Changes between Version 15 and Version 16 of DataParallel/Dec2010Release


Ignore:
Timestamp:
Dec 14, 2010 1:35:45 AM (5 years ago)
Author:
chak
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • DataParallel/Dec2010Release

    v15 v16  
    2121
    2222Changes:
    23  * -fdph-par should be the default (and sensible error message if the dph package is not available) '''[MANUEL]'''
    24  * -Odph should be equivalent to '-O2 -fsimplifier-phases=3 -fsimplifier-iterations=20' '''[MANUEL]'''
     23 * -fdph-par should be the default (and sensible error message if the dph package is not available) (done, but need to still sort out a build problem) '''[MANUEL]'''
     24 * -Odph should be equivalent to '-O2 -fsimplifier-phases=3 -fsimplifier-iterations=20' (done, not pushed yet) '''[MANUEL]'''
    2525 * Move GHC.PArr into  the DPH libs.  (Needed for Haddock.) '''[MANUEL]'''
    2626 * Find out if we still need the `NoSpecConstr` annotation and remove it if not '''[ROMAN]'''
     
    5151  * !SumSquares [FINE]
    5252  * Dot product [FINE]
    53   * Evens [OK (but more than 3 times slower than C)]
     53  * Evens [OK (but more than 3 times slower than C; any improvement since #4830 was fixed?)]
    5454    * rl reckons this is due to GHC compiling modulo of powers of two inefficiently; c.f., #3065 (in `packByTags`)
    5555  * SMVM (blocked on optimisation of lifted indexing) '''[BROKEN]''' '''[BEN & ROMAN]'''
    5656 * Dynamically-nested DPH programs without user-defined datatypes should run correctly, but not necessarily fast
    5757  * Quicksort '''[BROKEN (SpecConstr) & SLOW]''' '''[SIMON & BEN]'''
    58   * Quickhull '''[OK, but has  a SpecConstr problem that we want to fix]''' '''[ROMAN]'''
     58  * Quickhull '''[OK, but has a SpecConstr problem that we want to fix]''' '''[ROMAN]'''
    5959    * Probably affected by the same optimisation issue with the compilation of modulo operations as Evens
    60     * #4830 (fix incoming)
     60    * #4830: this fix wasn't sufficient, still doesn't optimise properly
    6161 * Dynamically-nested DPH programs with user-defined datatypes should run correctly, but not necessarily fast
    62   * Words '''[BROKEN (SpecConstr)]''' '''[ROMAN & SIMON]'''
     62  * Words '''[BROKEN (SpecConstr when using `-dph-seq`)]''' '''[ROMAN & SIMON]'''
    6363    * #4831
    6464  * !BarnesHut [OK]