- Added 2 new blockschedulers, a greedy algorithm and an "optimal" ILP that
[libfirm] / ir / be / bespillremat.c
2006-09-12 Adam Szalkowski* fixed issue with exceeded register pressure at bb end
2006-09-06 Adam Szalkowskiminor changes
2006-09-05 Adam Szalkowskiimproved repairing of schedule
2006-09-04 Adam Szalkowskidumping pressure graph and pressure stats only if requested
2006-09-04 Christian Würdigfixed profile support
2006-09-04 Christian Würdigfixed C99 style
2006-09-03 Adam Szalkowskifixed issue when original definition of a node is removed
2006-09-02 Adam Szalkowskifixed a bug
2006-09-01 Adam Szalkowskihandling barrier stuff
2006-08-31 Adam Szalkowskinow hopefully computing number of available registers...
2006-08-30 Adam Szalkowskihopefully fixed registerhypertony at bb end
2006-08-28 Adam Szalkowskikeine Ahnung
2006-08-26 Adam Szalkowskiperform_memory operand now takes a spill instead of...
2006-08-25 Adam Szalkowskischeint jetzt etwas gammlich zu sein :(
2006-08-24 Adam Szalkowskispill slots are handled by matze
2006-08-23 Adam Szalkowskischeint nun fluffig zu sein
2006-08-21 Adam Szalkowskibugs weggeflufft
2006-08-17 Adam Szalkowskiminor fixes
2006-08-10 Adam Szalkowskianother bug fixed
2006-08-10 Adam Szalkowskifixed bugs
2006-08-09 Matthias Braun- Don't perform memory operands merging too early
2006-08-09 Adam Szalkowskiusing lc_opts
2006-08-09 Sebastian HackSeveral changes
2006-08-09 Sebastian HackFixed some typos
2006-08-09 Adam Szalkowskiadded default values for vars
2006-08-09 Sebastian HackFixed a bug
2006-08-04 Matthias Braunwarning fix
2006-08-04 Matthias BraunRewrote spillslot handling:
2006-08-04 Adam Szalkowskichanged output format a little bit
2006-08-03 Adam Szalkowskioption added for only using remats á la Briggs
2006-08-01 Adam Szalkowskifixed consts
2006-08-01 Christian Würdigadded semicolon after label
2006-08-01 Adam Szalkowskinow supporting memory operands
2006-07-31 Adam Szalkowskiadapted to new liveness
2006-07-31 Sebastian HackAdapted to new liveness
2006-07-29 Adam Szalkowskifixed a bug
2006-07-17 Adam Szalkowskifurther bug fixes
2006-07-15 Adam Szalkowskiminor changes
2006-07-11 Adam Szalkowskionly remat if one operand dies
2006-07-11 Sebastian HackAdapted to changes in exec_freq.h
2006-07-03 Adam Szalkowskidumping execfreqs in pressure graph
2006-06-28 Adam Szalkowskigammlich
2006-06-28 Adam Szalkowskihm, I'm quite happy with this
2006-06-27 Adam Szalkowskivalues may die at every use
2006-06-19 Adam Szalkowskiseems to work again and yields better objvals
2006-06-18 Adam Szalkowskiseems to spill again a bit
2006-06-17 Adam Szalkowskimemcopies/interference stuff now in ILP
2006-06-15 Adam Szalkowskihopefully fixed a bug which caused too high register...
2006-06-12 Adam SzalkowskiSpilling Unknowns is cheap now
2006-06-12 Christian Würdigspill_env creation changed
2006-06-09 Adam Szalkowskipressure stats
2006-06-06 Adam Szalkowskiseems to work now (still spill slot splitting to be...
2006-06-06 Adam Szalkowskigoodwin gammlich
2006-06-05 Adam Szalkowskicomputing execfreq from bb, just to be sure
2006-06-05 Adam Szalkowskiwrong irn number in ilp variable name
2006-06-04 Adam Szalkowskinot functional goodwin reduction
2006-05-31 Adam Szalkowskiusing execfreq as default
2006-05-31 Sebastian HackMade it compile under vc2005
2006-05-31 Adam Szalkowskidumping inverse remats in cyan color
2006-05-30 Adam Szalkowskiuse loop depth for computing execution frequencies...
2006-05-30 Adam Szalkowskicommitting ilp based spilling