System Monitor (PySide6)

Fast, elegant, real‑time system monitor for desktop. Built with PySide6 + psutil. Clean dark UI, smooth charts, per‑core CPU, processes view, and optional NVIDIA GPU metrics.

快速、优雅的桌面实时系统监视器。基于 PySide6 与 psutil，深色主题、流畅曲线、按核心显示 CPU、进程视图，并可选支持 NVIDIA GPU 指标。

Screenshots | 截图

Features | 功能亮点

• Dashboard cards + time‑series charts for CPU / Memory / Network / Disk / GPU

• CPU/GPU frequency display (MHz) in dashboard cards and tabs

• CPU/GPU model names shown in dashboard cards (e.g., "AMD Ryzen 5800X", "RTX 3090")

• Separate per‑core CPU mini‑charts with distinct colors + per-core frequency labels

• Memory frequency display (when available, requires privileges on some systems)

• GPU VRAM usage charts with fixed y-axis based on total VRAM

• GPU temperature monitoring with dedicated chart (0-100°C)

• Hierarchical process tree: CPU Core → Process → Threads with lazy loading

• Process search/filter by name or PID with real-time filtering

• Pause/resume monitoring (keyboard shortcut: P)

• User‑configurable update intervals: global (default 100ms), GPU refresh, Process refresh

• Unit switcher for throughput: MB/s or MiB/s (formulas shown in UI)

• Non‑blocking GPU stats via NVML or background nvidia-smi

• Performance optimizations: Background threading for process enumeration, caching for expensive system info queries

• Efficient design patterns: Singleton cache, ThreadPoolExecutor for concurrency, Producer-Consumer pattern

• Keyboard shortcuts: P=Pause/Resume, Esc=Quit

• 总览卡片 + 曲线图，覆盖 CPU / 内存 / 网络 / 磁盘 / GPU

• CPU/GPU 频率显示（MHz）在仪表盘和各页显示

• CPU/GPU 型号名称显示在仪表盘卡片（如"AMD Ryzen 5800X"、"RTX 3090"）

• 按核心分离的 CPU 小图，每个核心不同颜色 + 每核心频率标签

• 内存频率显示（可用时，部分系统需权限）

• GPU 显存使用曲线，固定 y 轴范围基于总显存

• GPU 温度监控，专用曲线图（0-100°C）

• 分层进程树：CPU 核心 → 进程 → 线程，延迟加载

• 进程搜索/过滤，按名称或 PID 实时过滤

• 暂停/恢复监控（快捷键：P）

• 可配置更新间隔：全局（默认 100ms）、GPU 刷新、进程刷新

• 网络/磁盘速率单位可切换：MB/s 或 MiB/s（UI 显示换算公式）

• GPU 指标使用 NVML 或后台 nvidia-smi，不会阻塞界面

• 快捷键：P=暂停/恢复，Esc=退出

Install & Run | 安装与运行

• Python 3.9+
• Linux / Windows / macOS 均可（GPU 指标需 NVIDIA 工具；在 macOS/Apple Silicon 上可能不可用）

Install 依赖：

pip install -r requirements.txt
# Optional (可选，NVML 绑定):
pip install nvidia-ml-py

Run 运行：

python system_monitor/app.py
# Or use module syntax:
python -m system_monitor.app

Project Structure | 项目结构

system_monitor/
├── __init__.py
├── app.py                          # Main application entry point (154 lines)
├── core/                           # Core application logic
│   ├── __init__.py
│   ├── info_manager.py             # System information gathering (93 lines)
│   ├── metrics_collector.py        # Parallel metrics collection (126 lines)
│   ├── metrics_updater.py          # Real-time metrics update logic (233 lines)
│   ├── process_collector.py        # Background process collection (174 lines)
│   └── process_manager.py          # Process tree management (222 lines)
├── providers/                      # Data providers
│   ├── __init__.py
│   └── gpu_provider.py             # GPU metrics (NVML/nvidia-smi) (196 lines)
├── ui/                             # UI builders and event handlers
│   ├── __init__.py
│   ├── basic_tabs_builder.py       # Memory/Network/Disk tabs (79 lines)
│   ├── chart_factory.py            # Chart creation factory (53 lines)
│   ├── cpu_tab_builder.py          # CPU tab with per-core charts (106 lines)
│   ├── dashboard_builder.py        # Dashboard with metric cards (82 lines)
│   ├── event_handlers.py           # Event handling logic (85 lines)
│   ├── gpu_tab_builder.py          # GPU tab builder (89 lines)
│   ├── process_tab_builder.py      # Process/Info tabs (80 lines)
│   └── toolbar_builder.py          # Toolbar builder (78 lines)
├── utils/                          # Utility functions
│   ├── __init__.py
│   ├── cache.py                    # Caching singleton for expensive queries (98 lines)
│   ├── system_info.py              # System info helpers (188 lines)
│   └── theme.py                    # Dark theme styling (152 lines)
└── widgets/                        # Custom Qt widgets
    ├── __init__.py
    ├── metric_card.py              # Dashboard metric card (167 lines)
    └── time_series_chart.py        # Real-time chart widget (97 lines)

Usage tips | 使用提示

• Update intervals: Three controls in toolbar

  • Global interval (default 100ms): affects dashboard and all tabs
  • GPU refresh (default 100ms): separate control for GPU metrics
  • Process refresh (default 100ms): separate control for process tree updates

• Pause/Resume: Press P or click toolbar button to pause/resume monitoring

• Units (MB/s vs MiB/s): switch in Network/Disk tabs. Formulas shown in UI:

  • MB/s = bytes/s ÷ 1,000,000
  • MiB/s = bytes/s ÷ 1,048,576

• CPU tab: per‑core charts with frequency labels below each chart; summary shows processes/threads and Python asyncio task count

• Memory tab: displays RAM frequency (MHz) when available (may require root/admin on some systems)

• GPU tab: shows utilization, VRAM usage, temperature, and frequency for each GPU

• Processes tab: hierarchical tree (CPU Core → Process → Threads)

  • Click to expand CPU cores to see processes pinned to that core
  • Click processes to see their threads (lazy loaded)
  • Use search box to filter by name or PID

• Keyboard shortcuts: P = Pause/Resume, Esc = Quit

• 更新间隔：工具栏三个控制项

  • 全局间隔（默认 100ms）：影响仪表盘和所有页面
  • GPU 刷新（默认 100ms）：GPU 指标的独立控制
  • 进程刷新（默认 100ms）：进程树更新的独立控制

• 暂停/恢复：按 P 或点击工具栏按钮暂停/恢复监控

• 单位（MB/s 与 MiB/s）：在 网络/磁盘 页切换。UI 内显示换算公式：

  • MB/s = 字节/秒 ÷ 1,000,000
  • MiB/s = 字节/秒 ÷ 1,048,576

• CPU 页：每核心图表，每个图表下方有频率标签；底部显示进程/线程与 Python 协程数量

• 内存页：显示 RAM 频率（MHz），可用时（部分系统需要 root/管理员权限）

• GPU 页：显示每个 GPU 的利用率、显存使用、温度和频率

• 进程页：分层树结构（CPU 核心 → 进程 → 线程）

  • 点击展开 CPU 核心查看固定到该核心的进程
  • 点击进程查看其线程（延迟加载）
  • 使用搜索框按名称或 PID 过滤

• 快捷键：P = 暂停/恢复，Esc = 退出

Contributing | 参与贡献

PRs welcome! Fork → branch → change → PR with a short description. 欢迎提 PR！Fork → 新分支 → 修改 → 提交 PR，并附简要说明。

License | 许可证

No license specified yet. Open an issue or PR if you need one. 尚未指定许可证，如有需求请提 Issue 或 PR 讨论。