Changes between Version 16 and Version 17 of Status/SLPJ-Tickets


Ignore:
Timestamp:
Dec 3, 2008 1:00:26 PM (5 years ago)
Author:
simonpj
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • Status/SLPJ-Tickets

    v16 v17  
    1818 * #2078: INLINing improvement; ask Christian.Maeder@dfki.de to see if it improves CASL 
    1919 
     20== New code generator == 
     21 * #2253: NCG could do better. Look at this when John D’s ncg is working 
     22 * #2838: trouble with `CmmInt` masking; watch out for this 
     23 * #783: SRTs getting big        
     24 * #2289: cheap check at start of case alternatives      
     25 
    2026=== Other performance === 
    2127 * #2670: record selectors behaving badly wrt optimisation 
    2228 * #2731: avoiding unnecessary evaluation when unpacking constructors 
    23  * #2253: NCG could do better. Look at this when John D’s ncg is working 
     29 
    2430 * #1434: Slow conversion Double to Int  
    2531 * #2439: Strict dictionaries    
    2632 * #2132: Optimise nested comparisons: if you know x>0 then you know x>=1 etc.  Maybe a special pass that knows about arithmetic?        
    27  * #783: SRTs getting big        
    2833 * #1216: array indexing, inlining/arity bug     
    2934 * #149: float-out/CSE   
    30  * #2289, 2387:(a) cheap check at start of case alternatives, (b) nested CPR analysis    
     35 * #2289, 2387: nested CPR analysis      
    3136 * #2092: Possible quadratic-sized Eq instances. Does it really go quadratic, or does the join-point inlining machinery prevent it?  Still to check: delicacy wrt case-of-case 
    3237 * #2255, #2643: Improve '''`SpecConstr`''' for free variables, and for join points.