Re: [Boost-build] Unknown options passed to Intel Linux compiler

Subject: Re: [Boost-build] Unknown options passed to Intel Linux compiler
From: K. Noel Belcourt (kbelco_at_[hidden])
Date: 2009-12-23 16:23:02

Hi Volodya,

On Dec 23, 2009, at 2:12 PM, K. Noel Belcourt wrote:

> On Dec 23, 2009, at 1:48 PM, Vladimir Prus wrote:
>
>> K. Noel Belcourt wrote:
>>
>>> Hi,
>>>
>>> It appears that both -nologo and -Winvalid-pch are being passed to
>>> the intel 11.1.059 compiler on Linux.  These options are not
>>> supported under Linux (or under Darwin).
>>>
>>> icpc: command line warning #10006: ignoring unknown option '-nologo'
>>> icpc: command line warning #10156: ignoring option '-W'; no argument
>>> required
>>>
>>> Below is a representative compile line showing the options.  Note
>>> that the -use-pch option is also invalid though there is an icpc
>>> Linux option called -pch-use, perhaps that's what someone intended?
>>>
>>> Can someone please look into where these are coming from?
>>>
>>> -- Noel
>>>
>>>
>>> intel-linux.compile.c++ /scratch/boost/results/boost/bin.v2/libs/
>>> math/
>>> test/test_roots.test/intel-linux-11.1/debug/test_roots.o
>>>
>>>      "/home/sntools/extras/compilers/intel/Compiler/11.1/059/bin/
>>> intel64/icpc" -c -xc++ -O0 -g -w1 -inline-level=0 -fPIC -nologo -
>>> Winvalid-pch -DBOOST_ALL_NO_LIB=1 -DBOOST_BUILD_PCH_ENABLED -
>>> DBOOST_REGEX_DYN_LINK=1 -DBOOST_TEST_NO_AUTO_LINK=1 -
>>> DBOOST_UBLAS_UNSUPPORTED_COMPILER=0 -I".." -I"../libs/math/test"  -
>>> use-pch"/scratch/boost/results/boost/bin.v2/libs/math/test/intel-
>>> linux-11.1/debug/pch.hpp.pchi" -c -o "/scratch/boost/results/boost/
>>> bin.v2/libs/math/test/test_roots.test/intel-linux-11.1/debug/
>>> test_roots.o" "../libs/math/test/test_roots.cpp"
>>>
>>> icpc: command line warning #10006: ignoring unknown option '-nologo'
>>> icpc: command line warning #10156: ignoring option '-W'; no argument
>>> required
>>
>> What user-config.jam is in effect on the machine where this command
>> line is generated? -nologo can only come from msvc.jam. Now, it is
>> surely a bug if -nologo appears for intel-linux, but I'd like to
>> verify this is what I think, and no some random breakage.

In case it helps, here's how I'm invoking the tests.

python run.py --proxy="http://wwwproxy.sandia.gov:80" -- 
runner="Sandia-intel-11.1-64" --bjam-toolset=intel-linux --pjl- 
toolset=gcc --toolsets=intel-11.1 --bjam-options=-j8 >&!  
intel-11.1.boost.log

-- Noel

> [ user-config.jam ]
>
> import toolset : using ;
>
> using python
>    : 2.5
>    : /home/kbelco/bin/python/linux
>    ;
>
> # /home/sntools/extras/mpi/RHEL5/openmpi-1.4.0/intel-11.1.059-64Bit/
> using mpi
>    : mpicxx
>    ;
>
> using intel
>    : 11.1
>    : /home/sntools/extras/compilers/intel/Compiler/11.1/059/bin/
> intel64/icpc
>    ;
>
> using gcc
>    ;
>
>
>
> _______________________________________________
> Unsubscribe & other changes: http://lists.boost.org/mailman/ 
> listinfo.cgi/boost-build
>