similar to: [LLVMdev] Wondering how best to run inlining on a single function.

On May 26, 2009, at 3:15 PM, Jeffrey Yasskin wrote: > In Unladen Swallow we (intend to) compile each function as we > determine it's hot. To "compile" a function means to translate it from > CPython bytecode to LLVM IR, optimize the IR using a > FunctionPassManager, and JIT the IR to machine code. We'd like to > include inlining among our optimizations. Currently

Hi, Slightly off-topic (as it's not directly about using or developing LLVM): http://code.google.com/p/unladen-swallow/wiki/ProjectPlan "Our long-term proposal is to replace CPython's custom virtual machine with a JIT built on top of LLVM, while leaving the rest of the Python runtime relatively intact." Just curious, has anyone here heard more about this project? Regards,

On Thu, Mar 26, 2009 at 8:20 AM, Paul Melis <llvm at assumetheposition.nl> wrote: > Hi, > > Slightly off-topic (as it's not directly about using or developing LLVM): > > http://code.google.com/p/unladen-swallow/wiki/ProjectPlan > > "Our long-term proposal is to replace CPython's custom virtual machine > with a JIT built on top of LLVM, while leaving the

Hello All, I have some data tied to the basic blocks in a function, and after inlining that function, I'd like to recover that data in the inlined version. Is there some way to match up the inlined version of the function with the original basic blocks? Thanks, Jeremy -------------- next part -------------- An HTML attachment was scrubbed... URL:

Hi, I found the following wiki page in the Unladen Swallow project: http://code.google.com/p/unladen-swallow/wiki/CodeLifecycle This would appear to answer my question. Could someone confirm for me if it's definitely unsafe to attempt to optimise/JIT any Modules while a different thread is currently executing a JITed function which has been generated from them? Or am I just missing

Hi, Has anyone had any success with running different PassManagers on llvm::Modules they've already JITed and are executing? In detail: 1) getting the IR, in form of an llvm::Module 2) calling PassManager->run() on the module 3) calling getFunction() and getPointerToFunction() to JIT the module 4) executing the JITed code using the function pointer received in step 3 and then what I

Devang Patel wrote: > Hi John, > > On Wed, Nov 4, 2009 at 12:04 PM, John Criswell <criswell at uiuc.edu> wrote: > >> Dear All, >> >> 1) I recall reading somewhere that a few optimizations in LLVM 2.6 strip >> away debug information when such information interferes with >> optimization. Is this correct, >> > > Yes. > >

According to http://llvm.org/docs/DebuggingJITedCode.html only ELF objects are supported for gdb support of JIT generated code. Is this still true, or are Mach-O files now supported in 2.7 trunk? Thanks in advance Garrison

Hi, I am writing a CallGraphSCCPass that uses LoopInfo which is a FunctionPass. However, doing so results in the following error. **** Unable to schedule 'Natural Loop Information' required by '......' **** Google led me to this page, where Devang Patel suggests implementing the addLowerLevelRequiredPasses in CGPassManager in a manner similar to MPPassManager.

I have a whole slew of questions about optimization passes. Answers to any or all would be extremely helpful: How important are doInitialization/doFinalization? I can't detect any difference if I use them or not. Why does the function pass manager have doInitialization/doFinalization, but the global pass manager doesn't? If I am applying the function passes to many functions, do I

Larry, On Wed, Aug 11, 2010 at 4:55 PM, Larry Gritz <lg at larrygritz.com> wrote: > I have a whole slew of questions about optimization passes. Answers to any > or all would be extremely helpful: > > How important are doInitialization/doFinalization? Most of the passes do not use them. > I can't detect any difference if I use them or not. Say, if you are writing

Hi, In following link: https://www.youtube.com/watch?reload=9&v=6X12D46sRFw They have specified that the inliner can't use DomTree/LoopInfo/MemorySSA analysis. 1. What's the reason for this? 2. Why can't we do it using getAnalysisUsage() construct? 3. Can inline use this information in the new Pass Manager? 4. What all information can we derive from DomTree to be of help to

Hi all, I'm attempting to use LLVM for run-time code specialization, but I'm facing a performance hurdle. I'm currently performing the specialization during the AST to LLVM IR translation, but unfortunately this leads to relatively slow recompiles as LLVM has to perform all the heavy (optimization) passes over and over again. So I was hoping that by first creating unspecialized LLVM

I would like to start prototyping changes towards profile guided inlining. Before doing so, I wanted to get feedback from the community on the appropriate approach. I'm going to layout a strawman proposal, but I'm open to other ideas on how to approach the problem. Depending on what approach we settle on, I *may* be able to commit resources to actually implement this in the near

The patch below basically hammers out some ideas as to where I'd like to take the passmanager in LLVM. I've tried thinking things through, but I'm still a n00b, so some criticism would be more than welcome. =) Starting from line 191 down. If you're wondering why I created a patch, well that's because I found thinking in passmanagert.h the most productive. -- Regards.

Not sure, but it seems the page contains a number of out-of-date entries: Pypy => pypy.org (link stale) plus: there is no llvm backend for pypy at the moment (although LLVM backends have been attempted a number of times, all seem to have failed) Unladen Swallow => not being developed since 2011 (http://qinsb.blogspot.com/2011/03/unladen-swallow-retrospective.html) TIA, Andreas The

Hello Duncan, > thanks for your interesting email. Do you understand why PyPy is no longer > using LLVM, and why Unladen Swallow died? Does LLVM need to be improved in > some way? The answers to all these questions are linked: LLVM is not fast enough (for a JIT). Of course this is not the whole story, but it is the LLVM-relevant part. Let's have a look at some random performance

I want to connect each llvm ir for example: 1. Turn C/C++ language into C_llvmIR assembly language using Clang 2. Turn Fortran language into Fortran_llvmIR assembly language using Dragonegg 3. Turn Python language into Python_llvmIR assembly language using Unladen-Swallow 4. Connect each llvm IR Is this possible? Wonjun, Choi -------------- next part -------------- An HTML attachment was

Hi Travis, ... > LLVM is still very relevant to Python because of projects like Numba --- but you > should know that PyPy is no longer using LLVM and Unladen Swallow has not been > worked on for several years. The future of LLVM and Python I think is very > bright --- especially for the scientific and data-analysis user-base. thanks for your interesting email. Do you understand

I've got a call instruction: call void bitcast (void (%4 addrspace(1)*, <2 x i32>, <4 x float>)* @_Z12write_imagefPU3AS110_image2d_tDv2_iDv4_f to void (%9 addrspace(1)*, <2 x i32>, <4 x float>)*)(%9 addrspace(1)* %dstimg, <2 x i32> %28, <4 x float> %26) nounwind %4 and %9 are both (stripped) opaque structs. InlineFunction() does not inline this because