similar to: [LLVMdev] getting file name from Module

Hi, I'm playing with some experimental register allocators for LLVM. One of them needs to build a so-called register class tree for representing the aliasing information among register classes. This tree is not function or module specific. It is actually target specific, because it depends only on the register classes defined for a machine that is used as a target of the current compilation.

On 6/28/11 11:35 PM, kecheng at cecs.pdx.edu wrote: > Hi all, > > Llvm can automatically group a list passes. I want to do some > experiment to see the difference between grouping and non-grouping > execution. > Considering the following example, if I want to apply transformation > A, B and C. The execution order is: > A.doInitialization(); >

Thanks John, this is an easy way to achieve this goal, but I'm wondering if this is the only way? I have around 100 passes. Quoting John Criswell <criswell at illinois.edu>: > On 6/28/11 11:35 PM, kecheng at cecs.pdx.edu wrote: >> Hi all, >> >> Llvm can automatically group a list passes. I want to do some >> experiment to see the difference between grouping

Hi David, ----- Ursprüngliche Mail ---- > Von: David Greene <dag at cray.com> > An: llvmdev at cs.uiuc.edu > Gesendet: Dienstag, den 7. Oktober 2008, 18:14:45 Uhr > Betreff: Re: [LLVMdev] Getting target machine specific information at run-time > > On Tuesday 07 October 2008 06:57, Roman Levenstein wrote: > > Hi, > > > > I'm playing with some

Hello All, I have two questions, one more of an implementation question, the other more a design question. First: I noticed that if one moves the FPPassManager::doInitialization(Module) call from FPPassManager::runOnModule to MPPassManager::runOnModule (which is the new location I am aiming for to avoid the need for a doInitialization/doFinalization outside of the run methods, as preferred by

Hello All, I have two questions, one more of an implementation question, the other more a design question. First: I noticed that if one moves the FPPassManager::doInitialization(Module) call from FPPassManager::runOnModule to MPPassManager::runOnModule (which is the new location I am aiming for to avoid the need for a doInitialization/doFinalization outside of the run methods, as preferred by

Is there any way you could approximate the register/instruction usage and perform live-range analysis in a higher-level LLVM IR pass? I'm not sure how useful NVPTXRegisterInfo would be anyway. Unlike backends that target "real" ISAs, these structures do not contain any special properties about registers or instructions, like cost or scheduling information. Are you trying to figure

Hi all, Llvm can automatically group a list passes. I want to do some experiment to see the difference between grouping and non-grouping execution. Considering the following example, if I want to apply transformation A, B and C. The execution order is: A.doInitialization(); B.doInitialization(); C.doInitialization(); A.runOnFunction(); B.runOnFunction(); C.runOnFunction();

On Sun, 21 Jan 2007, Marcel Weiher wrote: > just installed the new llvm 1.9 build and noticed that my code no > longer worked. It seems something has changed with > addPassesToEmitFile(). First, the arguments to that method changed so > that it no longer takes a PassManager, but only a > FunctionPassManager. Instead there is a addPassesToEmitWholeFile() > method, but that is

On 8/9/11 6:49 AM, Bjorn Reese wrote: > I have an optimization pass (FunctionPass) where I need to add global > constructors. > > For cleaness sake I decided to add these in my own module. My module > is created in my FunctionPass constructor: This is not how I would do it. A FunctionPass has doInitialization() and doFinalization() methods that you can implement thatcan operate

Does this count have to be exact, or just an accurate approximation? The back-end may add/remove registers fairly late in the codegen process, so if you need an exact count you may need to run *just* before the assembly printer. Perhaps we could introduce a special machine node that represents a shared memory allocation. The node's value would be the shared address space pointer of the

OK. I know what you mean......... Simply speaking, I want to do some optimizations for PTX, and the information I need is similar to a register allocator. I know PTX is virtual ISA, but I will use PTX as the input of the simulator, gpgpu-sim, so it makes sense. Whether to insert shared memory is depend on the analysis that needs LiveAnalysis, PTX InstrInfo, PTX RegisterInfo, etc. That's why I

Yes, total number of PTX registers that will be emitted is exactly what I need. It's hard to figure out this in LLVM IR level. 2013/8/7 Justin Holewinski <justin.holewinski at gmail.com> > Is there any way you could approximate the register/instruction usage and > perform live-range analysis in a higher-level LLVM IR pass? I'm not sure > how useful NVPTXRegisterInfo

Yes, global variables are the only way to access shared memory. I'm just trying to get an idea of what you're aiming to accomplish to see if we can improve on the interface here. A MachineFunctionPass runs after instruction selection and relying on doInitialization to run before instruction selection is an implementation detail that I do not believe is guaranteed anywhere (I could be

I'm in a group tasked with improving the GC of LLVM for a 421 project. We are having trouble getting started with the given SemiSpace collector. We found the string llvm_gc_initialize called from a single source file ./test/Regression/CodeGen/Generic/GC/alloc_loop.ll which we tried with the following... (showing LLVM checked out from cvs a few days ago, similar output with release 1.3) $

Hi Alexei, I haven't tried to run code yet. The clang frontend can generate machine code for bpf but llc does not do the job. When i use : llc -march bpf -filetype=obj arith.bc It has this error message: Assertion failed: (isa<X>(Val) && "cast<Ty>() argument of incompatible type!"), function cast, file

On 2011-08-09 16:48, John Criswell wrote: > On 8/9/11 6:49 AM, Bjorn Reese wrote: >> I have an optimization pass (FunctionPass) where I need to add global >> constructors. >> >> For cleaness sake I decided to add these in my own module. My module >> is created in my FunctionPass constructor: > > This is not how I would do it. A FunctionPass has

Can you tell us a bit more about what you're trying to accomplish? Changes to the IR performed during MachineFunctionPass::doInitialization will likely propagate down through code generation, but at that point what is the purpose of using a MachineFunctionPass? You won't have any analysis or instruction information available until runOnMachineFunction. On Mon, Aug 5, 2013 at 12:00 PM,

Dear Contributors, I am asking help on the way how to solve a problem related to loops for that I always get confused with. I would like to perform the following procedure in a compact way. Consider that p is a matrix composed of 100 rows and three columns. I need to calculate the sum over some rows of each column separately, as follows: fa1<-(colSums(p[1:25,])) fa2<-(colSums(p[26:50,]))

I want to create share memory in my MachineFunctionPass, and insert load/store instruction for it. The way to create share memory is to add global variables which are in share memory address space (not sure if it is the only way). Therefore, I should add global variables in fixed size in doInitialization, and record its real size in other place like MachineModuleInfo. Then modify or query its real