Displaying 11 results from an estimated 11 matches for "tempscop".
2011 Nov 01
0
[LLVMdev] How to make Polly ignore some non-affine memory accesses
...tor::getPwAff(polly::ScopStmt const*, llvm::SCEV const*,
llvm::Value const*) + 57
12 LLVMPolly.dylib 0x000000010685d076
polly::ScopStmt::buildConditionSet(polly::Comparison const&) const +
54
13 LLVMPolly.dylib 0x000000010685d414
polly::ScopStmt::addConditionsToDomain(isl_set*, polly::TempScop&,
llvm::Region const&) const + 196
14 LLVMPolly.dylib 0x000000010685d519
polly::ScopStmt::buildDomain(polly::TempScop&, llvm::Region const&)
const + 169
15 LLVMPolly.dylib 0x000000010685d90d
polly::ScopStmt::ScopStmt(polly::Scop&, polly::TempScop&, llvm::Region
con...
2011 Oct 27
2
[LLVMdev] How to make Polly ignore some non-affine memory accesses
Perfect, thank you very much :)
2011/10/26 Tobias Grosser <tobias at grosser.es>:
> On 10/24/2011 11:32 PM, Marcello Maggioni wrote:
>>
>> Strange , with --enable-shared (I use auto tool by the way ...) it gives:
>>
>> MacBook-Pro-di-Marcello:examples Kariddi$ ./compile_ex.sh
>> not_so_simple_loop
>> clang (LLVM option parsing): Unknown command line
2011 Nov 14
0
[LLVMdev] How to make Polly ignore some non-affine memory accesses
...ale_down ()
#3 0x0000000101a5a175 in polly::MemoryAccess::MemoryAccess
(this=0x101911ed0, Access=@0x10190fc48, Statement=0x10190ffa0) at
/Users/Kariddi/Documents/Sviluppo/Tesi/llvm/tools/polly/lib/Analysis/ScopInfo.cpp:308
#4 0x0000000101a5b22a in polly::ScopStmt::buildAccesses
(this=0x10190ffa0, tempScop=@0x10190eb70, CurRegion=@0x10190eed0) at
/Users/Kariddi/Documents/Sviluppo/Tesi/llvm/tools/polly/lib/Analysis/ScopInfo.cpp:588
#5 0x0000000101a5bc08 in polly::ScopStmt::ScopStmt (this=0x10190ffa0,
parent=@0x10190fdf0, tempScop=@0x10190eb70, CurRegion=@0x10190eed0,
bb=@0x101905bd0, NestLoops=@0x7ff...
2011 Nov 02
5
[LLVMdev] How to make Polly ignore some non-affine memory accesses
...const*, llvm::SCEV const*,
> llvm::Value const*) + 57
> 12 LLVMPolly.dylib 0x000000010685d076
> polly::ScopStmt::buildConditionSet(polly::Comparison const&) const +
> 54
> 13 LLVMPolly.dylib 0x000000010685d414
> polly::ScopStmt::addConditionsToDomain(isl_set*, polly::TempScop&,
> llvm::Region const&) const + 196
> 14 LLVMPolly.dylib 0x000000010685d519
> polly::ScopStmt::buildDomain(polly::TempScop&, llvm::Region const&)
> const + 169
> 15 LLVMPolly.dylib 0x000000010685d90d
> polly::ScopStmt::ScopStmt(polly::Scop&, polly::T...
2012 Feb 08
2
[LLVMdev] Polly and non affine branches in ScoPs
...even if these branches are present,
because even if the analysis gets wrong thanks to these branches it is
not that big of a deal for my project.
Can't this feature be implemented like the one that ignores the non
affine memory subscripts in ScoPs? To implement this would be needed a
change in TempScop and ScopInfo or a change only in the code that
verifies the branches in ScopDetection would be enough?
Thanks
Marcello
2012 Feb 12
0
[LLVMdev] Polly and non affine branches in ScoPs
...esent,
> because even if the analysis gets wrong thanks to these branches it is
> not that big of a deal for my project.
>
> Can't this feature be implemented like the one that ignores the non
> affine memory subscripts in ScoPs? To implement this would be needed a
> change in TempScop and ScopInfo or a change only in the code that
> verifies the branches in ScopDetection would be enough?
OK. So here are several issues.
First of all, the following code is perfectly affine:
for (i
if (i % 2 > 5)
...
A modulo were the right side is an integer constant, can easily...
2012 Feb 14
1
[LLVMdev] Polly and non affine branches in ScoPs
...even if the analysis gets wrong thanks to these branches it is
>> not that big of a deal for my project.
>>
>> Can't this feature be implemented like the one that ignores the non
>> affine memory subscripts in ScoPs? To implement this would be needed a
>> change in TempScop and ScopInfo or a change only in the code that
>> verifies the branches in ScopDetection would be enough?
>
>
> OK. So here are several issues.
>
> First of all, the following code is perfectly affine:
>
> for (i
> if (i % 2 > 5)
> ...
>
> A modulo were...
2013 May 02
0
[LLVMdev] [Polly] GSoC Proposal: Reducing LLVM-Polly Compiling overhead
...ization step to
Typo -----> canonicalization
> allow Polly boil out early if it cannot benefit programs.
> StageV -- Improve other parts. [Week 11-12]
>
> Week 11: Improve other parts of Polly. Especially, I will focus on some
> expensive helper functions such as TempScop analysis. This helper
> function is critical and expensive.
How do you know TempScop is expensive?
> Week 12: Integrate all improvements
> and evaluate the whole Polly with multiple benchmarks.
I think the only way to do this project is to continuously evaluate your
changes on Polybenc...
2013 Jul 05
0
[LLVMdev] [Polly] Assert in Scope construction
...013 12:29 PM
> > To: 'Tobias Grosser'
> > Cc: 'llvmdev'
> > Subject: Re: [LLVMdev] [LNT] Question about results reliability in LNT
> > infrustructure
> >
> >
> > Tobias,
> >
> > I seem to trigger an assert in Polly lib/Analysis/TempScopInfo.cpp
> >
> > void TempScopInfo::buildAffineCondition(Value &V, bool inverted,
> > Comparison **Comp) const { ...
> > ICmpInst *ICmp = dyn_cast<ICmpInst>(&V);
> > assert(ICmp && "Only ICmpInst of...
2013 Apr 26
4
[LLVMdev] [Polly] GSoC Proposal: Reducing LLVM-Polly Compiling overhead
Hi all,
I have updated my GSoS proposal: "FastPolly: Reducing LLVM-Polly Compiling overhead" (https://gist.github.com/tanstar/5441808). I think the pass ordering problem you discussed early can be also investigated in this project!
Is there any comment or advice about my proposal? I appreciate all your help and advice.
Thanks,
Star Tan
Proposal:
2013 Jul 03
2
[LLVMdev] [Polly] Assert in Scope construction
...gei Larin
> Sent: Wednesday, July 03, 2013 12:29 PM
> To: 'Tobias Grosser'
> Cc: 'llvmdev'
> Subject: Re: [LLVMdev] [LNT] Question about results reliability in LNT
> infrustructure
>
>
> Tobias,
>
> I seem to trigger an assert in Polly lib/Analysis/TempScopInfo.cpp
>
> void TempScopInfo::buildAffineCondition(Value &V, bool inverted,
> Comparison **Comp) const { ...
> ICmpInst *ICmp = dyn_cast<ICmpInst>(&V);
> assert(ICmp && "Only ICmpInst of constant as condition su...