[LLVMbugs] [Bug 1282] NEW: Extremly slow codegen

bugzilla-daemon at cs.uiuc.edu bugzilla-daemon at cs.uiuc.edu
Wed Mar 28 01:30:03 PDT 2007


http://llvm.org/bugs/show_bug.cgi?id=1282

           Summary: Extremly slow codegen
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Backend: X86
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: asl at math.spbu.ru


Consider attached file. I have:

 $ ./llc -time-passes sha512-opt.bc
===-------------------------------------------------------------------------===
                      ... Pass execution timing report ...
===-------------------------------------------------------------------------===
  Total Execution Time: 553.6418 seconds (554.1256 wall clock)

   ---User Time---   --System Time--   --User+System--   ---Wall Time---  ---
Name ---
  553.4408 ( 99.9%)   0.0899 ( 91.8%)  553.5308 ( 99.9%)  554.0134 ( 99.9%)  X86
DAG->DAG Instruction Selection
   0.0439 (  0.0%)   0.0000 (  0.0%)   0.0439 (  0.0%)   0.0466 (  0.0%)  Linear
Scan Register Allocator
   0.0169 (  0.0%)   0.0000 (  0.0%)   0.0169 (  0.0%)   0.0213 (  0.0%)  Live
Interval Analysis
   0.0099 (  0.0%)   0.0019 (  2.0%)   0.0119 (  0.0%)   0.0126 (  0.0%)  X86
AT&T-Style Assembly Printer
   0.0119 (  0.0%)   0.0000 (  0.0%)   0.0119 (  0.0%)   0.0079 (  0.0%)  Live
Variable Analysis
   0.0019 (  0.0%)   0.0000 (  0.0%)   0.0019 (  0.0%)   0.0047 (  0.0%)  Module
Verifier
   0.0029 (  0.0%)   0.0000 (  0.0%)   0.0029 (  0.0%)   0.0035 (  0.0%)  Loop
Strength Reduction
   0.0009 (  0.0%)   0.0010 (  1.0%)   0.0019 (  0.0%)   0.0032 (  0.0%)  Live
Variable Analysis
   0.0040 (  0.0%)   0.0000 (  0.0%)   0.0040 (  0.0%)   0.0027 (  0.0%) 
Two-Address instruction pass
   0.0009 (  0.0%)   0.0000 (  0.0%)   0.0009 (  0.0%)   0.0023 (  0.0%) 
Machine Code Deleter
   0.0030 (  0.0%)   0.0019 (  2.0%)   0.0050 (  0.0%)   0.0013 (  0.0%) 
Prolog/Epilog Insertion & Frame Finalization
   0.0009 (  0.0%)   0.0010 (  1.0%)   0.0019 (  0.0%)   0.0005 (  0.0%)  Scalar
Evolution Analysis
   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0004 (  0.0%)  ET
Forest Construction
   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0004 (  0.0%)  ET
Forest Construction
   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0004 (  0.0%) 
Immediate Dominators Construction
   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0003 (  0.0%) 
Eliminate PHI nodes for register allocation
   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0003 (  0.0%) 
Control Flow Optimizer
   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0003 (  0.0%) 
Immediate Dominators Construction
   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0003 (  0.0%)  Label
Folder
   0.0009 (  0.0%)   0.0000 (  0.0%)   0.0009 (  0.0%)   0.0003 (  0.0%)  X86 FP
Stackifier
   0.0009 (  0.0%)   0.0000 (  0.0%)   0.0009 (  0.0%)   0.0003 (  0.0%) 
Dominator Set Construction
   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0003 (  0.0%) 
Natural Loop Construction
   0.0010 (  0.0%)   0.0000 (  0.0%)   0.0010 (  0.0%)   0.0002 (  0.0%) 
Natural Loop Construction
   0.0009 (  0.0%)   0.0000 (  0.0%)   0.0009 (  0.0%)   0.0002 (  0.0%) 
Dominator Tree Construction
   0.0009 (  0.0%)   0.0000 (  0.0%)   0.0009 (  0.0%)   0.0002 (  0.0%)  Remove
unreachable blocks from the CFG
   0.0000 (  0.0%)   0.0010 (  1.0%)   0.0010 (  0.0%)   0.0002 (  0.0%) 
Canonicalize natural loops
   0.0000 (  0.0%)   0.0010 (  1.0%)   0.0010 (  0.0%)   0.0002 (  0.0%)  Lower
invoke and unwind, for unwindless code generators
   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0001 (  0.0%)  Lower
GC intrinsics, for GCless code generators
   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0000 (  0.0%)   0.0001 (  0.0%)  Target
Data Layout
  553.5438 (100.0%)   0.0979 (100.0%)  553.6418 (100.0%)  554.1256 (100.0%)  TOTAL

This is something like 30% from all compilation time of Mozilla here.



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.



More information about the llvm-bugs mailing list