public boolean hasPrimaryKey() {
        return false;
    }

    @Override
    public boolean hasCompoundPrimaryKey() {
        return false;
    }

    @Override
    protected UniqueInfo cpui() {
        return null;
    }

    // ===================================================================================
    //                                                                           Type Name

streaming compression due to its stability and performance.

This algorithm is specifically optimized for machine generated content.
Write throughput is typically at least 500MB/s per CPU core,
and scales with the number of available CPU cores.
Decompression speed is typically at least 1GB/s.

This means that in cases where raw IO is below these numbers
compression will not only reduce disk usage but also help increase system throughput.

mkdir -p ~/.gradle/init.d && cp -v $WORKSPACE/.ci/init.gradle ~/.gradle/init.d
if [ "$(uname -m)" = "arm64" ] || [ "$(uname -m)" = "aarch64" ]; then
   MAX_WORKERS=16
elif [ -f /proc/cpuinfo ]; then
   MAX_WORKERS=`grep '^cpu\scores' /proc/cpuinfo  | uniq | sed 's/\s\+//g' |  cut -d':' -f 2`
else
   if [[ "$OSTYPE" == "darwin"* ]]; then
      MAX_WORKERS=`sysctl -n hw.physicalcpu | sed 's/\s\+//g'`
   else
      echo "Unsupported OS Type: $OSTYPE"

    // .backing_device of shape is CPU since the tensor is backed by CPU
    backing_device_name =
        TFE_TensorHandleBackingDeviceName(retvals[0], status.get());
    ASSERT_EQ(TF_OK, TF_GetCode(status.get())) << TF_Message(status.get());
    ASSERT_TRUE(absl::StrContains(backing_device_name, "CPU:0"))
        << backing_device_name;

    TFE_DeleteOp(shape_op);

  CHECK_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);

  ASSERT_EQ(2, TFE_TensorDebugInfoOnDeviceNumDims(debug_info));
  // Shape is the same for CPU tensors.
  EXPECT_EQ(3, TFE_TensorDebugInfoOnDeviceDim(debug_info, 0));
  EXPECT_EQ(2, TFE_TensorDebugInfoOnDeviceDim(debug_info, 1));

  TFE_DeleteTensorDebugInfo(debug_info);
  TFE_DeleteTensorHandle(h);

    public boolean hasPrimaryKey() {
        return false;
    }

    @Override
    public boolean hasCompoundPrimaryKey() {
        return false;
    }

    @Override
    protected UniqueInfo cpui() {
        return null;
    }

    // ===================================================================================
    //                                                                           Type Name

concurrency:
  group: ${{ github.workflow }}-${{ github.head_ref || github.ref }}
  # Don't cancel in-progress jobs for master branches.
  cancel-in-progress: ${{ github.ref != 'master' }}

jobs:
  #build-and-test-linux-x86-cpu:
  #  uses: ./.github/workflows/build-reusable.yml
  #  with:
  #    runner: 'linux-x86-n2-16'
  #    tfci: 'py313,linux_x86,rbe'

  #build-and-test-linux-x86-cuda:
  #  uses: ./.github/workflows/build-reusable.yml

          CI_DOCKER_BUILD_EXTRA_PARAMS="--build-arg py_major_minor_version=${{ matrix.pyver }} --build-arg is_nightly=${is_nightly} --build-arg tf_project_name=${tf_project_name}" \

        append(buf, "open", () -> processProbe.getOpenFileDescriptorCount()).append(',');
        append(buf, "max", () -> processProbe.getMaxFileDescriptorCount());
        buf.append("},");
        buf.append("\"cpu\":{");
        append(buf, "percent", () -> processProbe.getProcessCpuPercent()).append(',');
        append(buf, "total", () -> processProbe.getProcessCpuTotalTime());
        buf.append("},");

让我们回顾一下之前的部署概念：

* 安全性 - HTTPS
* 启动时运行
* 重新启动
* **复制（运行的进程数）**
* 内存
* 启动前的先前步骤

到目前为止，在文档中的所有教程中，您可能一直是在运行一个**服务器程序**，例如使用 `fastapi` 命令来启动 Uvicorn，而它默认运行的是**单进程模式**。

部署应用程序时，您可能希望进行一些**进程复制**，以利用**多核** CPU 并能够处理更多请求。

正如您在上一章有关[部署概念](concepts.md)中看到的，您可以使用多种策略。

在本章节中，我将向您展示如何使用 `fastapi` 命令或直接使用 `uvicorn` 命令以**多工作进程模式**运行 **Uvicorn**。

/// info | 信息