llvm dev - Feb 2015 - [LLVMdev] Is "clang -O1" the same as "clang -O0 + opt -O1"?

Hello,

I encounter a bug that pumped during execution of "clang -O1". However
the
bug cannot be reproduced by using "clang -O0 + opt -O1". It seems that
"clang -O1" is not the same as "clang -O0 + opt -O1".
According to the
generated LLVM IRs are large, I would like to use bugpoint with "clang
-O1"
directly instead of using "clang -O0" first and using bugpoint with
"opt
-O1". Could you tell me how I can do this?

Any suggestion is welcomed.

Thanks
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20150209/dfa9d5f3/attachment.html>

Hi,

No, it's not. Try:
  clang -O1 -mllvm -disable-llvm-optzns -S -emit-llvm

to get the IR, as it would appear to the optimizer, but without running the
optimizer itself as part of the clang invocation.

 -Hal

----- Original Message -----
> From: "Yi-Hong Lyu" <b95705030 at ntu.edu.tw>
> To: llvmdev at cs.uiuc.edu
> Sent: Sunday, February 8, 2015 8:32:32 PM
> Subject: [LLVMdev] Is "clang -O1" the same as "clang -O0 +
opt -O1"?
> 
> 
> 
> Hello,
> 
> I encounter a bug that pumped during execution of "clang -O1".
> However the bug cannot be reproduced by using "clang -O0 + opt
-O1".
> It seems that "clang -O1" is not the same as "clang -O0 +
opt -O1".
> According to the generated LLVM IRs are large, I would like to use
> bugpoint with "clang -O1" directly instead of using "clang
-O0"
> first and using bugpoint with "opt -O1". Could you tell me how I
can
> do this?
> 
> Any suggestion is welcomed.
> 
> Thanks
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
> 
-- 
Hal Finkel
Assistant Computational Scientist
Leadership Computing Facility
Argonne National Laboratory

Hello Hal,

Thansk for your reply. It seems that clang -O1 -mllvm -disable-llvm-optzns
is the same as "clang -O0 -mllvm -disable-llvm-optzns", right? Please
tell
me if I have mistake.

Although I use "clang -O0 -mllvm -disable-llvm-optzns + opt -O1", the
bug
is still not reproduced. I use "mllvm -debug" to dump the pass
manager's
arguments. It seems that the LLVM passes order is different between "clang
-O0 -mllvm -disable-llvm-optzns + opt -O1" and "clang -O1". The
bug seems
only occur when I am using "clang -O1". Is there any way to use
bugpoint
with clang (embedded bugpoint into clang)?

Thanks a lot.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20150210/2e775cd0/attachment.html>