llvm dev - Nov 2010 - [LLVMdev] tot clang/llvm and tot gcc performance comparision

Hi Evan,
> Thanks David. Unfortunately many of us cannot use GPL v3 gcc so it's
hard for us
> to investigate this. One question, can you tell if gcc is inlining
significantly
> more than llvm? We have reports that this is one of the issue plaguing eon
> performance.
are you allowed to look at assembler output by GPL v3 gcc?  If so, maybe someone
else can do the compiling for you, and just send you results and assembler?

Ciao,

Duncan.

On Nov 17, 2010, at 11:14 PM, Duncan Sands wrote:
> Hi Evan,
> 
>> Thanks David. Unfortunately many of us cannot use GPL v3 gcc so
it's hard for us
>> to investigate this. One question, can you tell if gcc is inlining
significantly
>> more than llvm? We have reports that this is one of the issue plaguing
eon
>> performance.
> 
> are you allowed to look at assembler output by GPL v3 gcc?  If so, maybe
someone
> else can do the compiling for you, and just send you results and assembler?
Looking at assembly is probably ok. What you described sounds like a great
improvement but it would be even better is someone can do some in-depth
analysis.

Evan
> 
> Ciao,
> 
> Duncan.
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev

Hi Evan,
>>> Thanks David. Unfortunately many of us cannot use GPL v3 gcc so
it's hard for us
>>> to investigate this. One question, can you tell if gcc is inlining
significantly
>>> more than llvm? We have reports that this is one of the issue
plaguing eon
>>> performance.
>>
>> are you allowed to look at assembler output by GPL v3 gcc?  If so,
maybe someone
>> else can do the compiling for you, and just send you results and
assembler?
>
> Looking at assembly is probably ok. What you described sounds like a great
improvement but it would be even better is someone can do some in-depth
analysis.
it would of course be best if someone without GPL v3 restrictions could analyse
the spec results [*].  If no-one steps forward, I could set up some kind of
build server to which you can submit preprocessed source: it would compile it
using gcc-4.5 and send you back the assembler (or something along those lines).

Ciao,

Duncan.

[*] Presumably Apple management can understand the usefulness of studying the
competition.  Perhaps someone can be appointed as "GPL v3 guy",
allowed to poke
at GPL v3 software to some extent.