Livermore Loop08 confuses CSE
Bugzilla Link | 903 |
Created on | Jan 16, 2012 20:10 |
Version | svn |
OS | Linux |
Architecture | PC |
Attachments | loop08.sac |
Extended Description
Created an attachment (id=841) source code to reproduce failure AWLFIfundef: AWLFI: Algebraic-With-Loop-Folding Inference in function _dup_23807_shift__Loop_58 ends SimplifySymb: SSE: Entering opt micro-cycle for function Loop8 SimplifySymb: SSE: LIR= 0, INL=0, CSE=10873, TUP=0, CF=37, VP=670, AS=0, AL=2, DL=0 SimplifySymb: SSE: LIR= 0, INL=0, CSE=195, TUP=0, CF=0, VP=14, AS=0, AL=0, DL=0 SimplifySymb: SSE: LIR= 0, INL=0, CSE=190, TUP=0, CF=0, VP=0, AS=0, AL=0, DL=0 SimplifySymb: SSE: LIR= 0, INL=0, CSE=190, TUP=0, CF=0, VP=0, AS=0, AL=0, DL=0 SimplifySymb: SSE: LIR= 0, INL=0, CSE=190, TUP=0, CF=0, VP=0, AS=0, AL=0, DL=0 SimplifySymb: SSE: LIR= 0, INL=0, CSE=190, TUP=0, CF=0, VP=0, AS=0, AL=0, DL=0 SimplifySymb: SSE: LIR= 0, INL=0, CSE=190, TUP=0, CF=0, VP=0, AS=0, AL=0, DL=0 SimplifySymb: SSE: LIR= 0, INL=0, CSE=190, TUP=0, CF=0, VP=0, AS=0, AL=0, DL=0 SimplifySymb: SSE: LIR= 0, INL=0, CSE=190, TUP=0, CF=0, VP=0, AS=0, AL=0, DL=0 SimplifySymb: SSE: LIR= 0, INL=0, CSE=190, TUP=0, CF=0, VP=0, AS=0, AL=0, DL=0 SimplifySymb: SSE: Stabilized at iteration 9 for function Loop8 AWLFIfundef: AWLFI: Algebraic-With-Loop-Folding Inference in function Loop8 ends This fails here, but I suspect it is not a new problem: sac2c -V sac2c v1.00-beta (Haggis And Apple) developer rev 17723 linux-gnu_x86_64 (Fri Jan 13 16:39:25 EST 2012 by sac) ~/sac/demos/benchmarks/livermore_loops/for_comparison/loop08$ sac2c loop08.sac -doawlf -nowlf -bopt -v4 -#d,AWLFI,SSE &>crud The SSE code in AWLFI calls the listed traversals in its own microcycle, including a DCR call not shown above. Oddly enough, CSE keeps finding the same expressions, even though they should be gone after the first pass. This causes the microcycle to run until maxoptcyc. Not a good thing.
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information