similar to: [LLVMdev] Modifing an operand of MDNode

On Jun 7, 2010, at 3:35 PM, Junjie Gu wrote: > I have a MDNode, and one of its operand is an int. I'd like to > assign a new int value to that operand during the llvm optimization. > What is the right way to do this ? (There is no setOperand() with > MDNode class.) Hi Junjie, MDNodes are uniqued, so you can't change them. Just make a new MDNode with the new operand list

I'd like to add a pragma support in llvm. I am thinking about using a llvm intrinsic to represent each pragma, such as llvm.pragma (metadata, ...) where metadata describes a pragma. So if an application has: #pragma p1 .. #pragma p2... for (...) The llvm IR would be llvm.pragma (metadata..) // for p1 llvm.pragma (metadata..) // for p2 llvm IR for "for

Hi Chris, Thanks. I will see what I can do for this. Junjie On Wed, May 19, 2010 at 3:45 PM, Chris Lattner <clattner at apple.com> wrote: > > On May 19, 2010, at 2:38 PM, Junjie Gu wrote: > >> Many compilers support per-loop pragma, such as loop unrolling (ie >> #pragma unroll=2).  Is there any LLVM project/effort going on >> in this area ? What is the expected

On Nov 12, 2013, at 4:59 PM, Chandler Carruth <chandlerc at google.com> wrote: > Other options are: > 1> Using a different LLVMContext for the destination module, but it didn’t work out since Linker was not designed to work with different LLVMContexts for source vs destination. > 2> removeUnusedMDNodes checks if a MDNode is used in a different way (i.e use_empty() &&

Thanks for comments/suggestions. Yes, attaching metadata to instructions will be good choices for many cases. But for loops, attaching metadata to back-edges requires that the front end to build loops, which is an additional task for the front end. And this task is really a backend's job, not the front end's. If the only concern is that it is hard for pragma intrinsics to stay with their

On Tue, Jun 1, 2010 at 5:42 PM, Junjie Gu <jgu222 at gmail.com> wrote: > Any suggestions/ideas ? IIUC, Chris suggested something like following ... header: br i1 %x, label %then, %label endif then: ... br i1 %y, label %loop_exit, label %header, !loop_pragma !1 endif: ... br i1 %z, label %loop_exit, label %header, !loop_pragma !2 loop_exit: ret i32 1 Where !1 and !2 are the

On Tue, Nov 12, 2013 at 4:46 PM, Manman Ren <manman.ren at gmail.com> wrote: > > > > On Tue, Nov 12, 2013 at 4:38 PM, Chandler Carruth <chandlerc at google.com>wrote: > >> >> On Tue, Nov 12, 2013 at 4:29 PM, Manman Ren <manman.ren at gmail.com> wrote: >> >>> Hi All, >>> >>> In LTO, we load in the source modules and

On Tue, Nov 12, 2013 at 4:38 PM, Chandler Carruth <chandlerc at google.com>wrote: > > On Tue, Nov 12, 2013 at 4:29 PM, Manman Ren <manman.ren at gmail.com> wrote: > >> Hi All, >> >> In LTO, we load in the source modules and link the source modules into a >> destination module. >> Lots of MDNodes are only used by the source modules, for example

On Wed, Sep 28, 2011 at 5:47 PM, Eli Friedman <eli.friedman at gmail.com> wrote: > On Wed, Sep 28, 2011 at 3:27 PM, Junjie Gu <jgu222 at gmail.com> wrote: >> Instrinsic llvm.memory.barrier does not work as expected.  Is it a bug >> or it has not been implemented yet ? > > It's going away in favor of the new fence instruction (and I'll remove > it as soon

If LLVM would like to support OpenMP pragma in the future, not sure if attaching metadata to instructions is still a good choice. Junjie On Wed, Jun 2, 2010 at 11:08 AM, Junjie Gu <jgu222 at gmail.com> wrote: > Thanks for comments/suggestions. > > Yes, attaching metadata to instructions will be good choices for many > cases. But for loops, > attaching metadata to back-edges

On Tue, Nov 12, 2013 at 6:02 PM, Chris Lattner <clattner at apple.com> wrote: > On Nov 12, 2013, at 4:59 PM, Chandler Carruth <chandlerc at google.com> > wrote: > > Other options are: >>>> 1> Using a different LLVMContext for the destination module, but it >>>> didn’t work out since Linker was not designed to work with different >>>>

On Tue, Nov 12, 2013 at 4:29 PM, Manman Ren <manman.ren at gmail.com> wrote: > Hi All, > > In LTO, we load in the source modules and link the source modules into a > destination module. > Lots of MDNodes are only used by the source modules, for example Xalan > used 649MB for MDNodes after loading and linking, but the actual > destination module only has 393MB of

Hi All, In LTO, we load in the source modules and link the source modules into a destination module. Lots of MDNodes are only used by the source modules, for example Xalan used 649MB for MDNodes after loading and linking, but the actual destination module only has 393MB of MDNodes. There are 649-393MB (40% of 649MB) not used. MDNodes belong to the Context, deleting modules will not release the

On Wed, Sep 28, 2011 at 3:27 PM, Junjie Gu <jgu222 at gmail.com> wrote: > Instrinsic llvm.memory.barrier does not work as expected.  Is it a bug > or it has not been implemented yet ? It's going away in favor of the new fence instruction (and I'll remove it as soon as dragonegg catches up). It should still work at the moment, though. > (1) false arguments do not work >

On Wed, Jun 2, 2010 at 11:08 AM, Junjie Gu <jgu222 at gmail.com> wrote: > Thanks for comments/suggestions. > > Yes, attaching metadata to instructions will be good choices for many > cases. But for loops, > attaching metadata to back-edges requires that the front end to build > loops, which is an > additional task for the front end. And this task is really a

Dear LLVM community, LLVM IR has a concept of 'LoopID' [1] which is a misnomer: (a) LoopIDs are not unique: Any pass that duplicates IR will do it including its metadata (e.g. LoopVersioning) such that thereafter multiple loops are linked with the same LoopID. There is even a test case (Transforms/LoopUnroll/unroll-pragmas-disabled.ll) for multiple loops with the same LoopID. (b)

I am trying to add an analysis pass as a FunctionPass, and let LICM (LoopPass) depends upon it. So in LICM.cpp, I have the following: virtual void getAnalysisUsage(AnalysisUsage &AU) const { AU.setPreservesCFG(); AU.addRequired<DominatorTree>(); AU.addRequired<LoopInfo>(); AU.addRequiredID(LoopSimplifyID); AU.addRequired<AliasAnalysis>();

On Mon, Jul 27, 2009 at 9:31 PM, Nick Lewycky<nicholas at mxc.ca> wrote: > Why not have a named GlobalValue with an MDNode initializer? How is this > different from what we had before? GlobalValue initializer accepts only Constants. - Devang

Many compilers support per-loop pragma, such as loop unrolling (ie #pragma unroll=2). Is there any LLVM project/effort going on in this area ? What is the expected way for implementing per-loop pragma, or general pragma ? Suggestions/comments ? Thanks Junjie

On Tue, Nov 12, 2013 at 4:59 PM, Chandler Carruth <chandlerc at google.com>wrote: > On Tue, Nov 12, 2013 at 4:46 PM, Manman Ren <manman.ren at gmail.com> wrote: > >> >> >> >> On Tue, Nov 12, 2013 at 4:38 PM, Chandler Carruth <chandlerc at google.com>wrote: >> >>> >>> On Tue, Nov 12, 2013 at 4:29 PM, Manman Ren <manman.ren