similar to: [LLVMdev] sccp pass with opt

This compiler does not have mem-SSA, as far as I know, only few pass can propagate value along memory. You need to promote those local variable into register first before sccp is invoked. e.g1. opt a.ll -basicaa -gvn -sccp -S eg.2. opt a.ll -sroa -sccp -S On 4/16/13 12:37 PM, Niko Zarzani wrote: > Hi all, > > I am trying to see how single llvm optimizations work by running them

Thank you, it was what I really was searching for :) However, I don't know if I well understood. I've find this link in the second link which seemed what i was looking for: http://nondot.org/sabre/LLVMNotes/BuiltinUnreachable.txt .If I put around the code block (inside my function with precondition (x>0 && y>0)) a contruct like that that use __builtin_unreachable: int foo(int

On Mon, Oct 22, 2012 at 9:33 PM, Niko Zarzani <koni10 at hotmail.it> wrote: > Thank you, it was what I really was searching for :) > > However, I don't know if I well understood. I've find this link in the > second link which seemed what i was looking for: > http://nondot.org/sabre/LLVMNotes/BuiltinUnreachable.txt . > If I put around the code block (inside my

Hi all,Is there any way to tell LLVM some additional information about the variables in the code in order to make better optimization?For example, if my function has a certain precondition (such as x>0) then it will be possible to better optimize the code given that information (which the compiler does not know).I am new in this field and I don't know if there are ways to tell the compiler

Hi all, I wrote 2 passes and I want to make run llvm run the passes in this order: -mem2reg -load=…/mypass1.dylib -mypass1 -load=…/mypass2.dylib -mypass2 -O1 -O2 -O3 I know I can do this by manually passing them as an argument to opt. Is there any way to force this sequence directly from clang? I am asking this because I am trying to compile a program and I can specify in the ./configure

Hi Niko, Do you mean branch prediction, i.e. __builtin_expect [1]? Many compilers support it, I think clang (LLVM's C/C++ frontend) is among them. - D. [1] http://gcc.gnu.org/onlinedocs/gcc/Other-Builtins.html 2012/10/23 Niko Zarzani <koni10 at hotmail.it>: > Hi all, > Is there any way to tell LLVM some additional information about the > variables in the code in order to

I tried it on a simple file test.c (as you did) but actually clang does not show me any output. In addition I tried to set the CFLAGS specifying the -mllvm -time-passes options. However I still have anything about the time passes informations in the output and I also get this warning during the build with make:clang: warning: argument unused during compilation: '-mllvm -time-passes'Maybe

You may want to check this out: http://lists.cs.uiuc.edu/pipermail/llvmdev/2012-October/053924.html and also http://llvm.org/PR810 - xi On 10/22/12 6:05 PM, Niko Zarzani wrote: > Hi all, > Is there any way to tell LLVM some additional information about the > variables in the code in order to make better optimization? > For example, if my function has a certain precondition (such

Hi all, I was interested in knowing which of the passes spends the most of the time compiling the dns server BIND.With the -CC option I selected clang as my compiler and, as I was expecting, the compilation gave me no problems.Now I have a question, is there any way to set an option similiar to "-time-passes" to clang in order to get those information I was interested in?Or do you think

To use options like "-time-passes" and "-stats" you need to prefix them with "-mllvm": clang -mllvm -time-passes -O3 test.c -o /dev/null Note that I believe you need asserts enabled or some other build configuration for this to work, the timing code is disabled in fully release builds IIRC. Hope this helps! ~Will On Thu, Sep 12, 2013 at 11:50 AM, Niko Zarzani

Why in the webpage http://llvm.org/docs/Passes.html some of the passes such as "correlated value propagation" or "early cse" are not described? Am I missing some other webpage with these informations? Thank you in advance, Niko -------------- next part -------------- An HTML attachment was scrubbed... URL:

On Oct 5, 2013, at 9:17 PM, Niko Zarzani <koni10 at hotmail.it> wrote: > Hi all, > > I wrote 2 passes and I want to make run llvm run the passes in this order: > -mem2reg -load=…/mypass1.dylib -mypass1 -load=…/mypass2.dylib -mypass2 -O1 -O2 -O3 > > I know I can do this by manually passing them as an argument to opt. > > Is there any way to force this sequence

> > I have written them by hand in the .s file in the same way of the IR > > reference (http://llvm.org/docs/LangRef.html#named-metadata) : > > > > define i32 @function(i32 %argInt, metadata !3) nounwind { > > This seems wrong. "metadata" is a type (like i32), and the exclamation > mark is only used to refer to metadata nodes, not to declare functions

Hi, I am trying to write a Pass that reads metadata passed as a argument to a function. I have written them by hand in the .s file in the same way of the IR reference (http://llvm.org/docs/LangRef.html#named-metadata) : define i32 @function(i32 %argInt, metadata !3) nounwind {entry: %argInt.addr = alloca i32, align 4 store i32 %argInt, i32* %argInt.addr, align 4 %1 = load i32* %argInt.addr,

Since variables in the source code are renamed in the IR, I wanted to ask you how and where this mapping is done in Clang.I am interested in converting the variable names in some C strings to the one in the IR.For example if at a certain point of the program I have a string like "x>0" I want to change it in "%x>0" (I already implemented a parser to recognize the

Chris Lattner wrote: > On Tue, 9 May 2006, Nick Lewycky wrote: > >>> You could do that, but SCCPSolver isn't really useful to mainline LLVM >>> for anything other than SCCP and IPSCCP, so we don't need it in a public >>> header. Out of curiosity, what are you looking to use the latice values >>> for? Why not just run SCCP and then look at the

Hi, I am interested in developing some passes using the Xcode IDE. I create a pass directory with a simple pass in it, add it to the lib/Trasform/ directory, changed and added some CMakeList.txt files and compiled using cmake (cmake -G Xcode). The compilation gives me no errors but then I've found that I have not a Debug+Asserts folder and opt has no option related to my pass...what can I do

On Wed, 10 May 2006, Nick Lewycky wrote: >> Then just run the SCCP pass, and check to see if any operands satisfy >> the predicate "isa<UndefValue>(V)". LLVM explicitly represents >> undefined values. > > I have a case where it doesn't, but perhaps the SCCP pass isn't to blame: > > extern void write_char(int); > > const char foo[4] =

Hi. I'm sending this email to -dev as this may be of interest of many/people may have opinions/want to try the change before it goes in to report problems. I've been recently working on a patch to integrate `undef` in the SCCP solver, in the hope of fixing a tail of latent bugs in SCCP which remained uncovered for many years. I think this is a decent time to propose, so that it can

Hi David, Looking at the original bug, it seems like a straightforward undef-propagation bug to me -- SCCP was folding "or undef, constant" to "undef", which is wrong. Why is changing that not the fix? That is, some variant of diff --git a/lib/Transforms/Scalar/SCCP.cpp b/lib/Transforms/Scalar/SCCP.cpp index 8a6be97..45f1241 100644 --- a/lib/Transforms/Scalar/SCCP.cpp +++