llvm dev - Jun 2013 - [LLVMdev] LLVM+Clang 3.3: clang: for the -vectorize-loops option: may only occur zero or one times!

Hi,

 

My application generates some C code which I then compile using LLVM+Clang
before running it. For the compilation side of things, I have a class which
one of the methods (the CompilerEngine::compileCode() method in
https://raw.github.com/opencor/opencor/llvm_3.3/src/plugins/misc/Compiler/sr
c/compilerengine.cpp) is based on the code from the clang-interpreter
example
(http://llvm.org/viewvc/llvm-project/cfe/trunk/examples/clang-interpreter/ma
in.cpp?view=co). The idea is therefore to create an instance of the
'compiler' class and, among other things, use its compileCode() whenever
needed.

 

All of this works perfectly fine using LLVM+Clang 3.2, but not with
LLVM+Clang 3.3. At least, not out of the box. Basically, the first time I
try to compile some C code, everything is fine. However, if I try to compile
some other C code, then I will get the following error message:

 

clang: for the -vectorize-loops option: may only occur zero or one times! 

 

FWIW, I use the -O3 option which I understand makes implicit use of the loop
vectorizer, and that seems to be the problem. Indeed, if I disable the loop
vectorizer (using the -fno-vectorize option), then everything is fine but I
clearly don't want to have to do that. So, right now, my solution consists
of commenting out the test for the number of option occurrences in the
Option::addOccurrence() method
(http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Support/CommandLine.cpp?
view=co), but it is clearly not satisfactory either (even though it's good
enough for my particular purpose).

 

So, would anyone have any idea about what I might be missing, if
anything?... 

 

Cheers, Alan.

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20130620/4ab9324e/attachment.html>

Hi,

I can confirm this problem. We use clang's libtooling ToolInvocation object
and the second module fails to compile even through it's a completely fresh
instance.

Adding -fno-vectorize explicitly seems to work around this problem, but
does anyone know a real fix?

ciao,
Mario

Am 20.06.13 15:07, schrieb Alan Garny:
> My application generates some C code which I then compile using LLVM+Clang
before running it. For the compilation side of things, I have a class which one
of the methods (the CompilerEngine::compileCode() method in
https://raw.github.com/opencor/opencor/llvm_3.3/src/plugins/misc/Compiler/sr
c/compilerengine.cpp) is based on the code from the clang-interpreter example
(http://llvm.org/viewvc/llvm-project/cfe/trunk/examples/clang-interpreter/ma
in.cpp?view=co). The idea is therefore to create an instance of the
'compiler' class and, among other things, use its compileCode() whenever
needed.
> 
> All of this works perfectly fine using LLVM+Clang 3.2, but not with
LLVM+Clang 3.3. At least, not out of the box. Basically, the first time I try to
compile some C code, everything is fine. However, if I try to compile some other
C code, then I will get the following error message:
> 
> clang: for the -vectorize-loops option: may only occur zero or one times!
> 
> FWIW, I use the -O3 option which I understand makes implicit use of the
loop vectorizer, and that seems to be the problem. Indeed, if I disable the loop
vectorizer (using the -fno-vectorize option), then everything is fine but I
clearly don't want to have to do that. So, right now, my solution consists
of commenting out the test for the number of option occurrences in the
Option::addOccurrence() method
(http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Support/CommandLine.cpp?
view=co), but it is clearly not satisfactory either (even though it's good
enough for my particular purpose).
> 
> So, would anyone have any idea about what I might be missing, if
anything?...
> 
> Cheers, Alan.

Thanks Mario, it's good (so to speak!) to know that I am not the only one
experiencing this problem!

I have had another look at the LLVM code and, as far as I can tell, it would
seem that it's related to llvm::cl:: RegisteredOptionList not being properly
reset or something. Anyway, I want to be able to take advantage of the loop
vectorizer, so I have decided to comment out the test for the number of
option occurrences in the Option::addOccurrence() method
(http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Support/CommandLine.cpp?
view=co). Not nice, but it seems to be doing the job for me and, unless I am
told/shown otherwise, there doesn't seem to be any other solution, so...

Alan
> -----Original Message-----
> From: Mario Schwalbe [mailto:m3o.s6e at googlemail.com]
> Sent: 21 June 2013 10:32
> To: Alan Garny
> Cc: LLVM Dev; Clang Dev; Mario Schwalbe
> Subject: Re: [LLVMdev] LLVM+Clang 3.3: clang: for the -vectorize-loops
option:
> may only occur zero or one times!
> 
> Hi,
> 
> I can confirm this problem. We use clang's libtooling ToolInvocation
object
> and the second module fails to compile even through it's a completely
fresh
> instance.
> 
> Adding -fno-vectorize explicitly seems to work around this problem, but
does
> anyone know a real fix?
> 
> ciao,
> Mario
> 
> Am 20.06.13 15:07, schrieb Alan Garny:
> > My application generates some C code which I then compile using
> LLVM+Clang before running it. For the compilation side of things, I have a
> class which one of the methods (the CompilerEngine::compileCode() method
> in
> https://raw.github.com/opencor/opencor/llvm_3.3/src/plugins/misc/Compiler
> /sr c/compilerengine.cpp) is based on the code from the clang-interpreter
> example (http://llvm.org/viewvc/llvm-project/cfe/trunk/examples/clang-
> interpreter/ma in.cpp?view=co). The idea is therefore to create an
instance of
> the 'compiler' class and, among other things, use its compileCode()
whenever
> needed.
> >
> > All of this works perfectly fine using LLVM+Clang 3.2, but not with
> LLVM+Clang 3.3. At least, not out of the box. Basically, the first time I
try to
> compile some C code, everything is fine. However, if I try to compile some
> other C code, then I will get the following error message:
> >
> > clang: for the -vectorize-loops option: may only occur zero or one
times!
> >
> > FWIW, I use the -O3 option which I understand makes implicit use of
the
> loop vectorizer, and that seems to be the problem. Indeed, if I disable
the
> loop vectorizer (using the -fno-vectorize option), then everything is fine
but I
> clearly don't want to have to do that. So, right now, my solution
consists
of
> commenting out the test for the number of option occurrences in the
> Option::addOccurrence() method (http://llvm.org/viewvc/llvm-
> project/llvm/trunk/lib/Support/CommandLine.cpp? view=co), but it is
clearly
> not satisfactory either (even though it's good enough for my particular
> purpose).	
> >
> > So, would anyone have any idea about what I might be missing, if
> anything?...
> >
> > Cheers, Alan.