I've added these tips to
https://www.gridpp.ac.uk/wiki/Benchmarking_procedure#What_is_HEPSPEC06.2C_and_how_do_I_use_it
BTW: Just going through it and it seems to work with the "compat and
devel rpms" and Alessandra's fix. Thanks for that.
Ste
On 02/04/2019 19:41, Alessandra Forti wrote:
> I've added the needed compat and devel rpms and added to the CXX line
> in linux32-gcc_cern.cfg
>
> -include cstddef *-std=c++11**
> *
> and it compiled everything now.
>
> cheers
> alessandra
>
> On 02/04/2019 15:38, Pete Gronbech wrote:
>> Requesting 1.2 is free to those already with a licence.
>> Pete
>>
>> --
>> ----------------------------------------------------------------------
>> Peter Gronbech
>>
>> Head of IT
>> Department of Physics,
>> University of Oxford,
>> Denys Wilkinson Building, Tel No. : 01865 272233
>> Keble Road, Oxford OX1 3RH, UK E-mail :[log in to unmask]
>> ----------------------------------------------------------------------
>>
>>
>>
>> -----Original Message-----
>> From: Testbed Support for GridPP member institutes<[log in to unmask]> On Behalf Of Alessandra Forti
>> Sent: 02 April 2019 14:54
>> To:[log in to unmask]
>> Subject: Re: HS06 results
>>
>> Hi,
>>
>> I still have 1.1 but 32bit still fails for me even with these changes.
>>
>> cheers
>> alessandra
>>
>>
>>
>> On 28/03/2019 10:16, Daniel Traynor wrote:
>>> As Simon says adding
>>>
>>> -include cstddef" to your CXXPORTABILITY
>>>
>>> enables 64bit compilation
>>>
>>> Also installing (as stated in
>>> https://www.gridpp.ac.uk/wiki/Benchmarking_procedure)
>>> glibc-devel.i686
>>> compat-libstdc++-33.i686
>>> libstdc++.i686
>>> libstdc++-devel.i686
>>>
>>> Seems to be enough to get 32 bit compilation up and running (note
>>> compat-libstdc++-296.i686 is not available in cenots7)
>>>
>>> for comparison
>>>
>>> for Intel Xeon gold 5118 (Dell R440 bios option set to performance per
>>> watt OS)
>>> 32 bit SL6 = ~575
>>> 32 bit centos7 = ~590
>>> 64 bit cenots7 = ~660
>>>
>>> * Dr Daniel Traynor, Grid cluster system manager
>>> * Tel +44(0)20 7882 6560, Particle Physics,QMUL
>>>
>>> ________________________________________
>>> From: Testbed Support for GridPP member institutes
>>> <[log in to unmask]> on behalf of Martin Bly - UKRI STFC
>>> <[log in to unmask]>
>>> Sent: 26 March 2019 22:20
>>> To:[log in to unmask]
>>> Subject: Re: HS06 results
>>>
>>>
>>> SPEC CPU2006 v1.2 is *required* for SL/CentOS7 - it fixes all the compile problems with v1.1.
>>>
>>> Martin.
>>> --
>>> Martin Bly
>>> RAL Tier1 Fabric Manager
>>>
>>>> -----Original Message-----
>>>> From: Testbed Support for GridPP member institutes [mailto:TB-
>>>> [log in to unmask]] On Behalf Of Simon Fayer
>>>> Sent: Tuesday, March 26, 2019 7:02 PM
>>>> To:[log in to unmask]
>>>> Subject: Re: HS06 results
>>>>
>>>> Hi Dan,
>>>>
>>>> Yes, I think this is almost certainly fixed in the newer spec version
>>>> (although I can't find any mention of it in the changelog). As a
>>>> workaround you could try adding "-include cstddef" to your
>>>> CXXPORTABILITY, which I think is the right header for the ptrdiff_t type.
>>>>
>>>> Regards,
>>>> Simon
>>>>
>>>>
>>>> On Tue, Mar 26, 2019 at 04:58:41PM +0000, Daniel Traynor wrote:
>>>>> For me i'm unable to compile dealII package. error below on centos7.6.1810.
>>>> it's the same with 32 or 64 bit
>>>>> my config file has
>>>>>
>>>>> 447.dealII=default=default=default:
>>>>> CXXPORTABILITY= -DSPEC_CPU_LINUX -include cstdlib -include cstring
>>>>>
>>>>>
>>>>> ./runspec.sh.new -a 64 -d "64 bit CentOS7 " -w -f $OUTPUT_DIR
>>>>>
>>>>> I note I'm using SPEC2006_v11.tar.bz2, but if i can get hold of an
>>>>> updated
>>>> version I can give it a try
>>>>> dan
>>>>>
>>>>> ....
>>>>> g++ -c -o block_vector.o -DSPEC_CPU -DNDEBUG -Iinclude -
>>>> DBOOST_DISABLE_THREADS -Ddeal_II_dimension=3 -O2 -fPIC -pthread -
>>>> DSPEC_CPU_LP64 -DSPEC_CPU_LINUX -include cstdlib -include cstring
>>>> block_vector.cc
>>>>> In file included from include/lac/block_vector.templates.h:19:0,
>>>>> from block_vector.cc:14:
>>>>> include/lac/block_vector.h:227:17: error: 'ptrdiff_t' does not name a type
>>>>> typedef ptrdiff_t difference_type;
>>>>> ^
>>>>> include/lac/block_vector.h:333:38: error: 'difference_type' does not
>>>>> name a
>>>> type
>>>>> reference operator [] (const difference_type d) const;
>>>>> ^
>>>>> include/lac/block_vector.h:333:38: note: (perhaps 'typename
>>>> std::iterator<std::random_access_iterator_tag, typename
>>>> internal::BlockVectorIterators::Types<number,
>>>> constness>::NumberType>::difference_type' was intended)
>>>>> include/lac/block_vector.h:333:54: error: ISO C++ forbids declaration of 'd'
>>>> with no type [-fpermissive]
>>>>> reference operator [] (const difference_type d) const;
>>>>> ^
>>>>> include/lac/block_vector.h:469:9: error: 'difference_type' does not
>>>>> name a
>>>> type
>>>>> difference_type operator - (const Iterator &i) const;
>>>>> ^
>>>>> include/lac/block_vector.h:469:9: note: (perhaps 'typename
>>>> std::iterator<std::random_access_iterator_tag, typename
>>>> internal::BlockVectorIterators::Types<number,
>>>> constness>::NumberType>::difference_type' was intended)
>>>>> include/lac/block_vector.h:475:9: error: 'difference_type' does not
>>>>> name a
>>>> type
>>>>> difference_type operator - (const InverseConstnessIterator &i) const;
>>>>> ^
>>>>> include/lac/block_vector.h:475:9: note: (perhaps 'typename
>>>> std::iterator<std::random_access_iterator_tag, typename
>>>> internal::BlockVectorIterators::Types<number,
>>>> constness>::NumberType>::difference_type' was intended)
>>>>> include/lac/block_vector.h:482:36: error: 'difference_type' does not
>>>>> name a
>>>> type
>>>>> Iterator operator + (const difference_type &d) const;
>>>>> ^
>>>>> include/lac/block_vector.h:482:36: note: (perhaps 'typename
>>>> std::iterator<std::random_access_iterator_tag, typename
>>>> internal::BlockVectorIterators::Types<number,
>>>> constness>::NumberType>::difference_type' was intended)
>>>>> include/lac/block_vector.h:482:53: error: ISO C++ forbids declaration of 'd'
>>>> with no type [-fpermissive]
>>>>> Iterator operator + (const difference_type &d) const;
>>>>> ^
>>>>> include/lac/block_vector.h:489:36: error: 'difference_type' does not
>>>>> name a
>>>> type
>>>>> Iterator operator - (const difference_type &d) const;
>>>>> ^
>>>>> include/lac/block_vector.h:489:36: note: (perhaps 'typename
>>>> std::iterator<std::random_access_iterator_tag, typename
>>>> internal::BlockVectorIterators::Types<number,
>>>> constness>::NumberType>::difference_type' was intended)
>>>>> include/lac/block_vector.h:489:53: error: ISO C++ forbids declaration of 'd'
>>>> with no type [-fpermissive]
>>>>> Iterator operator - (const difference_type &d) const;
>>>>> ^
>>>>> include/lac/block_vector.h:496:39: error: 'difference_type' does not
>>>>> name a
>>>> type
>>>>> Iterator & operator += (const difference_type &d);
>>>>> ^
>>>>> include/lac/block_vector.h:496:39: note: (perhaps 'typename
>>>> std::iterator<std::random_access_iterator_tag, typename
>>>> internal::BlockVectorIterators::Types<number,
>>>> constness>::NumberType>::difference_type' was intended)
>>>>> include/lac/block_vector.h:496:56: error: ISO C++ forbids declaration of 'd'
>>>> with no type [-fpermissive]
>>>>> Iterator & operator += (const difference_type &d);
>>>>> ^
>>>>> include/lac/block_vector.h:503:39: error: 'difference_type' does not
>>>>> name a
>>>> type
>>>>> Iterator & operator -= (const difference_type &d);
>>>>> ^
>>>>> include/lac/block_vector.h:503:39: note: (perhaps 'typename
>>>> std::iterator<std::random_access_iterator_tag, typename
>>>> internal::BlockVectorIterators::Types<number,
>>>> constness>::NumberType>::difference_type' was intended)
>>>>> include/lac/block_vector.h:503:56: error: ISO C++ forbids declaration of 'd'
>>>> with no type [-fpermissive]
>>>>> Iterator & operator -= (const difference_type &d);
>>>>> ^
>>>>> include/lac/block_vector.h:1494:52: error: 'difference_type' does
>>>>> not name a
>>>> type
>>>>> Iterator<number,constness>::operator [] (const difference_type d) const
>>>>> ^
>>>>> include/lac/block_vector.h:1494:52: note: (perhaps 'typename
>>>> std::iterator<std::random_access_iterator_tag, typename
>>>> internal::BlockVectorIterators::Types<number,
>>>> constness>::NumberType>::difference_type' was intended)
>>>>> include/lac/block_vector.h:1494:68: error: ISO C++ forbids declaration of 'd'
>>>> with no type [-fpermissive]
>>>>> Iterator<number,constness>::operator [] (const difference_type d) const
>>>>>
>>>>> ^
>>>>> include/lac/block_vector.h:1727:5: error: prototype for 'typename
>>>> internal::BlockVectorIterators::Iterator<number,
>>>> constness>::difference_type
>>>> internal::BlockVectorIterators::Iterator<number,
>>>> constness>::operator-(const internal::BlockVectorIterators::Iterator<number, constness>&) const' does not match any in class 'internal::BlockVectorIterators::Iterator<number, constness>'
>>>>> Iterator<number,constness>::
>>>>> ^
>>>>> include/lac/block_vector.h:489:18: error: candidate is:
>>>> internal::BlockVectorIterators::Iterator<number, constness>
>>>> internal::BlockVectorIterators::Iterator<number,
>>>> constness>::operator-(const
>>>> int&) const
>>>>> Iterator operator - (const difference_type &d) const;
>>>>> ^
>>>>> include/lac/block_vector.h:1741:5: error: prototype for 'typename
>>>> internal::BlockVectorIterators::Iterator<number,
>>>> constness>::difference_type
>>>> internal::BlockVectorIterators::Iterator<number,
>>>> constness>::operator-(const
>>>> InverseConstnessIterator&) const' does not match any in class
>>>> 'internal::BlockVectorIterators::Iterator<number, constness>'
>>>>> Iterator<number,constness>::
>>>>> ^
>>>>> include/lac/block_vector.h:489:18: error: candidate is:
>>>> internal::BlockVectorIterators::Iterator<number, constness>
>>>> internal::BlockVectorIterators::Iterator<number,
>>>> constness>::operator-(const
>>>> int&) const
>>>>> Iterator operator - (const difference_type &d) const;
>>>>> ^
>>>>> include/lac/block_vector.h:1756:23: error: 'difference_type' does
>>>>> not name a
>>>> type
>>>>> operator + (const difference_type &d) const
>>>>> ^
>>>>> include/lac/block_vector.h:1756:23: note: (perhaps 'typename
>>>> std::iterator<std::random_access_iterator_tag, typename
>>>> internal::BlockVectorIterators::Types<number,
>>>> constness>::NumberType>::difference_type' was intended)
>>>>> include/lac/block_vector.h:1756:40: error: ISO C++ forbids declaration of 'd'
>>>> with no type [-fpermissive]
>>>>> operator + (const difference_type &d) const
>>>>> ^
>>>>> include/lac/block_vector.h:1781:23: error: 'difference_type' does
>>>>> not name a
>>>> type
>>>>> operator - (const difference_type &d) const
>>>>> ^
>>>>> include/lac/block_vector.h:1781:23: note: (perhaps 'typename
>>>> std::iterator<std::random_access_iterator_tag, typename
>>>> internal::BlockVectorIterators::Types<number,
>>>> constness>::NumberType>::difference_type' was intended)
>>>>> include/lac/block_vector.h:1781:40: error: ISO C++ forbids declaration of 'd'
>>>> with no type [-fpermissive]
>>>>> operator - (const difference_type &d) const
>>>>> ^
>>>>> include/lac/block_vector.h:1806:24: error: 'difference_type' does
>>>>> not name a
>>>> type
>>>>> operator += (const difference_type &d)
>>>>> ^
>>>>> include/lac/block_vector.h:1806:24: note: (perhaps 'typename
>>>> std::iterator<std::random_access_iterator_tag, typename
>>>> internal::BlockVectorIterators::Types<number,
>>>> constness>::NumberType>::difference_type' was intended)
>>>>> include/lac/block_vector.h:1806:41: error: ISO C++ forbids declaration of 'd'
>>>> with no type [-fpermissive]
>>>>> operator += (const difference_type &d)
>>>>> ^
>>>>> include/lac/block_vector.h:1834:24: error: 'difference_type' does
>>>>> not name a
>>>> type
>>>>> operator -= (const difference_type &d)
>>>>> ^
>>>>> include/lac/block_vector.h:1834:24: note: (perhaps 'typename
>>>> std::iterator<std::random_access_iterator_tag, typename
>>>> internal::BlockVectorIterators::Types<number,
>>>> constness>::NumberType>::difference_type' was intended)
>>>>> include/lac/block_vector.h:1834:41: error: ISO C++ forbids declaration of 'd'
>>>> with no type [-fpermissive]
>>>>> operator -= (const difference_type &d)
>>>>> ^
>>>>> specmake: *** [block_vector.o] Error 1 Error with make 'specmake
>>>>> build': check file
>>>> '/scratch/tmp/hepspec/SPEC2006_v11/benchspec/CPU2006/447.dealII/run/b
>>>> uil
>>>> d_base_x64_gcc_cern.0000/make.err'
>>>>> Error with make!
>>>>> *** Error building 447.dealII
>>>>>
>>>>>
>>>>>
>>>>> * Dr Daniel Traynor, Grid cluster system manager
>>>>> * Tel +44(0)20 7882 6560, Particle Physics,QMUL
>>>>>
>>>>> ________________________________________
>>>>> From: Testbed Support for GridPP member institutes <TB-
>>>> [log in to unmask]> on behalf of Simon Fayer
>>>> <[log in to unmask]>
>>>>> Sent: 26 March 2019 16:30
>>>>> To:[log in to unmask]
>>>>> Subject: Re: HS06 results
>>>>>
>>>>> Hi,
>>>>>
>>>>> I ran HS06 (using SPEC06 v1.2 as the base) 32-bit on CentOS 7.6.1810
>>>>> recently without any problems: What are the exact issues/error
>>>>> messages you are seeing?
>>>>>
>>>>> Regards,
>>>>> Simon
>>>>>
>>>>>
>>>>> On Mon, Mar 25, 2019 at 02:41:25PM +0000, Daniel Traynor wrote:
>>>>>> yes I've been having problems
>>>>>>
>>>>>> I can compile the dealII package.
>>>>>>
>>>>>> Also as CentOs7.6 is our first release with our new deployment
>>>>>> system going
>>>> back to an old release is difficult
>>>>>> dan
>>>>>>
>>>>>> * Dr Daniel Traynor, Grid cluster system manager
>>>>>> * Tel +44(0)20 7882 6560, Particle Physics,QMUL
>>>>>>
>>>>>> ________________________________________
>>>>>> From: Testbed Support for GridPP member institutes <TB-
>>>> [log in to unmask]> on behalf of Alastair Dewhurst
>>>> <[log in to unmask]>
>>>>>> Sent: 25 March 2019 13:46
>>>>>> To:[log in to unmask]
>>>>>> Subject: HS06 results
>>>>>>
>>>>>> Hi All
>>>>>>
>>>>>> We are currently performing our HS06 benchmarking for the new WN at
>>>>>> the
>>>> Tier-1. The hardware was the same as last year, so I wasn't
>>>> expecting any surprises.... The work was being done by our very
>>>> talented apprentice (Dominic) and he found he was unable to run the
>>>> the test using the latest versions of SL7 that we run. This was because various 32 bit packages no longer get bundled.
>>>> He 'solved' the problem by changing the flag in the test to 64 bit
>>>> which gave us a
>>>> HS06 about 100 higher per box than before (~900 rather than ~800).
>>>> In order to get a HS06 score using the 32 bit packages we have had to
>>>> roll back to an older version of SL7, which seems to be a bit silly.
>>>>>> Has anybody else run into this or a similar problem? It might be a
>>>>>> relatively
>>>> minor problem for us this year, but this is only going to get worse. Do we need
>>>> to push (again) for an updated benchmark? I know it is HEPiX this week, so if
>>>> there is anybody in San Diego in the benchmarking WG, it might be a
>>>> good time to mention this again.
>>>>>> Thanks
>>>>>>
>>>>>> Alastair
>>>>>>
>>>>>>
>>>> #################################################################
>>>> #######
>>>>>> To unsubscribe from the TB-SUPPORT list, click the following link:
>>>>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
>>>>>>
>>>>>>
>>>> #################################################################
>>>> #######
>>>>>> To unsubscribe from the TB-SUPPORT list, click the following link:
>>>>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
>>>> #################################################################
>>>> #######
>>>>> To unsubscribe from the TB-SUPPORT list, click the following link:
>>>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
>>>>>
>>>>>
>>>> #################################################################
>>>> #######
>>>>> To unsubscribe from the TB-SUPPORT list, click the following link:
>>>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
>>>> #################################################################
>>>> #######
>>>>
>>>> To unsubscribe from the TB-SUPPORT list, click the following link:
>>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
>>> ######################################################################
>>> ##
>>>
>>> To unsubscribe from the TB-SUPPORT list, click the following link:
>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
>>>
>>> ######################################################################
>>> ##
>>>
>>> To unsubscribe from the TB-SUPPORT list, click the following link:
>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
>> --
>> Respect is a rational process. \\//
>> For Ur-Fascism, disagreement is treason. (U. Eco)
>>
>> ########################################################################
>>
>> To unsubscribe from the TB-SUPPORT list, click the following link:
>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
>>
>> ########################################################################
>>
>> To unsubscribe from the TB-SUPPORT list, click the following link:
>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
>
> --
> Respect is a rational process. \\//
> For Ur-Fascism, disagreement is treason. (U. Eco)
>
> ------------------------------------------------------------------------
>
> To unsubscribe from the TB-SUPPORT list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
>
--
Steve Jones [log in to unmask]
Grid System Administrator office: 220
High Energy Physics Division tel (int): 43396
Oliver Lodge Laboratory tel (ext): +44 (0)151 794 3396
University of Liverpool http://www.liv.ac.uk/physics/hep/
########################################################################
To unsubscribe from the TB-SUPPORT list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
|