劉溢茗
    • Create new note
    • Create a note from template
      • 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
        • Only me
        • Signed-in users
        • Everyone
        Only me Signed-in users Everyone
      • Write
        • Only me
        • Signed-in users
        • Everyone
        Only me 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
    • Save as template
    • 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 Create Help
Create Create new note Create a note from template
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
Only me
  • Only me
  • Signed-in users
  • Everyone
Only me Signed-in users Everyone
Write
Only me
  • Only me
  • Signed-in users
  • Everyone
Only me 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
    Subscribed
    • Any changes
      Be notified of any changes
    • Mention me
      Be notified of mention me
    • Unsubscribe
    Subscribe
    # 2019q1 Homework7 (skiplist) contributed by < `LiunuXXX` , `jeffcarl67`> ## 預期目標 1. 思考 Linux 核心內部的資料結構,特別是 cache-oblivious data structures 的考量 2. 實作 Skip list,設計對應的效能分析框架 3. 開發適用於使用者層級和核心層級的程式碼 ## 開發環境 <`jeffcarl67`> ``` $ lscpu 架構: x86_64 CPU 作業模式:  32-bit, 64-bit Byte Order: Little Endian Address sizes: 39 bits physical, 48 bits virtual CPU(s): 4 On-line CPU(s) list: 0-3 每核心執行緒數:   2 每通訊端核心數:    2 Socket(s): 1 NUMA 節點: 1 供應商識別號: GenuineIntel CPU 家族: 6 型號: 60 Model name: Intel(R) Core(TM) i5-4210M CPU @ 2.60GHz 製程: 3 CPU MHz: 859.812 CPU max MHz: 3200.0000 CPU min MHz: 800.0000 BogoMIPS: 5190.32 虛擬: VT-x L1d 快取: 32K L1i 快取: 32K L2 快取: 256K L3 快取: 3072K NUMA node0 CPU(s): 0-3 ``` ``` $ uname -r 4.19.45-1-MANJARO ``` ``` $ gcc --version gcc (GCC) 8.3.0 ``` ## 作業要求 * 完成 [第 11 週測驗題 (中)](https://hackmd.io/s/BkFJPHriE) 和所有延伸題目 * 在 Linux 核心原始程式碼使用 skip list 的案例,介紹其原理,設計 Linux 核心模組的實驗 * 需要涵蓋 kernel API 同步機制的運用 * 執行時期的分析 (提示: 可善用 eBPF) ## 作業區 ==完成 [第 11 週測驗題 (中)](https://hackmd.io/s/BkFJPHriE) 和所有延伸題目== ### 測驗 `1` 給定一個 circular linked list 實作如下: (檔案 `list.h`) ```cpp #ifndef INTERNAL_LIST_H #define INTERNAL_LIST_H /* circular doubly-linked list */ typedef struct __list_t { struct __list_t *prev, *next; } list_t; /* * Initialize a list to empty. Because these are circular lists, an "empty" * list is an entry where both links point to itself. This makes insertion * and removal simpler because they do not need any branches. */ static inline void list_init(list_t *list) { list->prev = list; list->next = list; } /* * Append the provided entry to the end of the list. This assumes the entry * is not in a list already because it overwrites the linked list pointers. */ static inline void list_push(list_t *list, list_t *entry) { list_t *prev = list->prev; entry->prev = prev; entry->next = list; prev->next = entry; list->prev = entry; } /* * Remove the provided entry from whichever list it is currently in. This * assumes that the entry is in a list. You do not need to provide the list * because the lists are circular, so the list's pointers will automatically * be updated if the first or last entries are removed. */ static inline void list_remove(list_t *entry) { list_t *prev = entry->prev; list_t *next = entry->next; prev->next = next; next->prev = prev; } /* * Remove and return the first entry in the list or NULL if the list is empty. */ static inline list_t *list_pop(list_t *list) { list_t *foo = list->prev; if (foo == list) return NULL; LL1 } #endif ``` 請參照上方程式註解,補完對應程式碼。 LL1 = ? --- 首先我們看看`list_pop`的要求 ```clike /* * Remove and return the first entry in the list or NULL if the list is empty. */ ``` 可知我們需要用到 `list_remove` 將 list 的 first entry 移除後並且回傳,若 list 為空就回傳 NULL 接著看 `list_remove` 的實作 ```clike /* * Remove the provided entry from whichever list it is currently in. This * assumes that the entry is in a list. You do not need to provide the list * because the lists are circular, so the list's pointers will automatically * be updated if the first or last entries are removed. */ static inline void list_remove(list_t *entry) { list_t *prev = entry->prev; list_t *next = entry->next; prev->next = next; next->prev = prev; } ``` * prev 記錄 entry 的前一個 node ```clike list_t *prev = entry->prev; ``` ![](https://i.imgur.com/IRbbWOR.jpg) --- * next 記錄 entry 的後一個 node ```clike list_t *next = entry->next; ``` ![](https://i.imgur.com/SEhNd5t.jpg) --- * prev 的 next 指標指向 next ```clike prev->next = next; ``` ![](https://i.imgur.com/peLAuI0.jpg) --- ```clike next->prev = prev; ```` * next 的 prev 指標指向 prev ![](https://i.imgur.com/wooGjJR.jpg) * 最終狀態如下圖 ![](https://i.imgur.com/9Me2JPT.jpg) --- 以上完成了 `list_remove` 的一次操作,可以注意到此方法僅改變前後指標, entry 並未被移除,其 prev 及 next 指標並未變動。 接下來回到 `list_pop` 中的第一行 ```clike list_t *foo = list->prev; ``` * foo 記錄 list 的前一 node ,此 node 即為 list 的 first entry ![](https://i.imgur.com/wgO8lMA.jpg) --- 因此我們用 `list_remove` 將 foo 移除,再將其回傳即可,如下圖 ![](https://i.imgur.com/d0Hi20l.jpg) 故答案選擇 `(e)` list_remove(foo); return foo; --- ## 延伸問題: ### 1. 研讀 2019q3 Week11 的 cache-oblivious data structures 參考資料,記錄所見所聞: --- [Maximize Cache Performance with this One Weird Trick: An Introduction to Cache-Oblivious Data Structures](https://rcoh.me/posts/cache-oblivious-datastructures/) 文章中提到,Postgres 利用固定大小的 blocks ,又稱為 "pages" 實作 B-Tree,其目的是為了與 OS 中的 page sizes 做對應,而 OS 又對應到硬體。 然而,上述做法實作於不同硬體時,必須針對其挑選一最適合的 page size 以達到較好的效能。 因此本篇的主題便是想設計不被 underlying cache size 影響的資料結構及演算法,如此我們不需自己調整 cache size,程式碼便會最佳化的利用 cache layers,擁有此性質的資料結構及演算法,我們稱之為 "cache-size oblivious"。 首先,我們來看一張 BST 的效能比較圖 ![](https://rcoh.me/images/cache-oblivious-graph.png) 上圖中我們可以看到三種不同策略的 BST 走訪: 1. blue line : level by level 的走訪,又稱為 "bread first traversal". 2. orange line : (preorder) "depth first traversal". 3. green line : 利用 "recusive blocking" 法實作上述的 cache obivious 概念. 由上圖我們發現,在 elements in tree 達到 1600 萬的時候,"recursive blocking" 法的效率為另外兩種方法的近兩倍之外。 接下來我們探討如何實作 green line 的方法,以及其如何效能遠高於其他兩種方法的原因。 #### A new scheme for analyzing algorithms 一般分析演算法效能的 Big-$O$ notation,是假設讀取成本與記憶體中的 byte 成正比,然而真實世界的運作方式是以 block 為單位,即便你只想讀少量 bytes 的資料,還是得付出其所在整個 block 的成本,反之,若你想從同一個 block 讀取更多的 bytes,幾乎不用負擔任何額外成本。 因此 cache 利用率較好的演算法應該有以下特質: When we read a block of data, we make use of every byte in that block. 然而使用一般演算法分析無法看出 "cache-oblivious" 的效益,因此我們介紹一種較接近真實世界運作 (work in blocks) 的演算法分析模型。 探討這種不受 cache size 影響又能最佳化運作的演算法時,我們假設 memory 有兩個區塊,small fast layer 及 big slow layer,其擁有以下特性: 1. memory 只可以從 slow layer 移動到 fast layer,且單位為固定大小的 block, $B$ 2. slow layer 的大小為無限大,而 fast layer 的大小為 $M$ 3. 我們分析演算法時以 $M$ 為 cache 的單位,$B$ 為 block 的單位 4. 分析時一單位的工作量為從 slow memory 載入 a block 至 fast memory,其他皆視為極快速的工作,不納入考量 回到前面那張 BST 效能比較圖 ![](https://rcoh.me/images/cache-oblivious-graph.png) 我們可以發現,在 Elements in Tree($N$)很小時,效能幾乎沒有差異,那是因為所有樹都在同一 block 內,讀取成本相差無幾,此外,blue and orange lines 中平坦的部分發生在特定的 internal cache sizes,而當$N$越大,cache-oblivious 的優勢就越大。 接著我們看看上圖三種方法的運作方式: #### Blue line : By Level (Breadth) ![](https://rcoh.me/images/breadthfirst.svg) * 其在 memory 內的觀點是以上圖的橘框為單位,elements 內的數字即為其在 array 內的位置 * 這種方法有個明顯的缺點,當 level 夠大時,每個 level 就代表一個 block,不但前面低階層的 block 利用率極低,如第一個 block 只存了 1 個 element ,且從樹根走訪至樹葉時,每經過一 level 就相當於讀取不同的 block,導致讀取成本高達 $\log_2N$ #### Orange line : Preorder (Depth First) ![](https://rcoh.me/images/depthfirst.svg) * 這種方法在讀取至下一個 block 之前,會持續往下走訪,在上圖的橘框內作樹根至樹葉走訪時,都在同一個 block 內讀取,為最佳情況。 * 最靠右下的 block 只存放一個 element,若是做最右邊的走訪,讀取成本也是 $\log_2N$ * 此方法僅比 BFS 好一點 #### Green line : Recursive Block Approach ![](https://rcoh.me/images/van-layout.svg) * 利用 divide and conquer strategy 實現 cache-oblivious data structures and algorithms * 從樹高 $(logN)$ 一半處 split ,則上半部有 $2^{\log_2(N)/2}=\sqrt{N}$ 個 nodes,而其 child nodes 也有大約 $\sqrt{N}$ 個 nodes,如此遞迴 split,如上圖 * 每個綠框為一個 block,橘框為 superblock * blocks 間以及 superblocks 間皆是連續的 * 再讀取下一個 block 前,隨著讀取的 block size 變大,能走訪的 level 也越大 例如 : 讀取一 block of size $B$,可以走訪 $\log_2B$ 的 levels 樹有 $\log_2N$ 個 levels,則讀出的 blocks 數為 $\frac{\log_2N}{\log_2B}=\log_BN=F(N)$ ### 2. 依據之前給定的 Linux 核心 linked list 實作,開發 cache-oblivious 的版本,並證實其效益 --- 一般在設計 cache-oblivious 的 data structures and algorithms 時有三個常用策略: 1. van Emde Boas (vEB) 2. Weighted 平衡樹 3. Packed Memory Structure 而前面所利用的 Recursive Block Approach 便是採用 vEM 策略,如下圖 ![](https://www.itread01.com/uploads/images/20161007/1475818713-3600.jpg) 而設計 cache-oblivious linked list 常用的策略為 Packed Memory Structure 在 [Cache-Oblivious Algorithms and Data Structures](http://erikdemaine.org/papers/BRICS2002/paper.pdf) 中提到: The packed-memory structure maintains N elements consecutive in memory with gaps of size $O(1)$, subject to insertions and deletions in $O(\log2N)$ amortized time and $O((\log_2N)/B)$ amortized 接著我們介紹在 [Data Structures Libraries](http://www.lsi.upc.edu/~lfrias/thesis/thesis-lfrias.pdf?fbclid=IwAR0r3a8tTkC2QDVxR6cyEWN41JhZiQwdnQht8IJBJz3cGjVhfrr2aEYGqT8) 中與 packed-memory structure 相關的三種資料結構,其概念是在 linked list 配置 array 儲存資料,並將整個結構稱為 bucket ![](https://i.imgur.com/HDqAKp6.png) ( a ) Contiguous: 在 linked list 內配置的 array 作連續性配置 * 優點:在同一個 bucket 內的 array 存取資料時,因為容易在同一個 memory block 內,存取速度會相較於原本的 linked list 快 * 缺點:和 array 一樣,插入或刪除資料時必須付出移動資料的成本 ( b ) With gaps: bucket 內部 array 的元素間允許存在空隙 * 減少因插入刪除而移動資料的成本,較 Contiguous 好一些 * 每一個 element 必須耗費額外空間記錄此空間是否存資料 ( c ) Linked: 內部元素間由 link 連接 * 須耗費額外空間儲存 link * 插入或刪除時不須負擔額外移動資料的成本 * scalability 較佳 --- [github](https://github.com/jeffcarl67/unrolled_linked_list) 以下是 contiguous 的 unrolled list 實作: * Data Structure of unrolled lsit ```clike= struct listitem { struct list_head list; int i[CAPACITY]; int stored; int cap; }; ``` * Initialization ```clike= void init_listitem(struct listitem *item) { if (!item) return; item->cap = CAPACITY; item->stored = 0; } ``` 每個 listitem 沿用了 list.h 的 list_head 之外,另有一陣列儲存資料,stored 記錄佔用格數,cap 為每個 bucket element 格數,以下是幾個基本操作 (a) insertion ```clike int insert_unrolled(struct list_head *list, int item) { struct listitem *tmp = NULL; struct listitem *last = NULL; if (!list) return 0; last = list_entry(list->prev, struct listitem, list); if (list_empty(list) || last->stored >= (last->cap / 2 + 1)) { tmp = (struct listitem *)malloc(sizeof(struct listitem)); if (!tmp) { printf("malloc error\n"); return -1; } init_listitem(tmp); list_add_tail(&tmp->list, list); last = list_entry(list->prev, struct listitem, list); } last->i[last->stored] = item; last->stored++; return 0; } ``` (b) deletion ```clike= int delete_unrolled(struct list_head *list, int index) { struct listitem *entry = NULL; int i = 0; list_for_each_entry(entry, list, list) { if (i <= index && i + entry->stored > index) { memcpy(entry->i + index - i, entry->i + index - i + 1, sizeof(int) * (i + entry->stored - index) - 1); entry->stored--; } i += entry->stored; } return 0; } ``` 備註: [演算法筆記](http://www.csie.ntnu.edu.tw/~u91029/Data.html?fbclid=IwAR3dF04fF4gthJ_RuaC1jYpPr-TaO9ChuptqolrmBHUUiiQau6g3xn7yWnM)的資料結構中,假設 $N$ 筆資料,分成 $A$ 個 Bucket,每個 bucket 約 $B=N/A$ 個元素,而在刪除資料時,若是相鄰兩塊小於等於 $B$ 就要合成一塊。 本實作演算法尚未加上上述合併的功能。 ### 效能分析 --- ### 開發環境 <`LiunuXXX`> ``` $ lscpu Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Byte Order: Little Endian CPU(s): 4 On-line CPU(s) list: 0-3 Thread(s) per core: 2 Core(s) per socket: 2 Socket(s): 1 NUMA node(s): 1 Vendor ID: GenuineIntel CPU family: 6 Model: 58 Model name: Intel(R) Core(TM) i5-3230M CPU @ 2.60GHz Stepping: 9 CPU MHz: 2934.994 CPU max MHz: 3200.0000 CPU min MHz: 1200.0000 BogoMIPS: 5188.25 Virtualization: VT-x L1d cache: 32K L1i cache: 32K L2 cache: 256K L3 cache: 3072K NUMA node0 CPU(s): 0-3 ``` ``` $ uname -r 4.15.0-51-generic ``` ``` $ gcc --version gcc (Ubuntu 5.4.0-6ubuntu1~16.04.11) 5.4.0 20160609 ``` ``` $ gnuplot gcc (Ubuntu 5.4.0-6ubuntu1~16.04.11) 5.4.0 20160609 ``` ### Traversal * Traversal Test of list ```clike= void test_traverse(struct list_head *list, int num) { struct timeval start; struct timeval end; struct timeval result = {0, 0}; struct listitem *entry, *next, *item; int i; int ret; INIT_LIST_HEAD(list); for (i = 0; i < num; i++) { item = (struct listitem *)malloc(sizeof(struct listitem)); item->i = i; list_add_tail(&item->list, list); } system("echo 3 > /proc/sys/vm/drop_caches"); gettimeofday(&start, NULL); list_for_each_entry(entry, list, list) {} gettimeofday(&end, NULL); timersub(&end, &start, &result); printf("%d %lu %lu\n", num, result.tv_sec, result.tv_usec); list_for_each_entry_safe(entry, next, list, list) { list_del(&entry->list); free(entry); } ``` * Traversal Test of unrolled list (aware) ```clike void test_traverse(struct list_head *list, int num) { struct timeval start; struct timeval end; struct timeval result = {0, 0}; struct listitem *entry, *next; int i; int ret; INIT_LIST_HEAD(list); for (i = 0; i < num; i++) { ret = insert_unrolled(list, i); } system("echo 3 > /proc/sys/vm/drop_caches"); gettimeofday(&start, NULL); list_for_each_entry(entry, list, list) { i = 0; while (i < entry->stored) { i++; } } gettimeofday(&end, NULL); timersub(&end, &start, &result); printf("%d %lu %lu\n", num, result.tv_sec, result.tv_usec); list_for_each_entry_safe(entry, next, list, list) { list_del(&entry->list); free(entry); } ``` 利用 timeval 分析走訪 list 及 unrolled list (aware) 的時間,結果如下圖 ![](https://i.imgur.com/pyd2Iuy.png) 推測此處差異主要是在時 unrolled list 內的 array 使其在讀取時較有優勢,而原本的 list 在存取時較常跳到不同的 block 而產生 cache miss 。 ### Insertion * Insertion test of list ```clike void test_insert(struct list_head *list, int num) { struct timeval start; struct timeval end; struct timeval result = {0, 0}; struct listitem *entry, *next, *item; int i; int ret; INIT_LIST_HEAD(list); system("echo 3 > /proc/sys/vm/drop_caches"); gettimeofday(&start, NULL); for (i = 0; i < num; i++) { item = (struct listitem *)malloc(sizeof(struct listitem)); item->i = i; list_add_tail(&item->list, list); } gettimeofday(&end, NULL); timersub(&end, &start, &result); printf("%d %lu %lu\n", num, result.tv_sec, result.tv_usec); list_for_each_entry_safe(entry, next, list, list) { list_del(&entry->list); free(entry); } } ``` * Insertion test of unrolled list (aware) ```clike void test_insert(struct list_head *list, int num) { struct timeval start; struct timeval end; struct timeval result = {0, 0}; struct listitem *entry, *next; int i; int ret; INIT_LIST_HEAD(list); system("echo 3 > /proc/sys/vm/drop_caches"); gettimeofday(&start, NULL); for (i = 0; i < num; i++) { ret = insert_unrolled(list, i); } gettimeofday(&end, NULL); timersub(&end, &start, &result); printf("%d %lu %lu\n", num, result.tv_sec, result.tv_usec); list_for_each_entry_safe(entry, next, list, list) { list_del(&entry->list); free(entry); } } ``` Insertion 效能比較如下圖 ![](https://i.imgur.com/zjsdW5P.png) 類似 traversal,效能差異來自於 unrolled list 內的 array 於循序寫入時的效益。 ### 3. 實作 Skip list,並且比照之前的 unit test 提供對應的測試、效能分析,還有改善機制 --- ==在 Linux 核心原始程式碼使用 skip list 的案例,介紹其原理,設計 Linux 核心模組的實驗== * 需要涵蓋 kernel API 同步機制的運用 * 執行時期的分析 (提示: 可善用 eBPF) # References 1. [Maximize Cache Performance with this One Weird Trick: An Introduction to Cache-Oblivious Data Structures](https://rcoh.me/posts/cache-oblivious-datastructures/#fnref:assum) 2. [關於算法,那些你不知道的事](https://www.itread01.com/articles/1475818714.html?fbclid=IwAR39mapd0cR55YSU4rgLAPqOjR0L6gBE9JxbTzdIq2H8sOF44kbiFs3xL9s) 3. [Cache-Oblivious Algorithms and Data Structures](http://erikdemaine.org/papers/BRICS2002/paper.pdf)

    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