Merge branch 'master' into pr/8836

.devops/llama-cli-cann.Dockerfile

@@ -0,0 +1,44 @@
+ARG ASCEND_VERSION=8.0.rc2.alpha003-910b-openeuler22.03-py3.8
+
+FROM cosdt/cann:$ASCEND_VERSION AS build
+
+WORKDIR /app
+
+COPY . .
+
+RUN yum install -y gcc g++ cmake make
+ENV ASCEND_TOOLKIT_HOME=/usr/local/Ascend/ascend-toolkit/latest
+ENV LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/lib64:$LIBRARY_PATH
+ENV LD_LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/lib64:${ASCEND_TOOLKIT_HOME}/lib64/plugin/opskernel:${ASCEND_TOOLKIT_HOME}/lib64/plugin/nnengine:${ASCEND_TOOLKIT_HOME}/opp/built-in/op_impl/ai_core/tbe/op_tiling:${LD_LIBRARY_PATH}
+ENV PYTHONPATH=${ASCEND_TOOLKIT_HOME}/python/site-packages:${ASCEND_TOOLKIT_HOME}/opp/built-in/op_impl/ai_core/tbe:${PYTHONPATH}
+ENV PATH=${ASCEND_TOOLKIT_HOME}/bin:${ASCEND_TOOLKIT_HOME}/compiler/ccec_compiler/bin:${PATH}
+ENV ASCEND_AICPU_PATH=${ASCEND_TOOLKIT_HOME}
+ENV ASCEND_OPP_PATH=${ASCEND_TOOLKIT_HOME}/opp
+ENV TOOLCHAIN_HOME=${ASCEND_TOOLKIT_HOME}/toolkit
+ENV ASCEND_HOME_PATH=${ASCEND_TOOLKIT_HOME}
+
+# find libascend_hal.so, because the drive hasn`t been mounted.
+ENV LD_LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/runtime/lib64/stub:$LD_LIBRARY_PATH
+
+RUN echo "Building with static libs" && \
+    source /usr/local/Ascend/ascend-toolkit/set_env.sh --force && \
+    cmake -B build -DGGML_CANN=ON -DBUILD_SHARED_LIBS=OFF  && \
+    cmake --build build --config Release --target llama-cli
+
+# TODO: use image with NNRT
+FROM cosdt/cann:$ASCEND_VERSION AS runtime
+COPY --from=build /app/build/bin/llama-cli /llama-cli
+
+ENV LC_ALL=C.utf8
+
+ENV ASCEND_TOOLKIT_HOME=/usr/local/Ascend/ascend-toolkit/latest
+ENV LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/lib64:$LIBRARY_PATH
+ENV LD_LIBRARY_PATH=${ASCEND_TOOLKIT_HOME}/lib64:${ASCEND_TOOLKIT_HOME}/lib64/plugin/opskernel:${ASCEND_TOOLKIT_HOME}/lib64/plugin/nnengine:${ASCEND_TOOLKIT_HOME}/opp/built-in/op_impl/ai_core/tbe/op_tiling:${LD_LIBRARY_PATH}
+ENV PYTHONPATH=${ASCEND_TOOLKIT_HOME}/python/site-packages:${ASCEND_TOOLKIT_HOME}/opp/built-in/op_impl/ai_core/tbe:${PYTHONPATH}
+ENV PATH=${ASCEND_TOOLKIT_HOME}/bin:${ASCEND_TOOLKIT_HOME}/compiler/ccec_compiler/bin:${PATH}
+ENV ASCEND_AICPU_PATH=${ASCEND_TOOLKIT_HOME}
+ENV ASCEND_OPP_PATH=${ASCEND_TOOLKIT_HOME}/opp
+ENV TOOLCHAIN_HOME=${ASCEND_TOOLKIT_HOME}/toolkit
+ENV ASCEND_HOME_PATH=${ASCEND_TOOLKIT_HOME}
+
+ENTRYPOINT ["/llama-cli" ]

README.md

@@ -425,6 +425,7 @@ Please refer to [Build llama.cpp locally](./docs/build.md)
 | [CUDA](./docs/build.md#cuda) | Nvidia GPU |
 | [hipBLAS](./docs/build.md#hipblas) | AMD GPU |
 | [Vulkan](./docs/build.md#vulkan) | GPU |
+| [CANN](./docs/build.md#cann) | Ascend NPU |
 
 ## Tools
 

docs/backend/CANN.md

@@ -0,0 +1,259 @@
+# llama.cpp for CANN
+
+ - [Background](#background)
+ - [News](#news)
+ - [OS](#os)
+ - [Hardware](#hardware)
+ - [Model Supports](#model-supports)
+ - [DataType Supports](#datatype-supports)
+ - [Docker](#docker)
+ - [Linux](#linux)
+ - [TODO](#todo)
+
+
+## Background
+
+**Ascend NPU** is a range of AI processors using Neural Processing Unit. It will efficiently handle matrix-matrix multiplication, dot-product and scalars.
+
+**CANN** (Compute Architecture for Neural Networks) is a heterogeneous computing architecture for AI scenarios, providing support for multiple AI frameworks on the top and serving AI processors and programming at the bottom. It plays a crucial role in bridging the gap between upper and lower layers, and is a key platform for improving the computing efficiency of Ascend AI processors. Meanwhile, it offers a highly efficient and easy-to-use programming interface for diverse application scenarios, allowing users to rapidly build AI applications and services based on the Ascend platform.
+
+**Llama.cpp + CANN**
+
+The llama.cpp CANN backend is designed to support Ascend NPU. It utilize the ability of AscendC and ACLNN which are intergrated to CANN Toolkit and kernels to using Ascend NPU directly.
+
+## News
+
+- 2024.8
+  - Support `Q4_0` and `Q8_0` data type for Ascend NPU.
+- 2024.7
+  - Create CANN backend for Ascend NPU.
+
+## OS
+
+| OS      | Status  | Verified                                       |
+|:-------:|:-------:|:----------------------------------------------:|
+| Linux   | Support | Ubuntu 22.04, OpenEuler22.03                   |
+
+
+## Hardware
+
+### Ascend NPU
+
+**Verified devices**
+| Ascend NPU                    | Status  |
+|:-----------------------------:|:-------:|
+| Atlas 300T A2                 | Support |
+
+*Notes:*
+
+- If you have trouble with Ascend NPU device, please create a issue with **[CANN]** prefix/tag.
+- If you run successfully with your Ascend NPU device, please help update the upper table.
+
+
+## Model Supports
+
+| Model Name                  | FP16  | Q8_0 | Q4_0 |
+|:----------------------------|:-----:|:----:|:----:|
+| AquilaChat2-7B              |   √   |   √  |   √  |
+| Baichuan-7b                 |   √   |   √  |   √  |
+| Baichuan2-7B-Chat           |   √   |   √  |   √  |
+| bitnet_b1_58-large          |   √   |   √  |   √  |
+| bloom-560m                  |   √   |   x  |   √  |
+| bloomz-alpaca-560m          |   √   |   x  |   √  |
+| c4ai-command-r-35B-v01      |   x   |   x  |   x  |
+| chatglm3-6B                 |   x   |   x  |   x  |
+| chinese-alpaca-2-1.3b       |   √   |   √  |   √  |
+| CodeShell-7B                |   √   |   √  |   √  |
+| deepseek-ai_deepseek-coder-1.3B-base | x |   x  |   x  |
+| deepseek-ai_DeepSeek-V2-Lite | x   |   x  |   x   |
+| deepseek-coder-6.7B-instruct | x   |   x  |   x   |
+| DeepSeek-V2-Lite-64x1.5B    |   x   |   x  |   x  |
+| falcon-7b-instruct          |   √   |   √  |   √  |
+| flan-t5-large               |   √   |   √  |   √  |
+| gemma-2-9b-it               |   √   |   √  |   √  |
+| glm-4-9B                    |   x   |   x  |   x  |
+| gpt2                        |   √   |   √  |   √  |
+| Gpt2-163M                   |   √   |   √  |   √  |
+| granite-3B-code-instruct    |   √   |   √  |   √  |
+| GritLM-7B                   |   √   |   √  |   √  |
+| internlm2_5-7b-chat         |   √   |   √  |   √  |
+| koala-7B-HF                 |   √   |   √  |   √  |
+| Llama-2-7b-chat-hf          |   √   |   √  |   √  |
+| Llama-3-Smaug-8B            |   √   |   √  |   √  |
+| Llama2-Chinese-7b-Chat      |   √   |   √  |   √  |
+| Llama3-8B                   |   √   |   √  |   √  |
+| Llama3-8b-chinese           |   √   |   √  |   √  |
+| mamba-130m-hf               |   √   |   √  |   √  |
+| Mistral-7B-Instruct-v0.2    |   √   |   √  |   √  |
+| Mixtral-8x7B-Instruct-v0.1  |   x   |   √  |   √  |
+| mpt-7B                      |   √   |   √  |   √  |
+| OLMo-1B-hf                  |   √   |   √  |   √  |
+| OpenELM-3B-Instruct         |   √   |   √  |   √  |
+| Orion-14b-base              |   √   |   √  |   √  |
+| phi1                        |   x   |   x  |   x  |
+| phi2                        |   x   |   x  |   x  |
+| Phi-3-mini-4k-instruct      |   √   |   √  |   √  |
+| plamo-13b                   |   √   |   √  |   √  |
+| pythia-70M                  |   x   |   x  |   x  |
+| Qwen-7B                     |   √   |   √  |   √  |
+| Qwen2-1.5B-Instruct         |   √   |   x  |   √  |
+| Refact-1_6B-fim             |   √   |   √  |   √  |
+| SmolLM-135M                 |   √   |   √  |   √  |
+| stablelm-zephyr             |   x   |   x  |   x  |
+| stablelm-2-zephyr-1_6b      |   x   |   x  |   x  |
+| starcoderbase-1b            |   √   |   √  |   √  |
+| starcoder2-3b               |   √   |   √  |   √  |
+| vigogne-7b-chat             |   √   |   √  |   √  |
+| xverse-7b-chat              |   √   |   √  |   √  |
+| Yi-6b-Chat                  |   √   |   √  |   √  |
+
+
+
+## DataType Supports
+
+| DataType               | Status  |
+|:----------------------:|:-------:|
+| FP16                   | Support |
+| Q8_0                   | Support |
+| Q4_0                   | Support |
+
+## Docker
+
+### Build Images
+You can get a image with llama.cpp in one command.
+```sh
+docker build -t llama-cpp-cann -f .devops/llama-cli-cann.Dockerfile .
+```
+
+### Run container
+
+```sh
+# Find all cards.
+npu-smi info
+
+# Select the cards that you want to use, make sure these cards are not used by someone.
+# Following using cards of device0.
+docker run --name llamacpp --device /dev/davinci0  --device /dev/davinci_manager --device /dev/devmm_svm --device /dev/hisi_hdc -v /usr/local/dcmi:/usr/local/dcmi -v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi -v /usr/local/Ascend/driver/lib64/:/usr/local/Ascend/driver/lib64/ -v /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info -v /PATH_TO_YOUR_MODELS/:/app/models -it llama-cpp-cann -m /app/models/MODEL_PATH -ngl 32 -p "Building a website can be done in 10 simple steps:"
+```
+
+*Notes:*
+
+- You may need to install Ascend Driver and firmware on the **host** machine *(Please refer to the [Linux configuration](#linux) for details)*.
+
+## Linux
+
+### I. Setup Environment
+
+1. **Install Ascend Driver and firmware**
+
+    ```sh
+    # create driver running user.
+    sudo groupadd -g HwHiAiUser
+    sudo useradd -g HwHiAiUser -d /home/HwHiAiUser -m HwHiAiUser -s /bin/bash
+    sudo usermod -aG HwHiAiUser $USER
+
+    # download driver from https://www.hiascend.com/hardware/firmware-drivers/community according to your system
+    # and install driver.
+    sudo sh Ascend-hdk-910b-npu-driver_x.x.x_linux-{arch}.run --full --install-for-all
+    ```
+
+    Once installed, run `npu-smi info` to check whether driver is installed successfully.
+    ```sh
+    +-------------------------------------------------------------------------------------------+
+    | npu-smi 24.1.rc2               Version: 24.1.rc2                                          |
+    +----------------------+---------------+----------------------------------------------------+
+    | NPU   Name           | Health        | Power(W)    Temp(C)           Hugepages-Usage(page)|
+    | Chip                 | Bus-Id        | AICore(%)   Memory-Usage(MB)  HBM-Usage(MB)        |
+    +======================+===============+====================================================+
+    | 2     xxx            | OK            | 64.4        51                15   / 15            |
+    | 0                    | 0000:01:00.0  | 0           1873 / 15077      0    / 32768         |
+    +======================+===============+====================================================+
+    | 5     xxx            | OK            | 64.0        52                15   / 15            |
+    | 0                    | 0000:81:00.0  | 0           1874 / 15077      0    / 32768         |
+    +======================+===============+====================================================+
+    | No running processes found in NPU 2                                                       |
+    +======================+===============+====================================================+
+    | No running processes found in NPU 5                                                       |
+    +======================+===============+====================================================+
+    ```
+
+2. **Install Ascend Firmware**
+    ```sh
+    # download driver from https://www.hiascend.com/hardware/firmware-drivers/community according to your system
+    # and install driver.
+    sudo sh Ascend-hdk-910b-npu-firmware_x.x.x.x.X.run --full
+    ```
+    If the following messaage appers, firmware is installed successfully.
+    ```sh
+    Firmware package installed successfully!
+    ```
+
+
+3. **Install CANN toolkit and kernels**
+
+    CANN toolkit and kernels can be obtained from the official [CANN Toolkit](https://www.hiascend.com/zh/developer/download/community/result?module=cann) page.
+
+    Please download the corresponding version that satified your system. The minimum version required is 8.0.RC2.alpha002 and here is the install command.
+    ```sh
+    pip3 install attrs numpy decorator sympy cffi pyyaml pathlib2 psutil protobuf scipy requests absl-py wheel typing_extensions
+    sh Ascend-cann-toolkit_8.0.RC2.alpha002_linux-aarch64.run --install
+    sh Ascend-cann-kernels-910b_8.0.RC2.alpha002_linux.run --install
+    ```
+
+    Set Ascend Variables:
+    ```sh
+    echo "source ~/Ascend/ascend-toolkit/set_env.sh" >> ~/.bashrc
+    source ~/.bashrc
+    ```
+
+Upon a successful installation, CANN is enabled for the available ascend devices.
+
+### II. Build llama.cpp
+
+```sh
+cmake -B build -DGGML_CANN=on -DCMAKE_BUILD_TYPE=release
+cmake --build build --config release
+```
+
+### III. Run the inference
+
+1. **Retrieve and prepare model**
+
+    You can refer to the general [*Prepare and Quantize*](../../README.md#prepare-and-quantize) guide for model prepration.
+
+    **Notes**:
+
+      - CANN backend only supports FP16/Q4_0/Q8_0 models currently.
+
+2. **Launch inference**
+
+    There are two device selection modes:
+
+    - Single device: Use one device target specified by the user.
+    - Multiple devices: Automatically choose the devices with the same backend.
+
+    | Device selection | Parameter                              |
+    |:----------------:|:--------------------------------------:|
+    | Single device    | --split-mode none --main-gpu DEVICE_ID |
+    | Multiple devices | --split-mode layer (default)           |
+
+    Examples:
+
+    - Use device 0:
+
+    ```sh
+    ./build/bin/llama-cli -m path_to_model -p "Building a website can be done in 10 simple steps:" -n 400 -e -ngl 33 -sm none -mg 0
+    ```
+
+    - Use multiple devices:
+
+    ```sh
+    ./build/bin/llama-cli -m path_to_model -p "Building a website can be done in 10 simple steps:" -n 400 -e -ngl 33 -sm layer
+    ```
+
+### **GitHub contribution**:
+Please add the **[CANN]** prefix/tag in issues/PRs titles to help the CANN-team check/address them without delay.
+
+
+## TODO
+- Support more models and data types.

docs/build.md

@@ -352,6 +352,31 @@ cmake --build build --config Release
 # ggml_vulkan: Using Intel(R) Graphics (ADL GT2) | uma: 1 | fp16: 1 | warp size: 32
 ```
 
+### CANN
+This provides NPU acceleration using the AI cores of your Ascend NPU. And [CANN](https://www.hiascend.com/en/software/cann) is a hierarchical APIs to help you to quickly build AI applications and service based on Ascend NPU.
+
+For more information about Ascend NPU in [Ascend Community](https://www.hiascend.com/en/).
+
+Make sure to have the CANN toolkit installed. You can download it from here: [CANN Toolkit](https://www.hiascend.com/developer/download/community/result?module=cann)
+
+Go to `llama.cpp` directory and build using CMake.
+```bash
+cmake -B build -DGGML_CANN=on -DCMAKE_BUILD_TYPE=release
+cmake --build build --config release
+```
+
+You can test with:
+
+`./build/llama-cli -m PATH_TO_MODEL -p "Building a website can be done in 10 steps:" -ngl 32`
+
+If the fllowing info is output on screen, you are using `llama.cpp by CANN backend`:
+```bash
+llm_load_tensors:       CANN buffer size = 13313.00 MiB
+llama_new_context_with_model:       CANN compute buffer size =  1260.81 MiB
+```
+
+For detailed info, such as model/device supports, CANN install, please refer to [llama.cpp for CANN](./backend/CANN.md).
+
 ### Android
 
 To read documentation for how to build on Android, [click here](./android.md)

ggml/src/ggml-rpc.cpp

@@ -82,17 +82,18 @@ static_assert(sizeof(rpc_tensor) % 8 == 0, "rpc_tensor size must be multiple of
 
 // RPC commands
 enum rpc_cmd {
-    ALLOC_BUFFER = 0,
+    RPC_CMD_ALLOC_BUFFER = 0,
-    GET_ALIGNMENT,
+    RPC_CMD_GET_ALIGNMENT,
-    GET_MAX_SIZE,
+    RPC_CMD_GET_MAX_SIZE,
-    BUFFER_GET_BASE,
+    RPC_CMD_BUFFER_GET_BASE,
-    FREE_BUFFER,
+    RPC_CMD_FREE_BUFFER,
-    BUFFER_CLEAR,
+    RPC_CMD_BUFFER_CLEAR,
-    SET_TENSOR,
+    RPC_CMD_SET_TENSOR,
-    GET_TENSOR,
+    RPC_CMD_GET_TENSOR,
-    COPY_TENSOR,
+    RPC_CMD_COPY_TENSOR,
-    GRAPH_COMPUTE,
+    RPC_CMD_GRAPH_COMPUTE,
-    GET_DEVICE_MEMORY,
+    RPC_CMD_GET_DEVICE_MEMORY,
+    RPC_CMD_COUNT,
 };
 
 // RPC data structures
@@ -330,7 +331,7 @@ GGML_CALL static void ggml_backend_rpc_buffer_free_buffer(ggml_backend_buffer_t
     uint64_t remote_ptr = ctx->remote_ptr;
     memcpy(input.data(), &remote_ptr, sizeof(remote_ptr));
     std::vector<uint8_t> output;
-    bool status = send_rpc_cmd(ctx->sock, FREE_BUFFER, input, output);
+    bool status = send_rpc_cmd(ctx->sock, RPC_CMD_FREE_BUFFER, input, output);
     GGML_ASSERT(status);
     GGML_ASSERT(output.empty());
     delete ctx;
@@ -346,7 +347,7 @@ GGML_CALL static void * ggml_backend_rpc_buffer_get_base(ggml_backend_buffer_t b
     uint64_t remote_ptr = ctx->remote_ptr;
     memcpy(input.data(), &remote_ptr, sizeof(remote_ptr));
     std::vector<uint8_t> output;
-    bool status = send_rpc_cmd(ctx->sock, BUFFER_GET_BASE, input, output);
+    bool status = send_rpc_cmd(ctx->sock, RPC_CMD_BUFFER_GET_BASE, input, output);
     GGML_ASSERT(status);
     GGML_ASSERT(output.size() == sizeof(uint64_t));
     // output serialization format: | base_ptr (8 bytes) |
@@ -405,7 +406,7 @@ GGML_CALL static void ggml_backend_rpc_buffer_set_tensor(ggml_backend_buffer_t b
     memcpy(input.data() + sizeof(rpc_tensor), &offset, sizeof(offset));
     memcpy(input.data() + sizeof(rpc_tensor) + sizeof(offset), data, size);
     std::vector<uint8_t> output;
-    bool status = send_rpc_cmd(ctx->sock, SET_TENSOR, input, output);
+    bool status = send_rpc_cmd(ctx->sock, RPC_CMD_SET_TENSOR, input, output);
     GGML_ASSERT(status);
 }
 
@@ -419,7 +420,7 @@ GGML_CALL static void ggml_backend_rpc_buffer_get_tensor(ggml_backend_buffer_t b
     memcpy(input.data() + sizeof(rpc_tensor), &offset, sizeof(offset));
     memcpy(input.data() + sizeof(rpc_tensor) + sizeof(offset), &size, sizeof(size));
     std::vector<uint8_t> output;
-    bool status = send_rpc_cmd(ctx->sock, GET_TENSOR, input, output);
+    bool status = send_rpc_cmd(ctx->sock, RPC_CMD_GET_TENSOR, input, output);
     GGML_ASSERT(status);
     GGML_ASSERT(output.size() == size);
     // output serialization format: | data (size bytes) |
@@ -444,7 +445,7 @@ GGML_CALL static bool ggml_backend_rpc_buffer_cpy_tensor(ggml_backend_buffer_t b
     memcpy(input.data(), &rpc_src, sizeof(rpc_src));
     memcpy(input.data() + sizeof(rpc_src), &rpc_dst, sizeof(rpc_dst));
     std::vector<uint8_t> output;
-    bool status = send_rpc_cmd(ctx->sock, COPY_TENSOR, input, output);
+    bool status = send_rpc_cmd(ctx->sock, RPC_CMD_COPY_TENSOR, input, output);
     GGML_ASSERT(status);
     // output serialization format: | result (1 byte) |
     GGML_ASSERT(output.size() == 1);
@@ -459,7 +460,7 @@ GGML_CALL static void ggml_backend_rpc_buffer_clear(ggml_backend_buffer_t buffer
     memcpy(input.data(), &ctx->remote_ptr, sizeof(ctx->remote_ptr));
     memcpy(input.data() + sizeof(ctx->remote_ptr), &value, sizeof(value));
     std::vector<uint8_t> output;
-    bool status = send_rpc_cmd(ctx->sock, BUFFER_CLEAR, input, output);
+    bool status = send_rpc_cmd(ctx->sock, RPC_CMD_BUFFER_CLEAR, input, output);
     GGML_ASSERT(status);
 }
 
@@ -488,7 +489,7 @@ GGML_CALL static ggml_backend_buffer_t ggml_backend_rpc_buffer_type_alloc_buffer
     memcpy(input.data(), &size, sizeof(size));
     std::vector<uint8_t> output;
     auto sock = get_socket(buft_ctx->endpoint);
-    bool status = send_rpc_cmd(sock, ALLOC_BUFFER, input, output);
+    bool status = send_rpc_cmd(sock, RPC_CMD_ALLOC_BUFFER, input, output);
     GGML_ASSERT(status);
     GGML_ASSERT(output.size() == 2*sizeof(uint64_t));
     // output serialization format: | remote_ptr (8 bytes) | remote_size (8 bytes) |
@@ -511,7 +512,7 @@ static size_t get_alignment(const std::shared_ptr<socket_t> & sock) {
     // input serialization format: | 0 bytes |
     std::vector<uint8_t> input;
     std::vector<uint8_t> output;
-    bool status = send_rpc_cmd(sock, GET_ALIGNMENT, input, output);
+    bool status = send_rpc_cmd(sock, RPC_CMD_GET_ALIGNMENT, input, output);
     GGML_ASSERT(status);
     GGML_ASSERT(output.size() == sizeof(uint64_t));
     // output serialization format: | alignment (8 bytes) |
@@ -529,7 +530,7 @@ static size_t get_max_size(const std::shared_ptr<socket_t> & sock) {
     // input serialization format: | 0 bytes |
     std::vector<uint8_t> input;
     std::vector<uint8_t> output;
-    bool status = send_rpc_cmd(sock, GET_MAX_SIZE, input, output);
+    bool status = send_rpc_cmd(sock, RPC_CMD_GET_MAX_SIZE, input, output);
     GGML_ASSERT(status);
     GGML_ASSERT(output.size() == sizeof(uint64_t));
     // output serialization format: | max_size (8 bytes) |
@@ -622,7 +623,7 @@ GGML_CALL static enum ggml_status ggml_backend_rpc_graph_compute(ggml_backend_t
     serialize_graph(cgraph, input);
     std::vector<uint8_t> output;
     auto sock = get_socket(rpc_ctx->endpoint);
-    bool status = send_rpc_cmd(sock, GRAPH_COMPUTE, input, output);
+    bool status = send_rpc_cmd(sock, RPC_CMD_GRAPH_COMPUTE, input, output);
     GGML_ASSERT(status);
     GGML_ASSERT(output.size() == 1);
     return (enum ggml_status)output[0];
@@ -636,7 +637,7 @@ GGML_CALL static bool ggml_backend_rpc_supports_op(ggml_backend_t backend, const
 }
 
 GGML_CALL static bool ggml_backend_rpc_supports_buft(ggml_backend_t backend, ggml_backend_buffer_type_t buft) {
-    if (buft->iface.get_name != ggml_backend_rpc_buffer_type_name) {
+    if (!buft || buft->iface.get_name != ggml_backend_rpc_buffer_type_name) {
         return false;
     }
     ggml_backend_rpc_buffer_type_context * buft_ctx = (ggml_backend_rpc_buffer_type_context *)buft->context;
@@ -678,6 +679,7 @@ GGML_API GGML_CALL ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const
     }
     auto sock = get_socket(endpoint);
     if (sock == nullptr) {
+        fprintf(stderr, "Failed to connect to %s\n", endpoint);
         return nullptr;
     }
     size_t alignment = get_alignment(sock);
@@ -719,7 +721,7 @@ static void get_device_memory(const std::shared_ptr<socket_t> & sock, size_t * f
     // input serialization format: | 0 bytes |
     std::vector<uint8_t> input;
     std::vector<uint8_t> output;
-    bool status = send_rpc_cmd(sock, GET_DEVICE_MEMORY, input, output);
+    bool status = send_rpc_cmd(sock, RPC_CMD_GET_DEVICE_MEMORY, input, output);
     GGML_ASSERT(status);
     GGML_ASSERT(output.size() == 2*sizeof(uint64_t));
     // output serialization format: | free (8 bytes) | total (8 bytes) |
@@ -1098,59 +1100,69 @@ static void rpc_serve_client(ggml_backend_t backend, sockfd_t sockfd, size_t fre
         if (!recv_data(sockfd, &cmd, 1)) {
             break;
         }
+        if (cmd >= RPC_CMD_COUNT) {
+            // fail fast if the command is invalid
+            fprintf(stderr, "Unknown command: %d\n", cmd);
+            break;
+        }
         std::vector<uint8_t> input;
         std::vector<uint8_t> output;
         uint64_t input_size;
         if (!recv_data(sockfd, &input_size, sizeof(input_size))) {
             break;
         }
-        input.resize(input_size);
+        try {
+            input.resize(input_size);
+        } catch (const std::bad_alloc & e) {
+            fprintf(stderr, "Failed to allocate input buffer of size %" PRIu64 "\n", input_size);
+            break;
+        }
         if (!recv_data(sockfd, input.data(), input_size)) {
             break;
         }
         bool ok = true;
         switch (cmd) {
-            case ALLOC_BUFFER: {
+            case RPC_CMD_ALLOC_BUFFER: {
                 ok = server.alloc_buffer(input, output);
                 break;
             }
-            case GET_ALIGNMENT: {
+            case RPC_CMD_GET_ALIGNMENT: {
                 server.get_alignment(output);
                 break;
             }
-            case GET_MAX_SIZE: {
+            case RPC_CMD_GET_MAX_SIZE: {
                 server.get_max_size(output);
                 break;
             }
-            case BUFFER_GET_BASE: {
+            case RPC_CMD_BUFFER_GET_BASE: {
                 ok = server.buffer_get_base(input, output);
                 break;
             }
-            case FREE_BUFFER: {
+            case RPC_CMD_FREE_BUFFER: {
                 ok = server.free_buffer(input);
                 break;
             }
-            case BUFFER_CLEAR: {
+            case RPC_CMD_BUFFER_CLEAR: {
                 ok = server.buffer_clear(input);
                 break;
             }
-            case SET_TENSOR: {
+            case RPC_CMD_SET_TENSOR: {
                 ok = server.set_tensor(input);
                 break;
             }
-            case GET_TENSOR: {
+            case RPC_CMD_GET_TENSOR: {
                 ok = server.get_tensor(input, output);
                 break;
             }
-            case COPY_TENSOR: {
+            case RPC_CMD_COPY_TENSOR: {
                 ok = server.copy_tensor(input, output);
                 break;
             }
-            case GRAPH_COMPUTE: {
+            case RPC_CMD_GRAPH_COMPUTE: {
                 ok = server.graph_compute(input, output);
                 break;
             }
-            case GET_DEVICE_MEMORY: {
+            case RPC_CMD_GET_DEVICE_MEMORY: {
                 // output serialization format: | free (8 bytes) | total (8 bytes) |
                 output.resize(2*sizeof(uint64_t), 0);
                 memcpy(output.data(), &free_mem, sizeof(free_mem));
@@ -1203,8 +1215,10 @@ void start_rpc_server(ggml_backend_t backend, const char * endpoint, size_t free
             return;
         }
         printf("Accepted client connection, free_mem=%zu, total_mem=%zu\n", free_mem, total_mem);
+        fflush(stdout);
         rpc_serve_client(backend, client_socket->fd, free_mem, total_mem);
         printf("Client connection closed\n");
+        fflush(stdout);
     }
 #ifdef _WIN32
     WSACleanup();