D11487: optimization of TextLineData::attribute

[Milian Wolff]

mwolff added a comment.


  @jtamate I just checked, the function is called with the same parameters for me locally. What output do you get for this:
  
    diff --git a/src/spellcheck/spellcheck.cpp b/src/spellcheck/spellcheck.cpp
    index 9e04d788..50e92885 100644
    --- a/src/spellcheck/spellcheck.cpp
    +++ b/src/spellcheck/spellcheck.cpp
    @@ -133,6 +133,7 @@ QList<QPair<KTextEditor::Range, QString> > KateSpellCheckManager::spellCheckWrtH
             bool singleLine,
             bool returnSingleRange)
     {
    +    qDebug() << range << dictionary << singleLine << returnSingleRange;
         QList<QPair<KTextEditor::Range, QString> > toReturn;
         if (range.isEmpty()) {
             return toReturn;
    @@ -172,6 +173,7 @@ QList<QPair<KTextEditor::Range, QString> > KateSpellCheckManager::spellCheckWrtH
                 }
                 const int start = (line == startLine) ? startColumn : 0;
                 const int end = (line == endLine) ? endColumn : kateTextLine->length();
    +            qDebug() << line << kateTextLine->attributesList().size() << (end - start);
                 for (int i = start; i < end;) { // WARNING: 'i' has to be incremented manually!
                     int attr = kateTextLine->attribute(i);
                     const KatePrefixStore &prefixStore = highlighting->getCharacterEncodingsPrefixStore(attr);
  
  For me this is the interesting bit:
  
    15.097 debug: unknown[unknown:0]: [ (0, 0)  ->  (0, 358720) ] "" true false
    15.097 debug: unknown[unknown:0]: 0 1979 358720
    23.013 debug: unknown[unknown:0]: [ (5, 0)  ->  (5, 162682) ] "" true false
    23.013 debug: unknown[unknown:0]: 5 12 162682
  
  So for line 0 I should actually get quite abysmal performance too, worst case 1979 * 358720, but it flies right through here... Now I wonder what kind of CPU you are using? I have a i7-5600U CPU @ 2.60GHz here in my laptop, which has:
  
    $ lstopo --of console
    Machine (11GB)
      Package L#0 + L3 L#0 (4096KB)
        L2 L#0 (256KB) + L1d L#0 (32KB) + L1i L#0 (32KB) + Core L#0
          PU L#0 (P#0)
          PU L#1 (P#1)
        L2 L#1 (256KB) + L1d L#1 (32KB) + L1i L#1 (32KB) + Core L#1
          PU L#2 (P#2)
          PU L#3 (P#3)
  
  The 1979 entries of attributes (each 12byte large) produce ~24KB of data, which fits nicely within my L1 cache. Maybe I'm just lucky and this is what hides this issue from me?

REPOSITORY
  R39 KTextEditor

REVISION DETAIL
  https://phabricator.kde.org/D11487

To: jtamate, #frameworks, #kate, cullmann
Cc: anthonyfieroni, dhaumann, mwolff, cullmann, michaelh, kevinapavew, ngraham, demsking, sars
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20180323/a6df2c20/attachment.html>