search for: rlwinm

...de didn't read correctly. FAIL: /usr/home/jeffc/llvm/obj/../test/Regression/CodeGen/PowerPC/rlwimi2.ll: Output/rlwimi2.ll.out.script: cannot open rlwimi2.ll: No such file or directory llc: bytecode didn't read correctly. FAIL: /usr/home/jeffc/llvm/obj/../test/Regression/CodeGen/PowerPC/rlwinm.ll: Output/rlwinm.ll.out.script: cannot open rlwinm.ll: No such file or directory llc: bytecode didn't read correctly. Output/rlwinm.ll.out.script: cannot open rlwinm.ll: No such file or directory llc: bytecode didn't read correctly. Output/rlwinm.ll.out.script: cannot open rlwinm.ll: No...

...ubyte %c8, label %loop_step [ ubyte 97, label %ret_true ubyte 98, label %ret_true ] Unfortunately, this generates really weird code on the LLVM 1.6 PowerPC backend: LBB_matches_1: ; regex6 lbz r4, 0(r3) LBB_matches_2: ; NodeBlock rlwinm r5, r4, 0, 24, 31 cmplwi cr0, r5, 98 blt cr0, LBB_matches_4 ; LeafBlock LBB_matches_3: ; LeafBlock1 rlwinm r4, r4, 0, 24, 31 cmpwi cr0, r4, 98 beq cr0, LBB_matches_8 ; ret_true b LBB_matches_5 ; NewDefault LBB_matches_4: ; LeafBlock...

...dags) and that complex pattern which I specify seems to be correct. It seems that the problem is with function: SDNode *PPCDAGToDAGISel::Select(SDOperand Op) in PPCISelDAGtoDAG.cpp file with case ISD::ADD when it checks whenever and'ed constant is rotated first. In such case he puts PPC::RLWINM instruction into DAG (return CurDAG- >SelectNodeTo(N, PPC::RLWINM, MVT::i32, Ops, 4);) My question is: can I use patterns to fetch that - if yes, then how (there is no RLWINM SDNode). When that function is used, after or before legalization? I don't see anything on my printouts and tha...

On Mar 17, 2006, at 7:54 AM, Eric Kidd wrote: > Unfortunately, this generates really weird code on the LLVM 1.6 > PowerPC backend: > > LBB_matches_1: ; regex6 > lbz r4, 0(r3) > LBB_matches_2: ; NodeBlock > rlwinm r5, r4, 0, 24, 31 > cmplwi cr0, r5, 98 > blt cr0, LBB_matches_4 ; LeafBlock > LBB_matches_3: ; LeafBlock1 > rlwinm r4, r4, 0, 24, 31 > cmpwi cr0, r4, 98 > beq cr0, LBB_matches_8 ; ret_true > b LBB_matches_5 ; NewDefault > LB...

On Thu, 16 Mar 2006, Eric Kidd wrote: > Hello! I'm compiling code which uses pointers as iterators. For some > reason--probably a silly misunderstanding of the docs--I can't eliminate > duplicate pointer loads. I'll probably figure this out eventually, but if > somebody else sees the answer instantly, I certainly won't complain. :-) There are no stupid questions.

Hello! I'm compiling code which uses pointers as iterators. For some reason--probably a silly misunderstanding of the docs--I can't eliminate duplicate pointer loads. I'll probably figure this out eventually, but if somebody else sees the answer instantly, I certainly won't complain. :-) Here are the optimizers I'm running: opt -f -simplifycfg -dce -instcombine

On Fri, 17 Mar 2006, Eric Kidd wrote: >> I'm particularly confused by the rlwinm instructions that keep turning up >> in PowerPC output, and the double test against 98. I don't have a problem >> or anything; I'm just trying to figure out what's going on. :-) > > Ah! The backend is running -lowerswitch, which does a binary search of the > case...

...n shorten the critical path but I hadn't really thought > about making that into a top-level goal. > What I said is sort of misleading in that the targets have a variety of "exotic" instructions that can effectively collapse multiple nodes in the critical path e.g. x86 BMI, PPC rlwinm, etc. Even stuff like whether the target has ANDN will affect the critical path (or an instruction taking a general truth table). Also the precise cost in a critical path of the individual instructions can vary a fair amount.to the extent that except for getting rid of mul's or div's it'...

> Interesting. Do you happen to have some examples laying around? Sorry, no, I'll have to re-run. I felt that the select-heavy results were sort of humorous and clever at first, but then just annoying. > Except for decode-limited situations, in general decreasing the critical path length is more important > than eliminating instructions. The critical path length is a

...document: http://llvm.cs.uiuc.edu/docs/LangRef.html#constants Target Specific Code Generator Improvements: 19. Misha implemented the machine code emitter for the PowerPC target. 20. Nate substantially improved instruction selection in the PowerPC backend, allowing it to generate rlwimi, rlwinm, and other instructions. 21. Jeff Cohen contributed patches to allow the X86 backend to fold the addresses of globals more aggressively into X86 memory references. 22. John contributed a patch to the C backend to make it emit code that has fewer dynamic copies for PHI nodes, speeding...