX"
-        IMPORTED_LOCATION "${llama_LIBRARY}"
-        INTERFACE_COMPILE_FEATURES cxx_std_11
-        POSITION_INDEPENDENT_CODE ON )
-
-check_required_components(Llama)
diff --git a/scripts/build-info.cmake b/scripts/build-info.cmake
deleted file mode 100644
index ea3dc55c..00000000
--- a/scripts/build-info.cmake
+++ /dev/null
@@ -1,58 +0,0 @@
-set(BUILD_NUMBER 0)
-set(BUILD_COMMIT "unknown")
-set(BUILD_COMPILER "unknown")
-set(BUILD_TARGET "unknown")
-
-# Look for git
-find_package(Git)
-if(NOT Git_FOUND)
-    find_program(GIT_EXECUTABLE NAMES git git.exe)
-    if(GIT_EXECUTABLE)
-        set(Git_FOUND TRUE)
-        message(STATUS "Found Git: ${GIT_EXECUTABLE}")
-    else()
-        message(WARNING "Git not found. Build info will not be accurate.")
-    endif()
-endif()
-
-# Get the commit count and hash
-if(Git_FOUND)
-    execute_process(
-        COMMAND ${GIT_EXECUTABLE} rev-parse --short HEAD
-        WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}
-        OUTPUT_VARIABLE HEAD
-        OUTPUT_STRIP_TRAILING_WHITESPACE
-        RESULT_VARIABLE RES
-    )
-    if (RES EQUAL 0)
-        set(BUILD_COMMIT ${HEAD})
-    endif()
-    execute_process(
-        COMMAND ${GIT_EXECUTABLE} rev-list --count HEAD
-        WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}
-        OUTPUT_VARIABLE COUNT
-        OUTPUT_STRIP_TRAILING_WHITESPACE
-        RESULT_VARIABLE RES
-    )
-    if (RES EQUAL 0)
-        set(BUILD_NUMBER ${COUNT})
-    endif()
-endif()
-
-if(MSVC)
-    set(BUILD_COMPILER "${CMAKE_C_COMPILER_ID} ${CMAKE_C_COMPILER_VERSION}")
-    set(BUILD_TARGET ${CMAKE_VS_PLATFORM_NAME})
-else()
-    execute_process(
-        COMMAND sh -c "$@ --version | head -1" _ ${CMAKE_C_COMPILER}
-        OUTPUT_VARIABLE OUT
-        OUTPUT_STRIP_TRAILING_WHITESPACE
-    )
-    set(BUILD_COMPILER ${OUT})
-    execute_process(
-        COMMAND ${CMAKE_C_COMPILER} -dumpmachine
-        OUTPUT_VARIABLE OUT
-        OUTPUT_STRIP_TRAILING_WHITESPACE
-    )
-    set(BUILD_TARGET ${OUT})
-endif()
diff --git a/scripts/build-info.sh b/scripts/build-info.sh
index 32682afb..fa9e7bac 100755
--- a/scripts/build-info.sh
+++ b/scripts/build-info.sh
@@ -8,20 +8,20 @@ build_compiler="unknown"
 build_target="unknown"
 
 if out=$(git rev-list --count HEAD); then
-  # git is broken on WSL so we need to strip extra newlines
-  build_number=$(printf '%s' "$out" | tr -d '\n')
+    # git is broken on WSL so we need to strip extra newlines
+    build_number=$(printf '%s' "$out" | tr -d '\n')
 fi
 
 if out=$(git rev-parse --short HEAD); then
-  build_commit=$(printf '%s' "$out" | tr -d '\n')
+    build_commit=$(printf '%s' "$out" | tr -d '\n')
 fi
 
 if out=$($CC --version | head -1); then
-  build_compiler=$out
+    build_compiler=$out
 fi
 
 if out=$($CC -dumpmachine); then
-  build_target=$out
+    build_target=$out
 fi
 
 echo "int LLAMA_BUILD_NUMBER = ${build_number};"
diff --git a/scripts/check-requirements.sh b/scripts/check-requirements.sh
index 0c6afdd5..d3bbded1 100755
--- a/scripts/check-requirements.sh
+++ b/scripts/check-requirements.sh
@@ -97,9 +97,9 @@ check_requirements() {
 }
 
 check_convert_script() {
-    local py=$1             # e.g. ./convert-hf-to-gguf.py
-    local pyname=${py##*/}  # e.g. convert-hf-to-gguf.py
-    pyname=${pyname%.py}    # e.g. convert-hf-to-gguf
+    local py=$1             # e.g. ./convert_hf_to_gguf.py
+    local pyname=${py##*/}  # e.g. convert_hf_to_gguf.py
+    pyname=${pyname%.py}    # e.g. convert_hf_to_gguf
 
     info "$py: beginning check"
 
@@ -108,6 +108,11 @@ check_convert_script() {
         fatal "$py missing requirements. Expected: $reqs"
     fi
 
+    # Check that all sub-requirements are added to top-level requirements.txt
+    if ! grep -qF "$reqs" requirements.txt; then
+        fatal "$reqs needs to be added to requirements.txt"
+    fi
+
     local venv="$workdir/$pyname-venv"
     python3 -m venv "$venv"
 
@@ -134,12 +139,7 @@ EOF
 
 readonly ignore_eq_eq='check_requirements: ignore "=="'
 
-for req in "$reqs_dir"/*; do
-    # Check that all sub-requirements are added to top-level requirements.txt
-    if ! grep -qF "$req" requirements.txt; then
-        fatal "$req needs to be added to requirements.txt"
-    fi
-
+for req in */**/requirements*.txt; do
     # Make sure exact release versions aren't being pinned in the requirements
     # Filters out the ignore string
     if grep -vF "$ignore_eq_eq" "$req" | grep -q '=='; then
@@ -166,12 +166,12 @@ if (( do_cleanup )); then
     rm -rf -- "$all_venv"
 fi
 
-check_convert_script examples/convert-legacy-llama.py
-for py in convert-*.py; do
-    # skip convert-hf-to-gguf-update.py
+check_convert_script examples/convert_legacy_llama.py
+for py in convert_*.py; do
+    # skip convert_hf_to_gguf_update.py
     # TODO: the check is failing for some reason:
     #       https://github.com/ggerganov/llama.cpp/actions/runs/8875330981/job/24364557177?pr=6920
-    [[ $py == convert-hf-to-gguf-update.py ]] && continue
+    [[ $py == convert_hf_to_gguf_update.py ]] && continue
 
     check_convert_script "$py"
 done
diff --git a/scripts/compare-commits.sh b/scripts/compare-commits.sh
index a45cd396..70679f4e 100755
--- a/scripts/compare-commits.sh
+++ b/scripts/compare-commits.sh
@@ -12,7 +12,7 @@ bench_args="${@:3}"
 
 rm -f llama-bench.sqlite > /dev/null
 
-# to test a backend, call the script with the corresponding environment variable (e.g. LLAMA_CUDA=1 ./scripts/compare-commits.sh ...)
+# to test a backend, call the script with the corresponding environment variable (e.g. GGML_CUDA=1 ./scripts/compare-commits.sh ...)
 
 git checkout $1 > /dev/null
 make clean > /dev/null
diff --git a/scripts/compare-llama-bench.py b/scripts/compare-llama-bench.py
index 513dde5e..92b9e682 100755
--- a/scripts/compare-llama-bench.py
+++ b/scripts/compare-llama-bench.py
@@ -123,13 +123,13 @@ builds = cursor.execute("SELECT DISTINCT build_commit FROM test;").fetchall()
 
 try:
     repo = git.Repo(".", search_parent_directories=True)
-except git.exc.InvalidGitRepositoryError:
+except git.InvalidGitRepositoryError:
     repo = None
 
 
-def find_parent_in_data(commit):
+def find_parent_in_data(commit: git.Commit):
     """Helper function to find the most recent parent measured in number of commits for which there is data."""
-    heap = [(0, commit)]
+    heap: list[tuple[int, git.Commit]] = [(0, commit)]
     seen_hexsha8 = set()
     while heap:
         depth, current_commit = heapq.heappop(heap)
@@ -144,7 +144,7 @@ def find_parent_in_data(commit):
     return None
 
 
-def get_all_parent_hexsha8s(commit):
+def get_all_parent_hexsha8s(commit: git.Commit):
     """Helper function to recursively get hexsha8 values for all parents of a commit."""
     unvisited = [commit]
     visited   = []
diff --git a/scripts/convert-gg.sh b/scripts/convert-gg.sh
deleted file mode 100755
index 8a016843..00000000
--- a/scripts/convert-gg.sh
+++ /dev/null
@@ -1,26 +0,0 @@
-#!/bin/bash
-
-set -e
-
-# LLaMA v1
-python3 examples/convert-legacy-llama.py ../llama1/7B  --outfile models/llama-7b/ggml-model-f16.gguf  --outtype f16
-python3 examples/convert-legacy-llama.py ../llama1/13B --outfile models/llama-13b/ggml-model-f16.gguf --outtype f16
-python3 examples/convert-legacy-llama.py ../llama1/30B --outfile models/llama-30b/ggml-model-f16.gguf --outtype f16
-python3 examples/convert-legacy-llama.py ../llama1/65B --outfile models/llama-65b/ggml-model-f16.gguf --outtype f16
-
-# LLaMA v2
-python3 examples/convert-legacy-llama.py ../llama2/llama-2-7b  --outfile models/llama-7b-v2/ggml-model-f16.gguf  --outtype f16
-python3 examples/convert-legacy-llama.py ../llama2/llama-2-13b --outfile models/llama-13b-v2/ggml-model-f16.gguf --outtype f16
-python3 examples/convert-legacy-llama.py ../llama2/llama-2-70b --outfile models/llama-70b-v2/ggml-model-f16.gguf --outtype f16
-
-# Code Llama
-python3 examples/convert-legacy-llama.py ../codellama/CodeLlama-7b/  --outfile models/codellama-7b/ggml-model-f16.gguf  --outtype f16
-python3 examples/convert-legacy-llama.py ../codellama/CodeLlama-13b/ --outfile models/codellama-13b/ggml-model-f16.gguf --outtype f16
-python3 examples/convert-legacy-llama.py ../codellama/CodeLlama-34b/ --outfile models/codellama-34b/ggml-model-f16.gguf --outtype f16
-
-# Falcon
-python3 convert-falcon-hf-to-gguf.py ../falcon/falcon-7b  1
-mv -v ../falcon/falcon-7b/ggml-model-f16.gguf models/falcon-7b/ggml-model-f16.gguf
-
-python3 convert-falcon-hf-to-gguf.py ../falcon/falcon-40b 1
-mv -v ../falcon/falcon-40b/ggml-model-f16.gguf models/falcon-40b/ggml-model-f16.gguf
diff --git a/scripts/debug-test.sh b/scripts/debug-test.sh
index 7b2b601a..91946c51 100755
--- a/scripts/debug-test.sh
+++ b/scripts/debug-test.sh
@@ -110,7 +110,7 @@ rm -rf "$build_dir" && mkdir "$build_dir" || abort "Failed to make $build_dir"
 ###########################################################
 
 # Note: test-eval-callback requires -DLLAMA_CURL
-cmake -B "./$build_dir" -DCMAKE_BUILD_TYPE=Debug -DLLAMA_CUDA=1 -DLLAMA_CURL=1 || abort "Failed to build enviroment"
+cmake -B "./$build_dir" -DCMAKE_BUILD_TYPE=Debug -DGGML_CUDA=1 -DLLAMA_CURL=1 || abort "Failed to build enviroment"
 pushd "$build_dir"
 make -j || abort "Failed to compile"
 popd > /dev/null || exit 1
diff --git a/scripts/gen-build-info-cpp.cmake b/scripts/gen-build-info-cpp.cmake
deleted file mode 100644
index d8933892..00000000
--- a/scripts/gen-build-info-cpp.cmake
+++ /dev/null
@@ -1,24 +0,0 @@
-include(${CMAKE_CURRENT_SOURCE_DIR}/scripts/build-info.cmake)
-
-set(TEMPLATE_FILE "${CMAKE_CURRENT_SOURCE_DIR}/common/build-info.cpp.in")
-set(OUTPUT_FILE "${CMAKE_CURRENT_SOURCE_DIR}/common/build-info.cpp")
-
-# Only write the build info if it changed
-if(EXISTS ${OUTPUT_FILE})
-    file(READ ${OUTPUT_FILE} CONTENTS)
-    string(REGEX MATCH "LLAMA_COMMIT = \"([^\"]*)\";" _ ${CONTENTS})
-    set(OLD_COMMIT ${CMAKE_MATCH_1})
-    string(REGEX MATCH "LLAMA_COMPILER = \"([^\"]*)\";" _ ${CONTENTS})
-    set(OLD_COMPILER ${CMAKE_MATCH_1})
-    string(REGEX MATCH "LLAMA_BUILD_TARGET = \"([^\"]*)\";" _ ${CONTENTS})
-    set(OLD_TARGET ${CMAKE_MATCH_1})
-    if (
-        NOT OLD_COMMIT   STREQUAL BUILD_COMMIT   OR
-        NOT OLD_COMPILER STREQUAL BUILD_COMPILER OR
-        NOT OLD_TARGET   STREQUAL BUILD_TARGET
-    )
-        configure_file(${TEMPLATE_FILE} ${OUTPUT_FILE})
-    endif()
-else()
-    configure_file(${TEMPLATE_FILE} ${OUTPUT_FILE})
-endif()
diff --git a/scripts/gen-unicode-data.py b/scripts/gen-unicode-data.py
index 890e4d7c..2d9bde01 100644
--- a/scripts/gen-unicode-data.py
+++ b/scripts/gen-unicode-data.py
@@ -1,3 +1,5 @@
+from __future__ import annotations
+
 import array
 import unicodedata
 import requests
@@ -133,7 +135,7 @@ table_nfd.sort()
 
 
 # group ranges with same flags
-ranges_flags = [(0, codepoint_flags[0])]  # start, flags
+ranges_flags: list[tuple[int, int]] = [(0, codepoint_flags[0])]  # start, flags
 for codepoint, flags in enumerate(codepoint_flags):
     if flags != ranges_flags[-1][1]:
         ranges_flags.append((codepoint, flags))
@@ -141,11 +143,11 @@ ranges_flags.append((MAX_CODEPOINTS, 0x0000))
 
 
 # group ranges with same nfd
-ranges_nfd = [(0, 0, 0)]  # start, last, nfd
+ranges_nfd: list[tuple[int, int, int]] = [(0, 0, 0)]  # start, last, nfd
 for codepoint, norm in table_nfd:
     start = ranges_nfd[-1][0]
     if ranges_nfd[-1] != (start, codepoint - 1, norm):
-        ranges_nfd.append(None)
+        ranges_nfd.append(None)  # type: ignore[arg-type]  # dummy, will be replaced below
         start = codepoint
     ranges_nfd[-1] = (start, codepoint, norm)
 
@@ -179,13 +181,13 @@ for codepoint in table_whitespace:
 out("};\n")
 
 out("const std::unordered_map<uint32_t, uint32_t> unicode_map_lowercase = {")
-for tuple in table_lowercase:
-    out("{0x%06X, 0x%06X}," % tuple)
+for tuple_lw in table_lowercase:
+    out("{0x%06X, 0x%06X}," % tuple_lw)
 out("};\n")
 
 out("const std::unordered_map<uint32_t, uint32_t> unicode_map_uppercase = {")
-for tuple in table_uppercase:
-    out("{0x%06X, 0x%06X}," % tuple)
+for tuple_up in table_uppercase:
+    out("{0x%06X, 0x%06X}," % tuple_up)
 out("};\n")
 
 out("const std::vector<range_nfd> unicode_ranges_nfd = {  // start, last, nfd")
diff --git a/scripts/pod-llama.sh b/scripts/pod-llama.sh
index 6ba499a2..6e56e1ed 100644
--- a/scripts/pod-llama.sh
+++ b/scripts/pod-llama.sh
@@ -42,7 +42,7 @@ git clone https://github.com/ggerganov/llama.cpp
 
 cd llama.cpp
 
-LLAMA_CUDA=1 make -j
+GGML_CUDA=1 make -j
 
 ln -sfn /workspace/TinyLlama-1.1B-Chat-v0.3  ./models/tinyllama-1b
 ln -sfn /workspace/CodeLlama-7b-hf           ./models/codellama-7b
@@ -60,7 +60,7 @@ cd /workspace/llama.cpp
 mkdir build-cublas
 cd build-cublas
 
-cmake -DLLAMA_CUDA=1 ../
+cmake -DGGML_CUDA=1 ../
 make -j
 
 if [ "$1" -eq "0" ]; then
@@ -75,7 +75,7 @@ if [ "$1" -eq "1" ]; then
 
     cd /workspace/llama.cpp
 
-    python3 examples/convert-legacy-llama.py ./models/tinyllama-1b  --outfile ./models/tinyllama-1b/ggml-model-f16.gguf  --outtype f16
+    python3 examples/convert_legacy_llama.py ./models/tinyllama-1b  --outfile ./models/tinyllama-1b/ggml-model-f16.gguf  --outtype f16
 
     ./llama-quantize ./models/tinyllama-1b/ggml-model-f16.gguf ./models/tinyllama-1b/ggml-model-q4_0.gguf q4_0
     ./llama-quantize ./models/tinyllama-1b/ggml-model-f16.gguf ./models/tinyllama-1b/ggml-model-q4_k.gguf q4_k
@@ -90,7 +90,7 @@ if [ "$1" -eq "2" ]; then
 
     cd /workspace/llama.cpp
 
-    python3 examples/convert-legacy-llama.py ./models/codellama-7b  --outfile ./models/codellama-7b/ggml-model-f16.gguf  --outtype f16
+    python3 examples/convert_legacy_llama.py ./models/codellama-7b  --outfile ./models/codellama-7b/ggml-model-f16.gguf  --outtype f16
 
     ./llama-quantize ./models/codellama-7b/ggml-model-f16.gguf ./models/codellama-7b/ggml-model-q4_0.gguf q4_0
     ./llama-quantize ./models/codellama-7b/ggml-model-f16.gguf ./models/codellama-7b/ggml-model-q4_k.gguf q4_k
@@ -105,7 +105,7 @@ if [ "$1" -eq "3" ]; then
 
     cd /workspace/llama.cpp
 
-    python3 examples/convert-legacy-llama.py ./models/codellama-13b --outfile ./models/codellama-13b/ggml-model-f16.gguf --outtype f16
+    python3 examples/convert_legacy_llama.py ./models/codellama-13b --outfile ./models/codellama-13b/ggml-model-f16.gguf --outtype f16
 
     ./llama-quantize ./models/codellama-13b/ggml-model-f16.gguf ./models/codellama-13b/ggml-model-q4_0.gguf q4_0
     ./llama-quantize ./models/codellama-13b/ggml-model-f16.gguf ./models/codellama-13b/ggml-model-q4_k.gguf q4_k
@@ -120,7 +120,7 @@ if [ "$1" -eq "4" ]; then
 
     cd /workspace/llama.cpp
 
-    python3 examples/convert-legacy-llama.py ./models/codellama-34b --outfile ./models/codellama-34b/ggml-model-f16.gguf --outtype f16
+    python3 examples/convert_legacy_llama.py ./models/codellama-34b --outfile ./models/codellama-34b/ggml-model-f16.gguf --outtype f16
 
     ./llama-quantize ./models/codellama-34b/ggml-model-f16.gguf ./models/codellama-34b/ggml-model-q4_0.gguf q4_0
     ./llama-quantize ./models/codellama-34b/ggml-model-f16.gguf ./models/codellama-34b/ggml-model-q4_k.gguf q4_k
@@ -135,7 +135,7 @@ if [ "$1" -eq "5" ]; then
 
     cd /workspace/llama.cpp
 
-    python3 examples/convert-legacy-llama.py ./models/codellama-7b-instruct  --outfile ./models/codellama-7b-instruct/ggml-model-f16.gguf  --outtype f16
+    python3 examples/convert_legacy_llama.py ./models/codellama-7b-instruct  --outfile ./models/codellama-7b-instruct/ggml-model-f16.gguf  --outtype f16
 
     ./llama-quantize ./models/codellama-7b-instruct/ggml-model-f16.gguf ./models/codellama-7b-instruct/ggml-model-q4_0.gguf q4_0
     ./llama-quantize ./models/codellama-7b-instruct/ggml-model-f16.gguf ./models/codellama-7b-instruct/ggml-model-q4_k.gguf q4_k
@@ -150,7 +150,7 @@ if [ "$1" -eq "6" ]; then
 
     cd /workspace/llama.cpp
 
-    python3 examples/convert-legacy-llama.py ./models/codellama-13b-instruct --outfile ./models/codellama-13b-instruct/ggml-model-f16.gguf --outtype f16
+    python3 examples/convert_legacy_llama.py ./models/codellama-13b-instruct --outfile ./models/codellama-13b-instruct/ggml-model-f16.gguf --outtype f16
 
     ./llama-quantize ./models/codellama-13b-instruct/ggml-model-f16.gguf ./models/codellama-13b-instruct/ggml-model-q4_0.gguf q4_0
     ./llama-quantize ./models/codellama-13b-instruct/ggml-model-f16.gguf ./models/codellama-13b-instruct/ggml-model-q4_k.gguf q4_k
@@ -165,7 +165,7 @@ if [ "$1" -eq "7" ]; then
 
     cd /workspace/llama.cpp
 
-    python3 examples/convert-legacy-llama.py ./models/codellama-34b-instruct --outfile ./models/codellama-34b-instruct/ggml-model-f16.gguf --outtype f16
+    python3 examples/convert_legacy_llama.py ./models/codellama-34b-instruct --outfile ./models/codellama-34b-instruct/ggml-model-f16.gguf --outtype f16
 
     ./llama-quantize ./models/codellama-34b-instruct/ggml-model-f16.gguf ./models/codellama-34b-instruct/ggml-model-q4_0.gguf q4_0
     ./llama-quantize ./models/codellama-34b-instruct/ggml-model-f16.gguf ./models/codellama-34b-instruct/ggml-model-q4_k.gguf q4_k
@@ -186,17 +186,17 @@ if [ "$1" -eq "1" ]; then
     # batched
     cd /workspace/llama.cpp
 
-    LLAMA_CUDA=1 make -j && ./llama-batched ./models/tinyllama-1b/ggml-model-f16.gguf "Hello, my name is" 8 128 999
+    GGML_CUDA=1 make -j && ./llama-batched ./models/tinyllama-1b/ggml-model-f16.gguf "Hello, my name is" 8 128 999
 
     # batched-bench
     cd /workspace/llama.cpp
 
-    LLAMA_CUDA=1 make -j && ./llama-batched-bench ./models/tinyllama-1b/ggml-model-f16.gguf 4608 1 99 0 512 128 1,2,3,4,5,6,7,8,16,32
+    GGML_CUDA=1 make -j && ./llama-batched-bench ./models/tinyllama-1b/ggml-model-f16.gguf 4608 1 99 0 512 128 1,2,3,4,5,6,7,8,16,32
 
     # parallel
     cd /workspace/llama.cpp
 
-    LLAMA_CUDA=1 make -j && ./llama-parallel -m ./models/tinyllama-1b/ggml-model-f16.gguf -t 1 -ngl 100 -c 4096 -b 512 -s 1 -np 8 -ns 128 -n 100 -cb
+    GGML_CUDA=1 make -j && ./llama-parallel -m ./models/tinyllama-1b/ggml-model-f16.gguf -t 1 -ngl 100 -c 4096 -b 512 -s 1 -np 8 -ns 128 -n 100 -cb
 
 fi
 
@@ -204,10 +204,9 @@ fi
 #if [ "$1" -eq "7" ]; then
 #    cd /workspace/llama.cpp
 #
-#    LLAMA_CUDA=1 make -j && ./llama-speculative -m ./models/codellama-34b-instruct/ggml-model-f16.gguf -md ./models/codellama-7b-instruct/ggml-model-q4_0.gguf -p "# Dijkstra's shortest path algorithm in Python (4 spaces indentation) + complexity analysis:\n\n" -e -ngl 999 -ngld 999 -t 4 -n 512 -c 4096 -s 21 --draft 16 -np 1 --temp 0.0
+#    GGML_CUDA=1 make -j && ./llama-speculative -m ./models/codellama-34b-instruct/ggml-model-f16.gguf -md ./models/codellama-7b-instruct/ggml-model-q4_0.gguf -p "# Dijkstra's shortest path algorithm in Python (4 spaces indentation) + complexity analysis:\n\n" -e -ngl 999 -ngld 999 -t 4 -n 512 -c 4096 -s 21 --draft 16 -np 1 --temp 0.0
 #fi
 
 # more benches
-#LLAMA_CUDA=1 make -j && ./llama-batched-bench ./models/codellama-7b/ggml-model-q4_k.gguf  4096 1 99 1 512,3200 128,128,800 1
-#LLAMA_CUDA=1 make -j && ./llama-batched-bench ./models/codellama-13b/ggml-model-q4_k.gguf 4096 1 99 1 512,3200 128,128,800 1
-
+#GGML_CUDA=1 make -j && ./llama-batched-bench ./models/codellama-7b/ggml-model-q4_k.gguf  4096 1 99 1 512,3200 128,128,800 1
+#GGML_CUDA=1 make -j && ./llama-batched-bench ./models/codellama-13b/ggml-model-q4_k.gguf 4096 1 99 1 512,3200 128,128,800 1
diff --git a/scripts/server-llm.sh b/scripts/server-llm.sh
index 19923244..802592a3 100644
--- a/scripts/server-llm.sh
+++ b/scripts/server-llm.sh
@@ -380,7 +380,7 @@ fi
 
 if [[ "$backend" == "cuda" ]]; then
     printf "[+] Building with CUDA backend\n"
-    LLAMA_CUDA=1 make -j llama-server $log
+    GGML_CUDA=1 make -j llama-server $log
 elif [[ "$backend" == "cpu" ]]; then
     printf "[+] Building with CPU backend\n"
     make -j llama-server $log
diff --git a/scripts/sync-ggml-am.sh b/scripts/sync-ggml-am.sh
index 9e34dc8b..ba3bedf2 100755
--- a/scripts/sync-ggml-am.sh
+++ b/scripts/sync-ggml-am.sh
@@ -53,7 +53,9 @@ while read c; do
     fi
 
     git format-patch -k $c~1..$c --stdout -- \
-        include/ggml/ggml*.h \
+        CMakeLists.txt \
+        src/CMakeLists.txt \
+        cmake/FindSIMD.cmake \
         src/ggml*.h \
         src/ggml*.c \
         src/ggml*.cpp \
@@ -61,6 +63,8 @@ while read c; do
         src/ggml*.metal \
         src/ggml*.cu \
         src/ggml-cuda/* \
+        src/ggml-sycl/* \
+        include/ggml*.h \
         tests/test-opt.cpp \
         tests/test-grad0.cpp \
         tests/test-quantize-fns.cpp \
@@ -93,30 +97,37 @@ if [ -f $SRC_LLAMA/ggml-src.patch ]; then
 
     # replace filenames:
     #
-    # src/ggml.c                  -> ggml.c
-    # src/ggml-alloc.c            -> ggml-alloc.c
-    # src/ggml-backend-impl.h     -> ggml-backend-impl.h
-    # src/ggml-backend.c          -> ggml-backend.c
-    # src/ggml-common.h           -> ggml-common.h
-    # src/ggml-cuda/*             -> ggml-cuda/
-    # src/ggml-cuda.cu            -> ggml-cuda.cu
-    # src/ggml-cuda.h             -> ggml-cuda.h
-    # src/ggml-impl.h             -> ggml-impl.h
-    # src/ggml-kompute.cpp        -> ggml-kompute.cpp
-    # src/ggml-kompute.h          -> ggml-kompute.h
-    # src/ggml-metal.h            -> ggml-metal.h
-    # src/ggml-metal.m            -> ggml-metal.m
-    # src/ggml-quants.c           -> ggml-quants.c
-    # src/ggml-quants.h           -> ggml-quants.h
-    # src/ggml-rpc.cpp            -> ggml-rpc.cpp
-    # src/ggml-rpc.h              -> ggml-rpc.h
-    # src/ggml-sycl.cpp           -> ggml-sycl.cpp
-    # src/ggml-sycl.h             -> ggml-sycl.h
-    # src/ggml-vulkan.cpp         -> ggml-vulkan.cpp
-    # src/ggml-vulkan.h           -> ggml-vulkan.h
-    # include/ggml/ggml.h         -> ggml.h
-    # include/ggml/ggml-alloc.h   -> ggml-alloc.h
-    # include/ggml/ggml-backend.h -> ggml-backend.h
+    # CMakelists.txt          -> ggml/CMakeLists.txt
+    # src/CMakeLists.txt      -> ggml/src/CMakeLists.txt
+    # cmake/FindSIMD.cmake    -> ggml/cmake/FindSIMD.cmake
+    #
+    # src/ggml.c              -> ggml/src/ggml.c
+    # src/ggml-alloc.c        -> ggml/src/ggml-alloc.c
+    # src/ggml-backend-impl.h -> ggml/src/ggml-backend-impl.h
+    # src/ggml-backend.c      -> ggml/src/ggml-backend.c
+    # src/ggml-common.h       -> ggml/src/ggml-common.h
+    # src/ggml-cuda/*         -> ggml/src/ggml-cuda/
+    # src/ggml-cuda.cu        -> ggml/src/ggml-cuda.cu
+    # src/ggml-impl.h         -> ggml/src/ggml-impl.h
+    # src/ggml-kompute.cpp    -> ggml/src/ggml-kompute.cpp
+    # src/ggml-metal.m        -> ggml/src/ggml-metal.m
+    # src/ggml-quants.c       -> ggml/src/ggml-quants.c
+    # src/ggml-quants.h       -> ggml/src/ggml-quants.h
+    # src/ggml-rpc.cpp        -> ggml/src/ggml-rpc.cpp
+    # src/ggml-sycl/*         -> ggml/src/ggml-sycl/
+    # src/ggml-sycl.cpp       -> ggml/src/ggml-sycl.cpp
+    # src/ggml-vulkan.cpp     -> ggml/src/ggml-vulkan.cpp
+    #
+    # include/ggml.h         -> ggml/include/ggml.h
+    # include/ggml-alloc.h   -> ggml/include/ggml-alloc.h
+    # include/ggml-backend.h -> ggml/include/ggml-backend.h
+    # include/ggml-blas.h    -> ggml/include/ggml-blas.h
+    # include/ggml-cuda.h    -> ggml/include/ggml-cuda.h
+    # include/ggml-kompute.h -> ggml/include/ggml-kompute.h
+    # include/ggml-metal.h   -> ggml/include/ggml-metal.h
+    # include/ggml-rpc.h     -> ggml/include/ggml-rpc.h
+    # include/ggml-sycl.h    -> ggml/include/ggml-sycl.h
+    # include/ggml-vulkan.h  -> ggml/include/ggml-vulkan.h
     #
     # tests/test-opt.cpp           -> tests/test-opt.cpp
     # tests/test-grad0.cpp         -> tests/test-grad0.cpp
@@ -124,41 +135,45 @@ if [ -f $SRC_LLAMA/ggml-src.patch ]; then
     # tests/test-quantize-perf.cpp -> tests/test-quantize-perf.cpp
     # tests/test-backend-ops.cpp   -> tests/test-backend-ops.cpp
     #
-    # LICENSE                      -> LICENSE
-    # scripts/gen-authors.sh       -> scripts/gen-authors.sh
-
-    cat ggml-src.patch | sed \
-        -e 's/src\/ggml\.c/ggml.c/g' \
-        -e 's/src\/ggml-alloc\.c/ggml-alloc.c/g' \
-        -e 's/src\/ggml-backend-impl\.h/ggml-backend-impl.h/g' \
-        -e 's/src\/ggml-backend\.c/ggml-backend.c/g' \
-        -e 's/src\/ggml-common\.h/ggml-common.h/g' \
-        -e 's/src\/ggml-cuda\//ggml-cuda\//g' \
-        -e 's/src\/ggml-cuda\.cu/ggml-cuda.cu/g' \
-        -e 's/src\/ggml-cuda\.h/ggml-cuda.h/g' \
-        -e 's/src\/ggml-impl\.h/ggml-impl.h/g' \
-        -e 's/src\/ggml-kompute\.cpp/ggml-kompute.cpp/g' \
-        -e 's/src\/ggml-kompute\.h/ggml-kompute.h/g' \
-        -e 's/src\/ggml-metal\.h/ggml-metal.h/g' \
-        -e 's/src\/ggml-metal\.m/ggml-metal.m/g' \
-        -e 's/src\/ggml-quants\.c/ggml-quants.c/g' \
-        -e 's/src\/ggml-quants\.h/ggml-quants.h/g' \
-        -e 's/src\/ggml-rpc\.cpp/ggml-rpc.cpp/g' \
-        -e 's/src\/ggml-rpc\.h/ggml-rpc.h/g' \
-        -e 's/src\/ggml-sycl\.cpp/ggml-sycl.cpp/g' \
-        -e 's/src\/ggml-sycl\.h/ggml-sycl.h/g' \
-        -e 's/src\/ggml-vulkan\.cpp/ggml-vulkan.cpp/g' \
-        -e 's/src\/ggml-vulkan\.h/ggml-vulkan.h/g' \
-        -e 's/include\/ggml\/ggml\.h/ggml.h/g' \
-        -e 's/include\/ggml\/ggml-alloc\.h/ggml-alloc.h/g' \
-        -e 's/include\/ggml\/ggml-backend\.h/ggml-backend.h/g' \
-        -e 's/tests\/test-opt\.cpp/tests\/test-opt.cpp/g' \
-        -e 's/tests\/test-grad0\.cpp/tests\/test-grad0.cpp/g' \
-        -e 's/tests\/test-quantize-fns\.cpp/tests\/test-quantize-fns.cpp/g' \
-        -e 's/tests\/test-quantize-perf\.cpp/tests\/test-quantize-perf.cpp/g' \
-        -e 's/tests\/test-backend-ops\.cpp/tests\/test-backend-ops.cpp/g' \
-        -e 's/LICENSE/LICENSE/g' \
-        -e 's/scripts\/gen-authors\.sh/scripts\/gen-authors.sh/g' \
+    # LICENSE                -> LICENSE
+    # scripts/gen-authors.sh -> scripts/gen-authors.sh
+
+    cat ggml-src.patch | sed -E \
+        -e 's/([[:space:]]|[ab]\/)CMakeLists.txt/\1ggml\/CMakeLists.txt/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/CMakeLists.txt/\1ggml\/src\/CMakeLists.txt/g' \
+        -e 's/([[:space:]]|[ab]\/)cmake\/FindSIMD.cmake/\1ggml\/cmake\/FindSIMD.cmake/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml\.c/\1ggml\/src\/ggml.c/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-alloc\.c/\1ggml\/src\/ggml-alloc.c/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-backend-impl\.h/\1ggml\/src\/ggml-backend-impl.h/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-backend\.c/\1ggml\/src\/ggml-backend.c/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-common\.h/\1ggml\/src\/ggml-common.h/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-cuda\//\1ggml\/src\/ggml-cuda\//g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-cuda\.cu/\1ggml\/src\/ggml-cuda.cu/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-impl\.h/\1ggml\/src\/ggml-impl.h/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-kompute\.cpp/\1ggml\/src\/ggml-kompute.cpp/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-metal\.m/\1ggml\/src\/ggml-metal.m/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-quants\.c/\1ggml\/src\/ggml-quants.c/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-quants\.h/\1ggml\/src\/ggml-quants.h/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-rpc\.cpp/\1ggml\/src\/ggml-rpc.cpp/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-sycl\//\1ggml\/src\/ggml-sycl\//g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-sycl\.cpp/\1ggml\/src\/ggml-sycl.cpp/g' \
+        -e 's/([[:space:]]|[ab]\/)src\/ggml-vulkan\.cpp/\1ggml\/src\/ggml-vulkan.cpp/g' \
+        -e 's/([[:space:]]|[ab]\/)include\/ggml\.h/\1ggml\/include\/ggml.h/g' \
+        -e 's/([[:space:]]|[ab]\/)include\/ggml-alloc\.h/\1ggml\/include\/ggml-alloc.h/g' \
+        -e 's/([[:space:]]|[ab]\/)include\/ggml-backend\.h/\1ggml\/include\/ggml-backend.h/g' \
+        -e 's/([[:space:]]|[ab]\/)include\/ggml-blas\.h/\1ggml\/include\/ggml-blas.h/g' \
+        -e 's/([[:space:]]|[ab]\/)include\/ggml-cuda\.h/\1ggml\/include\/ggml-cuda.h/g' \
+        -e 's/([[:space:]]|[ab]\/)include\/ggml-kompute\.h/\1ggml\/include\/ggml-kompute.h/g' \
+        -e 's/([[:space:]]|[ab]\/)include\/ggml-metal\.h/\1ggml\/include\/ggml-metal.h/g' \
+        -e 's/([[:space:]]|[ab]\/)include\/ggml-rpc\.h/\1ggml\/include\/ggml-rpc.h/g' \
+        -e 's/([[:space:]]|[ab]\/)include\/ggml-sycl\.h/\1ggml\/include\/ggml-sycl.h/g' \
+        -e 's/([[:space:]]|[ab]\/)include\/ggml-vulkan\.h/\1ggml\/include\/ggml-vulkan.h/g' \
+        -e 's/([[:space:]]|[ab]\/)examples\/common\.h/examples\/common.h/g' \
+        -e 's/([[:space:]]|[ab]\/)examples\/common\.cpp/examples\/common.cpp/g' \
+        -e 's/([[:space:]]|[ab]\/)examples\/common-ggml\.h/examples\/common-ggml.h/g' \
+        -e 's/([[:space:]]|[ab]\/)examples\/common-ggml\.cpp/examples\/common-ggml.cpp/g' \
+        -e 's/([[:space:]]|[ab]\/)LICENSE/LICENSE/g' \
+        -e 's/([[:space:]]|[ab]\/)scripts\/gen-authors\.sh/scripts\/gen-authors.sh/g' \
         > ggml-src.patch.tmp
     mv ggml-src.patch.tmp ggml-src.patch
 
diff --git a/scripts/sync-ggml.last b/scripts/sync-ggml.last
index b6c57ec5..80159b70 100644
--- a/scripts/sync-ggml.last
+++ b/scripts/sync-ggml.last
@@ -1 +1 @@
-5653a195935ea3ac54652644c9daf154dbc1571b
+e3b3846976c94163f2b3dd128cc959782653edbb
diff --git a/scripts/sync-ggml.sh b/scripts/sync-ggml.sh
index 4843f8a4..402446ef 100755
--- a/scripts/sync-ggml.sh
+++ b/scripts/sync-ggml.sh
@@ -1,34 +1,43 @@
 #!/bin/bash
 
-cp -rpv ../ggml/src/ggml.c                  ./ggml.c
-cp -rpv ../ggml/src/ggml-alloc.c            ./ggml-alloc.c
-cp -rpv ../ggml/src/ggml-backend-impl.h     ./ggml-backend-impl.h
-cp -rpv ../ggml/src/ggml-backend.c          ./ggml-backend.c
-cp -rpv ../ggml/src/ggml-common.h           ./ggml-common.h
-cp -rpv ../ggml/src/ggml-cuda/*             ./ggml-cuda/
-cp -rpv ../ggml/src/ggml-cuda.cu            ./ggml-cuda.cu
-cp -rpv ../ggml/src/ggml-cuda.h             ./ggml-cuda.h
-cp -rpv ../ggml/src/ggml-impl.h             ./ggml-impl.h
-cp -rpv ../ggml/src/ggml-kompute.cpp        ./ggml-kompute.cpp
-cp -rpv ../ggml/src/ggml-kompute.h          ./ggml-kompute.h
-cp -rpv ../ggml/src/ggml-metal.h            ./ggml-metal.h
-cp -rpv ../ggml/src/ggml-metal.m            ./ggml-metal.m
-cp -rpv ../ggml/src/ggml-metal.metal        ./ggml-metal.metal
-cp -rpv ../ggml/src/ggml-quants.c           ./ggml-quants.c
-cp -rpv ../ggml/src/ggml-quants.h           ./ggml-quants.h
-cp -rpv ../ggml/src/ggml-rpc.cpp            ./ggml-rpc.cpp
-cp -rpv ../ggml/src/ggml-rpc.h              ./ggml-rpc.h
-cp -rpv ../ggml/src/ggml-sycl.cpp           ./ggml-sycl.cpp
-cp -rpv ../ggml/src/ggml-sycl.h             ./ggml-sycl.h
-cp -rpv ../ggml/src/ggml-vulkan.cpp         ./ggml-vulkan.cpp
-cp -rpv ../ggml/src/ggml-vulkan.h           ./ggml-vulkan.h
-cp -rpv ../ggml/include/ggml/ggml.h         ./ggml.h
-cp -rpv ../ggml/include/ggml/ggml-alloc.h   ./ggml-alloc.h
-cp -rpv ../ggml/include/ggml/ggml-backend.h ./ggml-backend.h
+cp -rpv ../ggml/CMakeLists.txt       ./ggml/CMakeLists.txt
+cp -rpv ../ggml/src/CMakeLists.txt   ./ggml/src/CMakeLists.txt
+cp -rpv ../ggml/cmake/FindSIMD.cmake ./ggml/cmake/FindSIMD.cmake
 
-cp -rpv ../ggml/tests/test-opt.cpp         ./tests/test-opt.cpp
-cp -rpv ../ggml/tests/test-grad0.cpp       ./tests/test-grad0.cpp
-cp -rpv ../ggml/tests/test-backend-ops.cpp ./tests/test-backend-ops.cpp
+cp -rpv ../ggml/src/ggml.c              ./ggml/src/ggml.c
+cp -rpv ../ggml/src/ggml-alloc.c        ./ggml/src/ggml-alloc.c
+cp -rpv ../ggml/src/ggml-backend-impl.h ./ggml/src/ggml-backend-impl.h
+cp -rpv ../ggml/src/ggml-backend.c      ./ggml/src/ggml-backend.c
+cp -rpv ../ggml/src/ggml-common.h       ./ggml/src/ggml-common.h
+cp -rpv ../ggml/src/ggml-cuda/*         ./ggml/src/ggml-cuda/
+cp -rpv ../ggml/src/ggml-cuda.cu        ./ggml/src/ggml-cuda.cu
+cp -rpv ../ggml/src/ggml-impl.h         ./ggml/src/ggml-impl.h
+cp -rpv ../ggml/src/ggml-kompute.cpp    ./ggml/src/ggml-kompute.cpp
+cp -rpv ../ggml/src/ggml-metal.m        ./ggml/src/ggml-metal.m
+cp -rpv ../ggml/src/ggml-metal.metal    ./ggml/src/ggml-metal.metal
+cp -rpv ../ggml/src/ggml-quants.c       ./ggml/src/ggml-quants.c
+cp -rpv ../ggml/src/ggml-quants.h       ./ggml/src/ggml-quants.h
+cp -rpv ../ggml/src/ggml-rpc.cpp        ./ggml/src/ggml-rpc.cpp
+cp -rpv ../ggml/src/ggml-sycl/*         ./ggml/src/ggml-sycl/
+cp -rpv ../ggml/src/ggml-sycl.cpp       ./ggml/src/ggml-sycl.cpp
+cp -rpv ../ggml/src/ggml-vulkan.cpp     ./ggml/src/ggml-vulkan.cpp
 
-cp -rpv ../LICENSE                         ./LICENSE
-cp -rpv ../ggml/scripts/gen-authors.sh     ./scripts/gen-authors.sh
+cp -rpv ../ggml/include/ggml.h         ./ggml/include/ggml.h
+cp -rpv ../ggml/include/ggml-alloc.h   ./ggml/include/ggml-alloc.h
+cp -rpv ../ggml/include/ggml-backend.h ./ggml/include/ggml-backend.h
+cp -rpv ../ggml/include/ggml-blas.h    ./ggml/include/ggml-blas.h
+cp -rpv ../ggml/include/ggml-cuda.h    ./ggml/include/ggml-cuda.h
+cp -rpv ../ggml/include/ggml-kompute.h ./ggml/include/ggml-kompute.h
+cp -rpv ../ggml/include/ggml-metal.h   ./ggml/include/ggml-metal.h
+cp -rpv ../ggml/include/ggml-rpc.h     ./ggml/include/ggml-rpc.h
+cp -rpv ../ggml/include/ggml-sycl.h    ./ggml/include/ggml-sycl.h
+cp -rpv ../ggml/include/ggml-vulkan.h  ./ggml/include/ggml-vulkan.h
+
+cp -rpv ../ggml/tests/test-opt.cpp           ./tests/test-opt.cpp
+cp -rpv ../ggml/tests/test-grad0.cpp         ./tests/test-grad0.cpp
+cp -rpv ../ggml/tests/test-quantize-fns.cpp  ./tests/test-quantize-fns.cpp
+cp -rpv ../ggml/tests/test-quantize-perf.cpp ./tests/test-quantize-perf.cpp
+cp -rpv ../ggml/tests/test-backend-ops.cpp   ./tests/test-backend-ops.cpp
+
+cp -rpv ../LICENSE                     ./LICENSE
+cp -rpv ../ggml/scripts/gen-authors.sh ./scripts/gen-authors.sh