Merge branch 'ggml-org:master' into master

Author: Gadflyii

.github/workflows/build-amd.yml

@@ -0,0 +1,52 @@
+name: CI (AMD)
+
+on:
+  workflow_dispatch: # allows manual triggering
+  push:
+    branches:
+      - master
+    paths: [
+      '.github/workflows/build-amd.yml',
+      '**/CMakeLists.txt',
+      '**/.cmake',
+      '**/*.h',
+      '**/*.hpp',
+      '**/*.c',
+      '**/*.cpp',
+      '**/*.cu',
+      '**/*.cuh',
+      '**/*.comp'
+    ]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.head_ref && github.ref || github.run_id }}
+  cancel-in-progress: true
+
+jobs:
+  ggml-ci-x64-amd-vulkan:
+    runs-on: [self-hosted, Linux, X64, AMD]
+
+    steps:
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v4
+
+      - name: Test
+        id: ggml-ci
+        run: |
+          vulkaninfo --summary
+          GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
+
+  ggml-ci-x64-amd-rocm:
+    runs-on: [self-hosted, Linux, X64, AMD]
+
+    steps:
+      - name: Clone
+        id: checkout
+        uses: actions/checkout@v4
+
+      - name: Test
+        id: ggml-ci
+        run: |
+          amd-smi static
+          GG_BUILD_ROCM=1 GG_BUILD_AMDGPU_TARGETS="gfx1101" bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp

.github/workflows/build-linux-cross.yml

@@ -253,3 +253,47 @@ jobs:
                          -DCMAKE_FIND_ROOT_PATH_MODE_INCLUDE=BOTH
 
           cmake --build build --config Release -j $(nproc)
+
+  ubuntu-24-riscv64-cpu-spacemit-ime-cross:
+    runs-on: ubuntu-24.04
+
+    env:
+      SPACEMIT_IME_TOOLCHAIN_VERSION: "1.1.2"
+      SPACEMIT_IME_TOOLCHAIN_PATH: "spacemit-toolchain-linux-glibc-x86_64"
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Cache Toolchain
+        uses: actions/cache@v4
+        id: cache-spacemit-ime-cross-toolchain
+        with:
+          path: ./${{ env.SPACEMIT_IME_TOOLCHAIN_PATH }}
+          key: ${{ runner.os }}-spacemit-ime-toolchain-v${{ env.SPACEMIT_IME_TOOLCHAIN_VERSION }}
+
+      - name: Setup Toolchain
+        if: steps.cache-spacemit-ime-cross-toolchain.outputs.cache-hit != 'true'
+        run: |
+          wget --quiet --no-check-certificate https://archive.spacemit.com/toolchain/spacemit-toolchain-linux-glibc-x86_64-v${{ env.SPACEMIT_IME_TOOLCHAIN_VERSION }}.tar.xz -O ${{ env.SPACEMIT_IME_TOOLCHAIN_PATH }}.tar.xz
+          rm -rf ${{ env.SPACEMIT_IME_TOOLCHAIN_PATH }}
+          mkdir -p ${{ env.SPACEMIT_IME_TOOLCHAIN_PATH }}
+          tar xf ${{ env.SPACEMIT_IME_TOOLCHAIN_PATH }}.tar.xz -C ${{ env.SPACEMIT_IME_TOOLCHAIN_PATH }} --strip-components=1
+          rm -rf ${{ env.SPACEMIT_IME_TOOLCHAIN_PATH }}.tar.xz
+
+      - name: Build
+        run: |
+          export RISCV_ROOT_PATH=${PWD}/${{ env.SPACEMIT_IME_TOOLCHAIN_PATH }}
+          cmake -B build -DLLAMA_CURL=OFF \
+                         -DCMAKE_BUILD_TYPE=Release \
+                         -DGGML_OPENMP=OFF \
+                         -DLLAMA_BUILD_EXAMPLES=ON \
+                         -DLLAMA_BUILD_TOOLS=ON \
+                         -DLLAMA_BUILD_TESTS=OFF \
+                         -DGGML_CPU_RISCV64_SPACEMIT=ON \
+                         -DGGML_RVV=ON \
+                         -DGGML_RV_ZFH=ON \
+                         -DGGML_RV_ZICBOP=ON \
+                         -DRISCV64_SPACEMIT_IME_SPEC=RISCV64_SPACEMIT_IME1 \
+                         -DCMAKE_TOOLCHAIN_FILE=${PWD}/cmake/riscv64-spacemit-linux-gnu-gcc.cmake
+
+          cmake --build build --config Release -j $(nproc)

.github/workflows/build-riscv-native.yml

@@ -58,3 +58,63 @@ jobs:
             -DCMAKE_FIND_ROOT_PATH_MODE_INCLUDE=BOTH
 
           cmake --build build --config Release -j $(nproc)
+
+  # debian-13-riscv64-spacemit-ime-native: # Bianbu 2.2
+  #   runs-on: [self-hosted, RISCV64]
+
+  #   steps:
+  #     - name: Install prerequisites
+  #       run: |
+  #         sudo apt-get update || true
+  #         sudo apt-get install -y libatomic1
+  #     - uses: actions/checkout@v4
+  #     - name: Setup Riscv
+  #       run: |
+  #         sudo apt-get update || true
+  #         sudo apt-get install -y --no-install-recommends \
+  #                 build-essential \
+  #                 gcc-14-riscv64-linux-gnu \
+  #                 g++-14-riscv64-linux-gnu \
+  #                 ccache \
+  #                 cmake
+  #         sudo apt-get upgrade binutils -y
+
+  #     - name: Setup ccache
+  #       run: |
+  #         mkdir -p $HOME/.ccache
+  #         ccache -M 5G -d $HOME/.ccache
+  #         export CCACHE_LOGFILE=/home/runneruser/ccache_debug/ccache.log
+  #         export CCACHE_DEBUGDIR="/home/runneruser/ccache_debug"
+  #         echo "$GITHUB_WORKSPACE"
+  #         echo "CCACHE_LOGFILE=$CCACHE_LOGFILE" >> $GITHUB_ENV
+  #         echo "CCACHE_DEBUGDIR=$CCACHE_DEBUGDIR" >> $GITHUB_ENV
+  #         echo "CCACHE_BASEDIR=$GITHUB_WORKSPACE" >> $GITHUB_ENV
+  #         echo "CCACHE_DIR=$HOME/.ccache" >> $GITHUB_ENV
+
+  #     - name: Build
+  #       run: |
+  #         cmake -B build \
+  #           -DLLAMA_CURL=OFF \
+  #           -DCMAKE_BUILD_TYPE=Release \
+  #           -DGGML_OPENMP=OFF \
+  #           -DLLAMA_BUILD_EXAMPLES=ON \
+  #           -DLLAMA_BUILD_TOOLS=ON \
+  #           -DLLAMA_BUILD_TESTS=OFF \
+  #           -DCMAKE_SYSTEM_NAME=Linux \
+  #           -DCMAKE_SYSTEM_PROCESSOR=riscv64 \
+  #           -DCMAKE_C_COMPILER=riscv64-linux-gnu-gcc-14 \
+  #           -DCMAKE_CXX_COMPILER=riscv64-linux-gnu-g++-14 \
+  #           -DCMAKE_C_COMPILER_LAUNCHER=ccache \
+  #           -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
+  #           -DCMAKE_POSITION_INDEPENDENT_CODE=ON \
+  #           -DCMAKE_FIND_ROOT_PATH=/usr/lib/riscv64-linux-gnu \
+  #           -DCMAKE_FIND_ROOT_PATH_MODE_PROGRAM=NEVER \
+  #           -DCMAKE_FIND_ROOT_PATH_MODE_LIBRARY=ONLY \
+  #           -DCMAKE_FIND_ROOT_PATH_MODE_INCLUDE=BOTH \
+  #           -DGGML_RVV=ON \
+  #           -DGGML_RV_ZFH=ON \
+  #           -DGGML_RV_ZICBOP=ON \
+  #           -DGGML_CPU_RISCV64_SPACEMIT=ON \
+  #           -DRISCV64_SPACEMIT_IME_SPEC=RISCV64_SPACEMIT_IME1
+
+  #         cmake --build build --config Release -j $(nproc)

.github/workflows/build.yml

@@ -1461,34 +1461,6 @@ jobs:
         run: |
           bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
 
-#  ggml-ci-x64-amd-vulkan:
-#    runs-on: [self-hosted, Linux, X64, AMD]
-#
-#    steps:
-#      - name: Clone
-#        id: checkout
-#        uses: actions/checkout@v4
-#
-#      - name: Test
-#        id: ggml-ci
-#        run: |
-#          vulkaninfo --summary
-#          GG_BUILD_VULKAN=1 bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
-#
-#  ggml-ci-x64-amd-rocm:
-#    runs-on: [self-hosted, Linux, X64, AMD]
-#
-#    steps:
-#      - name: Clone
-#        id: checkout
-#        uses: actions/checkout@v4
-#
-#      - name: Test
-#        id: ggml-ci
-#        run: |
-#          amd-smi static
-#          GG_BUILD_ROCM=1 GG_BUILD_AMDGPU_TARGETS="gfx1101" bash ./ci/run.sh ~/results/llama.cpp /mnt/llama.cpp
-
   ggml-ci-mac-metal:
     runs-on: [self-hosted, macOS, ARM64]
 

CODEOWNERS

@@ -50,6 +50,7 @@
 /ggml/src/ggml-blas/                    @slaren
 /ggml/src/ggml-common.h                 @ggerganov @slaren
 /ggml/src/ggml-cpu/                     @ggerganov @slaren
+/ggml/src/ggml-cpu/spacemit/            @alex-spacemit
 /ggml/src/ggml-cuda/common.cuh          @slaren
 /ggml/src/ggml-cuda/fattn*              @JohannesGaessler
 /ggml/src/ggml-cuda/ggml-cuda.cu        @slaren

ci/run.sh

@@ -114,6 +114,7 @@ if [ ! -z ${GG_BUILD_NO_SVE} ]; then
     # arm 9 and newer enables sve by default, adjust these flags depending on the cpu used
     CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_NATIVE=OFF -DGGML_CPU_ARM_ARCH=armv8.5-a+fp16+i8mm"
 fi
+
 ## helpers
 
 # download a file if it does not exist or if it is outdated

cmake/riscv64-spacemit-linux-gnu-gcc.cmake

@@ -0,0 +1,29 @@
+set(CMAKE_SYSTEM_NAME Linux)
+set(CMAKE_SYSTEM_PROCESSOR riscv64)
+set(CMAKE_SYSTEM_VERSION 1)
+
+if (CMAKE_HOST_SYSTEM_PROCESSOR MATCHES "^(riscv)")
+    message(STATUS "HOST SYSTEM ${CMAKE_HOST_SYSTEM_PROCESSOR}")
+else()
+    set(GNU_MACHINE riscv64-unknown-linux-gnu CACHE STRING "GNU compiler triple")
+    if (DEFINED ENV{RISCV_ROOT_PATH})
+        file(TO_CMAKE_PATH $ENV{RISCV_ROOT_PATH} RISCV_ROOT_PATH)
+    else()
+        message(FATAL_ERROR "RISCV_ROOT_PATH env must be defined")
+    endif()
+
+    set(RISCV_ROOT_PATH ${RISCV_ROOT_PATH} CACHE STRING "root path to riscv toolchain")
+    set(CMAKE_C_COMPILER ${RISCV_ROOT_PATH}/bin/riscv64-unknown-linux-gnu-gcc)
+    set(CMAKE_CXX_COMPILER ${RISCV_ROOT_PATH}/bin/riscv64-unknown-linux-gnu-g++)
+    set(CMAKE_STRIP ${RISCV_ROOT_PATH}/bin/riscv64-unknown-linux-gnu-strip)
+    set(CMAKE_FIND_ROOT_PATH "${RISCV_ROOT_PATH}/riscv64-unknown-linux-gnu")
+    set(CMAKE_SYSROOT "${RISCV_ROOT_PATH}/sysroot")
+endif()
+
+set(CMAKE_FIND_ROOT_PATH_MODE_PROGRAM NEVER)
+set(CMAKE_FIND_ROOT_PATH_MODE_LIBRARY ONLY)
+set(CMAKE_FIND_ROOT_PATH_MODE_INCLUDE ONLY)
+set(CMAKE_FIND_ROOT_PATH_MODE_PACKAGE ONLY)
+set(CMAKE_C_FLAGS "-march=rv64gcv_zfh_zba_zicbop -mabi=lp64d ${CMAKE_C_FLAGS}")
+set(CMAKE_CXX_FLAGS "-march=rv64gcv_zfh_zba_zicbop -mabi=lp64d ${CXX_FLAGS}")
+set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -latomic")

common/chat.cpp

@@ -1616,17 +1616,36 @@ static common_chat_params common_chat_params_init_gpt_oss(const common_chat_temp
                 );
             });
 
-            auto recipient_in_role = builder.add_rule("recipient_in_role",
-                "\"<|start|>assistant\"? \" to=functions.\" ( " +
-                string_join(tool_rules_recipient_in_role, " | ") + " )"
-            );
-
             auto recipient_in_channel = builder.add_rule("recipient_in_channel",
                 channel + " \" to=functions.\" ( " +
                 string_join(tool_rules_recipient_in_channel, " | ") + " )"
             );
 
-            builder.add_rule("root", recipient_in_role + " | " + recipient_in_channel);
+            if (data.grammar_lazy) {
+                auto recipient_in_role = builder.add_rule("recipient_in_role",
+                    "\"<|start|>assistant\"? \" to=functions.\" ( " +
+                    string_join(tool_rules_recipient_in_role, " | ") + " )"
+                );
+
+                builder.add_rule("root", recipient_in_role + " | " + recipient_in_channel);
+            } else {
+                auto not_end = builder.add_rule("not-end",
+                    "[^<] | \"<\" [^|] | \"<|\" [^e] | \"<|e\" [^n] | \"<|en\" [^d] | \"<|end\" [^|] | \"<|end|\" [^>]");
+                auto analysis = builder.add_rule("analysis",
+                    "\"<|channel|>analysis<|message|>\" ( " + not_end + " )* \"<|end|>\"");
+                auto commentary = builder.add_rule("commentary",
+                    "\"<|channel|>commentary<|message|>\" ( " + not_end + " )* \"<|end|>\"");
+
+                auto recipient_in_role = builder.add_rule("recipient_in_role",
+                    "\" to=functions.\" ( " + string_join(tool_rules_recipient_in_role, " | ") + " )"
+                );
+
+                builder.add_rule("root",
+                    "( " + analysis + " \"<|start|>assistant\" )? " +
+                    "( " + commentary + " \"<|start|>assistant\" )? " +
+                    "( " + recipient_in_role + " | " + recipient_in_channel + " )"
+                );
+            }
 
             // Trigger on tool calls that appear in the commentary channel
             data.grammar_triggers.push_back({

docs/build-riscv64-spacemit.md

@@ -0,0 +1,89 @@
+> [!IMPORTANT]
+> This build documentation is specific only to RISC-V SpacemiT SOCs.
+
+## Build llama.cpp locally (for riscv64)
+
+1. Prepare Toolchain For RISCV
+~~~
+wget https://archive.spacemit.com/toolchain/spacemit-toolchain-linux-glibc-x86_64-v1.1.2.tar.xz
+~~~
+
+2. Build
+Below is the build script: it requires utilizing RISC-V vector instructions for acceleration. Ensure the `GGML_CPU_RISCV64_SPACEMIT` compilation option is enabled. The currently supported optimization version is `RISCV64_SPACEMIT_IME1`, corresponding to the `RISCV64_SPACEMIT_IME_SPEC` compilation option. Compiler configurations are defined in the `riscv64-spacemit-linux-gnu-gcc.cmake` file. Please ensure you have installed the RISC-V compiler and set the environment variable via `export RISCV_ROOT_PATH={your_compiler_path}`.
+```bash
+
+cmake -B build \
+    -DCMAKE_BUILD_TYPE=Release \
+    -DGGML_CPU_RISCV64_SPACEMIT=ON \
+    -DLLAMA_CURL=OFF \
+    -DGGML_RVV=ON \
+    -DGGML_RV_ZFH=ON \
+    -DGGML_RV_ZICBOP=ON \
+    -DRISCV64_SPACEMIT_IME_SPEC=RISCV64_SPACEMIT_IME1 \
+    -DCMAKE_TOOLCHAIN_FILE=${PWD}/cmake/riscv64-spacemit-linux-gnu-gcc.cmake \
+    -DCMAKE_INSTALL_PREFIX=build/installed
+
+cmake --build build --parallel $(nproc) --config Release
+
+pushd build
+make install
+popd
+```
+
+## Simulation
+You can use QEMU to perform emulation on non-RISC-V architectures.
+
+1. Download QEMU
+~~~
+wget https://archive.spacemit.com/spacemit-ai/qemu/jdsk-qemu-v0.0.14.tar.gz
+~~~
+
+2. Run Simulation
+After build your llama.cpp, you can run the executable file via QEMU for simulation, for example:
+~~~
+export QEMU_ROOT_PATH={your QEMU file path}
+export RISCV_ROOT_PATH_IME1={your RISC-V compiler path}
+
+${QEMU_ROOT_PATH}/bin/qemu-riscv64 -L ${RISCV_ROOT_PATH_IME1}/sysroot -cpu max,vlen=256,elen=64,vext_spec=v1.0 ${PWD}/build/bin/llama-cli -m ${PWD}/models/Qwen2.5-0.5B-Instruct-Q4_0.gguf -t 1
+~~~
+## Performance
+#### Quantization Support For Matrix
+~~~
+model name      : Spacemit(R) X60
+isa             : rv64imafdcv_zicbom_zicboz_zicntr_zicond_zicsr_zifencei_zihintpause_zihpm_zfh_zfhmin_zca_zcd_zba_zbb_zbc_zbs_zkt_zve32f_zve32x_zve64d_zve64f_zve64x_zvfh_zvfhmin_zvkt_sscofpmf_sstc_svinval_svnapot_svpbmt
+mmu             : sv39
+uarch           : spacemit,x60
+mvendorid       : 0x710
+marchid         : 0x8000000058000001
+~~~
+
+Q4_0
+|   Model    |   Size   | Params | backend | threads | test | t/s |
+| -----------| -------- | ------ | ------- | ------- | ---- |------|
+Qwen2.5 0.5B |403.20 MiB|630.17 M|   cpu   |    4    | pp512|64.12 ± 0.26|
+Qwen2.5 0.5B |403.20 MiB|630.17 M|   cpu   |    4    | tg128|10.03 ± 0.01|
+Qwen2.5 1.5B |1011.16 MiB| 1.78 B |   cpu   |    4    | pp512|24.16 ± 0.02|
+Qwen2.5 1.5B |1011.16 MiB| 1.78 B |   cpu   |    4    | tg128|3.83 ± 0.06|
+Qwen2.5 3B   | 1.86 GiB  | 3.40 B |   cpu   |    4    | pp512|12.08 ± 0.02|
+Qwen2.5 3B   | 1.86 GiB  | 3.40 B |   cpu   |    4    | tg128|2.23 ± 0.02|
+
+Q4_1
+|   Model    |   Size   | Params | backend | threads | test | t/s |
+| -----------| -------- | ------ | ------- | ------- | ---- |------|
+Qwen2.5 0.5B |351.50 MiB|494.03 M|   cpu   |    4    | pp512|62.07 ± 0.12|
+Qwen2.5 0.5B |351.50 MiB|494.03 M|   cpu   |    4    | tg128|9.91 ± 0.01|
+Qwen2.5 1.5B |964.06 MiB| 1.54 B |   cpu   |    4    | pp512|22.95 ± 0.25|
+Qwen2.5 1.5B |964.06 MiB| 1.54 B |   cpu   |    4    | tg128|4.01 ± 0.15|
+Qwen2.5 3B   | 1.85 GiB | 3.09 B |   cpu   |    4    | pp512|11.55 ± 0.16|
+Qwen2.5 3B   | 1.85 GiB | 3.09 B |   cpu   |    4    | tg128|2.25 ± 0.04|
+
+
+Q4_K
+|   Model    |   Size   | Params | backend | threads | test | t/s |
+| -----------| -------- | ------ | ------- | ------- | ---- |------|
+Qwen2.5 0.5B |462.96 MiB|630.17 M|   cpu   |    4    | pp512|9.29 ± 0.05|
+Qwen2.5 0.5B |462.96 MiB|630.17 M|   cpu   |    4    | tg128|5.67 ± 0.04|
+Qwen2.5 1.5B | 1.04 GiB | 1.78 B |   cpu   |    4    | pp512|10.38 ± 0.10|
+Qwen2.5 1.5B | 1.04 GiB | 1.78 B |   cpu   |    4    | tg128|3.17 ± 0.08|
+Qwen2.5 3B   | 1.95 GiB | 3.40 B |   cpu   |    4    | pp512|4.23 ± 0.04|
+Qwen2.5 3B   | 1.95 GiB | 3.40 B |   cpu   |    4    | tg128|1.73 ± 0.00|

ggml/CMakeLists.txt

@@ -4,8 +4,7 @@ project("ggml" C CXX ASM)
 ### GGML Version
 set(GGML_VERSION_MAJOR 0)
 set(GGML_VERSION_MINOR 9)
-set(GGML_VERSION_PATCH 0)
-set(GGML_VERSION_DEV "-dev")  # "-dev" for development, "" for releases
+set(GGML_VERSION_PATCH 3)
 set(GGML_VERSION_BASE "${GGML_VERSION_MAJOR}.${GGML_VERSION_MINOR}.${GGML_VERSION_PATCH}")
 
 find_program(GIT_EXE NAMES git git.exe NO_CMAKE_FIND_ROOT_PATH)
@@ -26,8 +25,8 @@ if(GIT_EXE)
     )
 endif()
 
-# Build the version string with optional -dev suffix and dirty flag
-set(GGML_VERSION "${GGML_VERSION_BASE}${GGML_VERSION_DEV}")
+# Build the version string with optional dirty flag
+set(GGML_VERSION "${GGML_VERSION_BASE}")
 if(GGML_GIT_DIRTY AND NOT GGML_GIT_DIRTY EQUAL 0)
     set(GGML_VERSION "${GGML_VERSION}-dirty")
 endif()