sysprog
      • Sharing URL Link copied
      • /edit
      • View mode
        • Edit mode
        • View mode
        • Book mode
        • Slide mode
        Edit mode View mode Book mode Slide mode
      • Customize slides
      • Note Permission
      • Read
        • Owners
        • Signed-in users
        • Everyone
        Owners Signed-in users Everyone
      • Write
        • Owners
        • Signed-in users
        • Everyone
        Owners Signed-in users Everyone
      • Engagement control Commenting, Suggest edit, Emoji Reply
    • Invite by email
      Invitee

      This note has no invitees

    • Publish Note

      Share your work with the world Congratulations! 🎉 Your note is out in the world Publish Note

      Your note will be visible on your profile and discoverable by anyone.
      Your note is now live.
      This note is visible on your profile and discoverable online.
      Everyone on the web can find and read all notes of this public team.
      See published notes
      Unpublish note
      Please check the box to agree to the Community Guidelines.
      View profile
    • Commenting
      Permission
      Disabled Forbidden Owners Signed-in users Everyone
    • Enable
    • Permission
      • Forbidden
      • Owners
      • Signed-in users
      • Everyone
    • Suggest edit
      Permission
      Disabled Forbidden Owners Signed-in users Everyone
    • Enable
    • Permission
      • Forbidden
      • Owners
      • Signed-in users
    • Emoji Reply
    • Enable
    • Versions and GitHub Sync
    • Note settings
    • Note Insights
    • Engagement control
    • Transfer ownership
    • Delete this note
    • Insert from template
    • Import from
      • Dropbox
      • Google Drive
      • Gist
      • Clipboard
    • Export to
      • Dropbox
      • Google Drive
      • Gist
    • Download
      • Markdown
      • HTML
      • Raw HTML
Menu Note settings Versions and GitHub Sync Note Insights Sharing URL Help
Menu
Options
Engagement control Transfer ownership Delete this note
Import from
Dropbox Google Drive Gist Clipboard
Export to
Dropbox Google Drive Gist
Download
Markdown HTML Raw HTML
Back
Sharing URL Link copied
/edit
View mode
  • Edit mode
  • View mode
  • Book mode
  • Slide mode
Edit mode View mode Book mode Slide mode
Customize slides
Note Permission
Read
Owners
  • Owners
  • Signed-in users
  • Everyone
Owners Signed-in users Everyone
Write
Owners
  • Owners
  • Signed-in users
  • Everyone
Owners Signed-in users Everyone
Engagement control Commenting, Suggest edit, Emoji Reply
  • Invite by email
    Invitee

    This note has no invitees

  • Publish Note

    Share your work with the world Congratulations! 🎉 Your note is out in the world Publish Note

    Your note will be visible on your profile and discoverable by anyone.
    Your note is now live.
    This note is visible on your profile and discoverable online.
    Everyone on the web can find and read all notes of this public team.
    See published notes
    Unpublish note
    Please check the box to agree to the Community Guidelines.
    View profile
    Engagement control
    Commenting
    Permission
    Disabled Forbidden Owners Signed-in users Everyone
    Enable
    Permission
    • Forbidden
    • Owners
    • Signed-in users
    • Everyone
    Suggest edit
    Permission
    Disabled Forbidden Owners Signed-in users Everyone
    Enable
    Permission
    • Forbidden
    • Owners
    • Signed-in users
    Emoji Reply
    Enable
    Import from Dropbox Google Drive Gist Clipboard
       owned this note    owned this note      
    Published Linked with GitHub
    1
    Subscribed
    • Any changes
      Be notified of any changes
    • Mention me
      Be notified of mention me
    • Unsubscribe
    Subscribe
    # Linux 核心專題: 分析 BitNet 的系統效能表現 > 執行人: willy-liu > [專題解說影片](https://youtu.be/vMq8v2Kho-4) ## 任務簡述 BitNet b1.58 的運行效率極高。在 Apple M2 這類通用 CPU 上可達 29 毫秒延遲,且記憶體佔用僅為 0.4GB,遠小於例如 Gemma 3 1B 所需的 1.4GB。微軟團隊在基準測試中將其與 Llama 3.2 1B、Gemma 3 1B 與 Qwen 2.5 1.5B 等全精度模型對比,發現 BitNet 即使在參數較少、權重精度較低的情況下,仍能在 MMLU、GSM8K、MATH 等任務中維持穩定表現,並在 GSM8K 上奪得最佳成績。BitNet 於 30 億參數以上的規模下,其性能已能接近 FP16 模型,並於 70 億參數規模達成高達 4 倍的推理速度提升與 7 倍的記憶體節省,顯示此技術具備高度擴展潛力。 本任務嘗試運用 Linux 核心提供的效能分析工具,定位出 BitNet 運行時期的效能瓶頸,歸納出推理過程中的程式熱點並予以分析 ## TODO: 探討 BitNet 的關鍵突破和原理 BitNet是微軟研究院與中國科學院研究人員在2023年發表的一個 [1-bit transformers for LLM](https://arxiv.org/abs/2310.11453) 架構,原生訓練純二值(1‑bit,權重僅為 ±1)BitLinear 模組,和其他研究不同,BitNet是原生用 1 Bit 從頭訓練,而不是先訓練全精度再量化,目標是達到與 FP16/FP32 Transformers 相當的效能,同時顯著減少記憶體佔用與能源消耗。 在 2024 年 2 月,團隊進一步提出 [BitNet b1.58 2B4T](https://arxiv.org/abs/2504.12285) 模型,使用1.58-bit權重量化(即三值 {-1, 0, +1})與8-bit激活值(float32->int8)方案,在2B參數規模、4兆token語料上從零訓練,效能可與同級全精度開源模型媲美,同時支援將訓練好的模型在bitnet.cpp(llama.cpp)框架等進行高效CPU/GPU推理;可以用在更輕量的平台。 (p.s. 累加和矩陣運算仍以 int32 或 float32 處理,這樣能兼顧訓練精度與推理速度。) ## TODO: 定位 BitNet 運行時期的效能瓶頸 > 運用 Linux 核心提供的效能分析工具,如 perf,也搭配 [uftrace](https://github.com/namhyung/uftrace) 一類使用者層級的工具,歸納出推理過程中的程式熱點並予以分析 :::spoiler 實驗環境 ```bash (bitnet-cpp) willy@linux2025:~/Desktop/linux2025$ cat /etc/os-release PRETTY_NAME="Ubuntu 24.04.2 LTS" NAME="Ubuntu" VERSION_ID="24.04" VERSION="24.04.2 LTS (Noble Numbat)" VERSION_CODENAME=noble ID=ubuntu ID_LIKE=debian HOME_URL="https://www.ubuntu.com/" SUPPORT_URL="https://help.ubuntu.com/" BUG_REPORT_URL="https://bugs.launchpad.net/ubuntu/" PRIVACY_POLICY_URL="https://www.ubuntu.com/legal/terms-and-policies/privacy-policy" UBUNTU_CODENAME=noble LOGO=ubuntu-logo (bitnet-cpp) willy@linux2025:~/Desktop/linux2025$ lscpu 架構: x86_64 CPU 作業模式: 32-bit, 64-bit Address sizes: 39 bits physical, 48 bits virtual Byte Order: Little Endian CPU(s): 12 On-line CPU(s) list: 0-11 供應商識別號: GenuineIntel Model name: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz CPU 家族: 6 型號: 158 每核心執行緒數: 2 每通訊端核心數: 6 Socket(s): 1 製程: 10 CPU(s) scaling MHz: 60% CPU max MHz: 4600.0000 CPU min MHz: 800.0000 BogoMIPS: 6399.96 Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe sys call nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni p clmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_dea dline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb pti ssbd ibrs ibpb stibp tpr_shadow flexpri ority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx rdseed adx smap clflushopt intel_pt xsaveopt xsavec xgetbv1 xsaves dtherm ida arat pln pts hwp hwp_notify hwp_act_window hwp_epp vnmi md_clear flush_l1d arch_capabilit ies Virtualization features: 虛擬: VT-x Caches (sum of all): L1d: 192 KiB (6 instances) L1i: 192 KiB (6 instances) L2: 1.5 MiB (6 instances) L3: 12 MiB (1 instance) NUMA: NUMA 節點: 1 NUMA node0 CPU(s): 0-11 Vulnerabilities: Gather data sampling: Mitigation; Microcode Itlb multihit: KVM: Mitigation: VMX disabled L1tf: Mitigation; PTE Inversion; VMX conditional cache flushes, SMT vulnerable Mds: Mitigation; Clear CPU buffers; SMT vulnerable Meltdown: Mitigation; PTI Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable Reg file data sampling: Not affected Retbleed: Mitigation; IBRS Spec rstack overflow: Not affected Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization Spectre v2: Mitigation; IBRS; IBPB conditional; STIBP conditional; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected Srbds: Mitigation; Microcode Tsx async abort: Mitigation; TSX disabled (bitnet-cpp) willy@linux2025:~/Desktop/linux2025$ gcc --version gcc (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0 Copyright (C) 2023 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. (bitnet-cpp) willy@linux2025:~/Desktop/linux2025$ conda info active environment : bitnet-cpp active env location : /home/willy/miniconda3/envs/bitnet-cpp shell level : 2 user config file : /home/willy/.condarc populated config files : /home/willy/miniconda3/.condarc conda version : 25.1.1 conda-build version : not installed python version : 3.12.9.final.0 solver : libmamba (default) virtual packages : __archspec=1=skylake __conda=25.1.1=0 __cuda=12.4=0 __glibc=2.39=0 __linux=6.11.0=0 __unix=0=0 base environment : /home/willy/miniconda3 (writable) conda av data dir : /home/willy/miniconda3/etc/conda conda av metadata url : None channel URLs : https://repo.anaconda.com/pkgs/main/linux-64 https://repo.anaconda.com/pkgs/main/noarch https://repo.anaconda.com/pkgs/r/linux-64 https://repo.anaconda.com/pkgs/r/noarch package cache : /home/willy/miniconda3/pkgs /home/willy/.conda/pkgs envs directories : /home/willy/miniconda3/envs /home/willy/.conda/envs platform : linux-64 user-agent : conda/25.1.1 requests/2.32.3 CPython/3.12.9 Linux/6.11.0-26-generic ubuntu/24.04.2 glibc/2.39 solver/libmamba conda-libmamba-solver/25.1.1 libmambapy/2.0.5 aau/0.5.0 c/. s/. e/. UID:GID : 1000:1000 netrc file : None offline mode : False ``` ::: 參考其他人的實驗: - [alanhc](https://hackmd.io/@alanhc/2025q1-term-project) - [Denny0097](https://hackmd.io/dXidLYjqT3SLwAyY2LSRpA?view) ### [microsoft/BitNet](https://github.com/microsoft/BitNet) 實驗方向 - [ ] 也許可以改變 chunk size 進行記憶體、 cache miss 和時間的分析 - [x] Huge Page - [x] CPU Affinity - [ ] 排程 - [ ] NUMA 調整 #### perf(kernel level) - perf record - https://gist.github.com/willy-liu/8d1708483690ab71f5048f8c68e7a59e 分析程式的熱點函式 | 佔比(%) | 函式名稱 | 所屬模組 | 推測用途與說明 | | ------ | ---------------------------------- | ------------- | ------------------------------------------------------------ | | 54.33% | `ggml_vec_dot_i2_i8_s` | `libggml.so` | **量化矩陣乘法核心**:針對 i2/i8 格式做向量內積運算,加速推論;通常為低精度 INT2/INT8 的權重乘法。 | | 36.03% | `ggml_vec_dot_f16` | `libggml.so` | **向量內積(半精度)**:FP16 格式的內積操作,是未量化模型的基礎乘法核心。 | | 26.41% | `_start` → `main` → `llama_decode` | `llama-cli` | **主要執行流程**:程式從主函式 `main` 開始,呼叫 LLaMA 解碼(推論)流程。 | | 25.92% | `llama_decode_internal` | `libllama.so` | **核心推論邏輯**:負責 LLaMA 模型解碼,包括 token-by-token 的預測計算。 | | 17.98% | `ggml_graph_compute_thread` | `libggml.so` | **背景執行緒**:處理計算圖執行的多執行緒核心。 | | 1.74% | `ggml_vec_dot_i2_i8_s`(其他執行緒) | `libggml.so` | 同上,分佈在其他 thread 中。 | | 0.75% | `ggml_barrier` | `libggml.so` | **執行緒同步點**:等待所有執行緒完成,類似 barrier 的功能。 | | 0.74% | `quantize_row_i8_s` | `libggml.so` | **量化處理**:將浮點資料轉換為 INT8 表示,通常在加速部署時使用。 | 火焰圖 ![flamegraph](https://hackmd.io/_uploads/SJGbVZpEee.svg) 函式用途分類 | 分類 | 函式(代表) | 說明 | | ----------- | ------------------------------------------------------------------------------------------ | -------------------------------------- | | **矩陣/內積計算** | `ggml_vec_dot_i2_i8_s`, `ggml_vec_dot_f16` | 是整個模型計算的最核心部分(矩陣乘法的 dot product)。 | | **矩陣乘法流程** | `ggml_compute_forward_mul_mat`, `ggml_compute_forward_mul_mat_one_chunk` | 表示計算圖中 forward pass 的一部分,針對不同區塊處理矩陣乘法。 | | **模型執行主體** | `llama_decode_internal`, `llama_graph_compute` | 控制推論 token 的運作流程、計算圖排程等。 | | **排程與執行緒** | `ggml_graph_compute_thread`, `ggml_graph_compute_secondary_thread`, `ggml_backend_sched_*` | 代表 GGML 用來安排 CPU 多執行緒運算的排程與背景執行邏輯。 | | **同步與量化** | `ggml_barrier`, `quantize_row_i8_s` | 在前後處理中同步各 thread,或處理資料格式轉換。 | - perf stat - 預設執行 https://gist.github.com/willy-liu/3bbee1d5492d02b919e7445b59366a44 - 開啟 huge page https://gist.github.com/willy-liu/409f33add5bb65db3dce68d28cfd74b0 - huge page + CPU affinity https://gist.github.com/willy-liu/626c8a19fc4e82754ef8181e693b1d97 整體效能指標 | 項目 | 數值 | 解釋與分析 | | ---------------------- | ----------------------- | -------------------------------------------------------------- | | `cycles` | **337,403,155,184** | CPU 時鐘週期數,越多表示執行時間越長。 | | `instructions` | **375,024,850,451** | 指令數量,代表總共執行的 CPU 指令。 | | `insn per cycle (IPC)` | **1.11** | 每 cycle 執行 1.11 條指令,<br>理想值越高越好,通常 > 1 表示 CPU 沒有完全被 memory 拖慢。 | | `time elapsed` | **16.46 秒**(wall clock) | 真實執行時間。注意和下方 `user`/`sys` 是不同角度。 | | `user` time | **84.49 秒** | 表示使用者空間總 CPU 使用時間,<br>多執行緒程式會大於實際時間(共用 CPU 核心)。 | | `sys` time | **0.27 秒** | 核心空間(系統呼叫)耗費的時間,低代表系統開銷小。 | 快取 | 項目 | 數值 | 解釋與分析 | | -------------------- | ---------- | ------------------------------------------------------ | | `cache-references` | 6.13B | Cache 存取次數。高代表 CPU 積極使用快取。 | | `cache-misses` | 1.68B | Cache miss 數量,越高代表從 RAM 讀取次數多。 | | **cache miss ratio** | **27.37%** | 高達 27% → **快取命中率差,效能可能受限於記憶體存取延遲**。可能是模型參數太大或資料存取模式不佳。 | TLB(Translation Lookaside Buffer) | 項目 | 數值 | 解釋 | | ------------------ | ------------- | ---------------------------------------------------- | | `dTLB-loads` | 155B | data TLB 存取次數 | | `dTLB-load-misses` | 19M(0.01%) | 幾乎沒有 miss,代表資料分頁效率很好 | | `iTLB-loads` | 7.59M | instruction TLB 存取次數 | | `iTLB-load-misses` | 1.05M(13.89%) | **iTLB miss 比例偏高** → 程式碼/指令分布廣、跳躍多,可能影響效能(常見於大型模型推論) | 效能比較 | 項目 | ❌ 預設 | ✅ 開啟 THP | ✅ THP + 綁定核心<br>`taskset -c 0-5` | | ------------------ | --------------- | ---------- | -------------------------------- | | `time elapsed`(秒) | 16.46 | 15.64 | 15.53 | | `instructions` | 375.02B | 373.47B | 373.01B | | `cycles` | 337.4B | 324.7B | 320.1B | | `IPC` | 1.11 | 1.15 | 1.17 | | `cache-miss ratio` | 27.37% | 27.86% | 28.17% | | `iTLB-load-misses` | 1.05M | **0.64M** | 0.97M | | `iTLB miss rate` | 13.89% | **13.30%** | **18.84%** 🔺 | #### uftrace 使用uftrace前,要重新 make 專案,並加入 - cmake - -pg - 插入 mcount() 呼叫點,讓 uftrace 能靜態追蹤每個函式的呼叫與耗時。 - uftrace 預設是靠這個符號來追蹤函式層級的動態呼叫圖。 - -g - 加入 debug 符號(如函式名稱、原始碼行號),讓 uftrace report 顯示人類可讀的函式名,而不是地址。 - 若省略會導致 report 顯示不出原始碼函式。 - fno-omit-frame-pointer - 保留堆疊追蹤資訊,對 uftrace、perf、gdb 都有幫助。 - 預設編譯器會優化掉 frame pointer,這會讓函式追蹤和火焰圖解析不準。 - -O0(選擇性) - 關閉編譯最佳化,讓追蹤結果更接近原始碼結構。 - 缺點是效能變慢,不建議在大型推論中長時間使用,但可用於小範例測試。 修改`./setup_env.py` ```diff - run_command(["cmake", "-B", "build", *COMPILER_EXTRA_ARGS[arch], *OS_EXTRA_ARGS.get(platform.system(), []), "-DCMAKE_C_COMPILER=clang", "-DCMAKE_CXX_COMPILER=clang++"], log_step="generate_build_files") + run_command([ + "cmake", + "-B", "build", + *COMPILER_EXTRA_ARGS[arch], + *OS_EXTRA_ARGS.get(platform.system(), []), + "-DCMAKE_C_COMPILER=clang", + "-DCMAKE_CXX_COMPILER=clang++", + "-DCMAKE_BUILD_TYPE=RelWithDebInfo", + "-DCMAKE_C_FLAGS=-pg -fno-omit-frame-pointer", + "-DCMAKE_CXX_FLAGS=-pg -fno-omit-frame-pointer", + ], log_step="generate_build_files") ``` 再重新編譯 `python setup_env.py -md models/BitNet-b1.58-2B-4T -q i2_s` 執行結果 - https://gist.github.com/willy-liu/66e9c3f60170ffd657abcdfea2f20d46 熱點分析 | # | Function 名稱 | 呼叫次數 | 總時間 | 主要功能說明 | | -: | ---------------------------------------- | ------: | ------: | ---------------------------- | | 1 | `ggml_vec_dot_i2_i8_s` | \~37M | 40.35s | INT2xINT8 量化矩陣乘法 dot product | | 2 | `ggml_vec_dot_f16` | \~11.7M | 18.87s | float16 dot product,應為主推理路徑 | | 3 | `ggml_compute_forward_mul_mat_one_chunk` | \~715K | 6.46s | 單一 chunk 的矩陣乘法計算 | | 4 | `ggml_barrier` | \~424K | 20.05s | 執行緒同步,thread barrier | | 5 | `ggml_graph_compute_thread` | 306 次 | 85.57ms | 主要 graph 計算執行緒 | | 6 | `ggml_graph_compute_secondary_thread` | 10 次 | 119μs | 次要執行緒調度 | | 7 | `ggml_thread_cpu_relax` | \~219M | 23.36s | idle 等待同步,表示有大量同步等待時間 | | 8 | `llama_decode` | 102 次 | 41.93s | 單一 token 的推理過程 | | 9 | `llama_decode_internal` | 51 次 | 20.97s | `llama_decode` 的核心實作 | | 10 | `std::map::emplace` 等 STL | - | 10s+ | token 對應、編碼解碼查表等 | 主要耗時在: - 向量運算與矩陣乘法核心 - ggml_vec_dot_i2_i8_s, ggml_vec_dot_f16, ggml_compute_forward_mul_mat_one_chunk - 大量執行緒同步 - ggml_barrier, ggml_thread_cpu_relax 表示 CPU threads 在等待任務分配與同步 - token 解碼 - llama_decode, llama_token_to_piece 系列是處理 token 字串的轉換與生成 #### 嘗試修改chunk_size,觀察ggml_compute_forward_mul_mat的cache和時間變化 [tiled matrix mul筆記](https://hackmd.io/@willy-liu/ryh9rcJree)

    Import from clipboard

    Paste your markdown or webpage here...

    Advanced permission required

    Your current role can only read. Ask the system administrator to acquire write and comment permission.

    This team is disabled

    Sorry, this team is disabled. You can't edit this note.

    This note is locked

    Sorry, only owner can edit this note.

    Reach the limit

    Sorry, you've reached the max length this note can be.
    Please reduce the content or divide it to more notes, thank you!

    Import from Gist

    Import from Snippet

    or

    Export to Snippet

    Are you sure?

    Do you really want to delete this note?
    All users will lose their connection.

    Create a note from template

    Create a note from template

    Oops...
    This template has been removed or transferred.
    Upgrade
    All
    • All
    • Team
    No template.

    Create a template

    Upgrade

    Delete template

    Do you really want to delete this template?
    Turn this template into a regular note and keep its content, versions, and comments.

    This page need refresh

    You have an incompatible client version.
    Refresh to update.
    New version available!
    See releases notes here
    Refresh to enjoy new features.
    Your user state has changed.
    Refresh to load new user state.

    Sign in

    Forgot password

    or

    By clicking below, you agree to our terms of service.

    Sign in via Facebook Sign in via Twitter Sign in via GitHub Sign in via Dropbox Sign in with Wallet
    Wallet ( )
    Connect another wallet

    New to HackMD? Sign up

    Help

    • English
    • 中文
    • Français
    • Deutsch
    • 日本語
    • Español
    • Català
    • Ελληνικά
    • Português
    • italiano
    • Türkçe
    • Русский
    • Nederlands
    • hrvatski jezik
    • język polski
    • Українська
    • हिन्दी
    • svenska
    • Esperanto
    • dansk

    Documents

    Help & Tutorial

    How to use Book mode

    Slide Example

    API Docs

    Edit in VSCode

    Install browser extension

    Contacts

    Feedback

    Discord

    Send us email

    Resources

    Releases

    Pricing

    Blog

    Policy

    Terms

    Privacy

    Cheatsheet

    Syntax Example Reference
    # Header Header 基本排版
    - Unordered List
    • Unordered List
    1. Ordered List
    1. Ordered List
    - [ ] Todo List
    • Todo List
    > Blockquote
    Blockquote
    **Bold font** Bold font
    *Italics font* Italics font
    ~~Strikethrough~~ Strikethrough
    19^th^ 19th
    H~2~O H2O
    ++Inserted text++ Inserted text
    ==Marked text== Marked text
    [link text](https:// "title") Link
    ![image alt](https:// "title") Image
    `Code` Code 在筆記中貼入程式碼
    ```javascript
    var i = 0;
    ```
    var i = 0;
    :smile: :smile: Emoji list
    {%youtube youtube_id %} Externals
    $L^aT_eX$ LaTeX
    :::info
    This is a alert area.
    :::

    This is a alert area.

    Versions and GitHub Sync
    Get Full History Access

    • Edit version name
    • Delete

    revision author avatar     named on  

    More Less

    Note content is identical to the latest version.
    Compare
      Choose a version
      No search result
      Version not found
    Sign in to link this note to GitHub
    Learn more
    This note is not linked with GitHub
     

    Feedback

    Submission failed, please try again

    Thanks for your support.

    On a scale of 0-10, how likely is it that you would recommend HackMD to your friends, family or business associates?

    Please give us some advice and help us improve HackMD.

     

    Thanks for your feedback

    Remove version name

    Do you want to remove this version name and description?

    Transfer ownership

    Transfer to
      Warning: is a public team. If you transfer note to this team, everyone on the web can find and read this note.

        Link with GitHub

        Please authorize HackMD on GitHub
        • Please sign in to GitHub and install the HackMD app on your GitHub repo.
        • HackMD links with GitHub through a GitHub App. You can choose which repo to install our App.
        Learn more  Sign in to GitHub

        Push the note to GitHub Push to GitHub Pull a file from GitHub

          Authorize again
         

        Choose which file to push to

        Select repo
        Refresh Authorize more repos
        Select branch
        Select file
        Select branch
        Choose version(s) to push
        • Save a new version and push
        • Choose from existing versions
        Include title and tags
        Available push count

        Pull from GitHub

         
        File from GitHub
        File from HackMD

        GitHub Link Settings

        File linked

        Linked by
        File path
        Last synced branch
        Available push count

        Danger Zone

        Unlink
        You will no longer receive notification when GitHub file changes after unlink.

        Syncing

        Push failed

        Push successfully