similar to: [LLVMdev] Move instruction

r2->replaceAllUsesWith(r1); r2->eraseFromParent(); LLVM makes it easy! ;-) --Owen On Nov 15, 2008, at 11:02 PM, Karhu, Abhinav R wrote: > Hi owen, > Can you please elaborate as to what should I do? Find all the > instruction which have r2 in it and replace all of them with r1 and > then remove the load instruction? > Thanks > Abhinav > ----- Original Message

Hi Eli, I want to do redundant load elimination. A simple example for that would be load (a1) , r1; --- --- --- load (a2) , r2; Now if a1 and a2 are alias of each other and the value stored in a1 and a2 remains the same then I can say that the second load is redundant. I can replace the second load with these instructions mov r0 r1; mov r0 r2; and remove the load instruction. This is the thing

Hi all, Can anybody tell me how to generate a mov instruction in llvm. I found out there is a MemMove Inst but I dont think that is what I want. I want to bypass a load instruction. So i need to generate a mov instruction which will help me remove the redundant load. Thanks -- Abhinav Karhu MS Computer Science Georgia Institute of Technology

On Fri, Nov 14, 2008 at 2:13 PM, Karhu, Abhinav R <akarhu3 at gatech.edu> wrote: > > Hi all, > Can anybody tell me how to generate a mov instruction in llvm. I found out there is a MemMove Inst but I dont think that is what I want. > I want to bypass a load instruction. So i need to generate a mov instruction which will help me remove the redundant load. At the LLVM IL level,

Hey all, I compiled the code using both the -c and -S options. I can get the human readable IR but I still cannot run it using lli. The output of the IR looks like this. Can some of the information tell me about why the program is not running. ; ModuleID = 'Hel.c' target datalayout = "e-p:32:32" target endian = little target pointersize = 32 target triple =

Hi all, I am facing new problem with lli now. I am having a sample code in which I am doing malloc for two variables. If I compile the code with normal gcc the program runs without any warning of any sorts. If I compile the program to convert it into a bytecode file and then run it through lli it segfaults and the program aborts. This is the stack trace after execution. lli((anonymous

I have compiled the program using the following option. llvm-gcc --emit-llvm Hello.c -o Hel.bc I have used llvm in the past but this sort of problem is new for me. I dont know how to go about it. Thanks for the reply. -- Abhinav Karhu MS Computer Science Georgia Institute of Technology

Hey all, I am getting this weird error which suddenly popped up today. Whenever I am trying to execute a piece of code using lli I get this following error. Error loading program 'sample.bc': Invalid bytecode signature: 464C457F (Vers=0, Pos=4) I dont know what to do..Please reply to this question. Please... -- Abhinav Karhu MS Computer Science Georgia Institute of Technology

Hi Jingyue, Convergent is not intended to prevent inlining. It’s tricky to formalize this inter-procedurally, but the intended interpretation is that a convergent operation cannot be move either into or out of a conditionally executed region. Normal inlining would not violate that. I would imagine that it would make sense to use a combination of convergent and noduplicate for barrier-like

Below is a proposal for a new "convergent" intrinsic attribute and MachineInstr property, needed for correctly modeling many SPMD/SIMT programming models in LLVM. Comments and feedback welcome. —Owen In order to make LLVM more suitable for programming models variously called SPMD and SIMT, we would like to propose a new intrinsic and MachineInstr annotation called

Hi Mehdi, My reading of it is that if you have a convergent instruction A, it is legal to duplicate it to instruction B if (assuming B is after A in program flow) A dominates B and B post-dominates A. James On Fri, 14 Aug 2015 at 08:32 Mehdi Amini via llvm-dev < llvm-dev at lists.llvm.org> wrote: > On Aug 13, 2015, at 9:43 PM, Owen Anderson via llvm-dev < > llvm-dev at

Why is this a regalloc problem? I assume in the example below the "r0" is somehow forced by the ABI? Because otherwise moving the texture2d operation into the branch wouldn't matter as long as we assign different registers to the two branches and use a technique like lib/Target/R600/SIFixSGPRLiveRanges.cpp. - Matthias > On May 13, 2015, at 6:00 PM, Philip Reames <listmail at

Hey All, I am running into a problem with running the program using LLVM JIT. I get the following error. lli: error loading program 'Hel.bc': Bitcode stream should be a multiple of 4 bytes in length. Hel.bc is the bytecode for my simple HelloWorld program. Please give me some insight into this. Thanks -- Abhinav Karhu MS Computer Science Georgia Institute of Technology

Hey all, I am trying to see if we can gather alias analysis information during profiling, similar to block profiling or edge profiling info which can be gathered. Is there some way that can be achieved? A pointer on this topic would be much appreciated. Thanks -- Abhinav Karhu MS Computer Science Georgia Institute of Technology

Hey all, I want to run the profile.pl file in the utils folder but I am having trouble running the script because the profile_rt.so file is not generated properly. So in order to generate that I am running make in the /runtime/libprofile. when I do that I get the following message. llvm[0]: Linking Release Lodable Module profile_rt.so llvm[0]: Compiling BasicBlockTracing.c for Release build

Alright, go ahead with it. —Owen > On Nov 17, 2014, at 4:58 AM, Daniil Troshkov <troshkovdanil at gmail.com> wrote: > > Hi! > > I have not found test case. (It is because we have no target using "ExpandInsertToVectorThroughStack"). > But I tested it for target currently not included in llvm trunk. > > This fix correct and trivial, so I'm offering

Surprisingly enough, libatomic_ops doesn't define just a hardware memory fence call as far as I can tell. --Owen On May 16, 2009, at 3:00 PM, Zoltan Varga wrote: > Hi, > > You might want to use this: > > http://www.hpl.hp.com/research/linux/atomic_ops/ > > Zoltan > > On Sat, May 16, 2009 at 11:11 PM, Owen Anderson <resistor at

BTW. I checked help page of contour and maybe it could mention a note about akima package or interp function. Petr ------- Forwarded message follows ------- From: Petr Pikal <petr.pikal at precheza.cz> To: "Abhinav Verma" <abhinav1205 at gmail.com>, r-help at stat.math.ethz.ch Subject: Re: [R] Reading xyz data from a file and plotting a

What would you do with a just-hardware memory fence? If the compiler's free to move operations over the hardware fence, that seems to defeat the purpose. C++0X provides a compiler-only fence, and a hardware+compiler fence, but no hardware-only fence, I believe for this reason. See <http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2009/n2857.pdf>, section 29.8. On Sat, May 16, 2009 at

I detected this bug using test case from platform which is not currently supported on llvm targets. (Our team is porting llvm on new target). Creating the test case will take some extra time. I'll try to do it ASAP. Have you any ideas about the test case? (targets using ExpandInsertToVectorThroughStack, etc...) On Wed, Nov 12, 2014 at 8:29 PM, Owen Anderson <resistor at mac.com> wrote: