Ross/Robert,
Thank you so much for your help/assistance.
I managed to install PMEMD successfully in SMP IA64 with changing the intel MKL lib file from libmkl_intel.a to libmkl_intel_lp64.a in the config.h
Again, thank you so much. I really appreciate both of your help. ;p
Cheers !
Vijay Manickam Achari
(Phd Student c/o Prof Rauzah Hashim)
Chemistry Department,
University of Malaya,
Malaysia
vjramana.gmail.com
--- On Wed, 8/4/09, Ross Walker <ross.rosswalker.co.uk> wrote:
> From: Ross Walker <ross.rosswalker.co.uk>
> Subject: RE: [AMBER] pmemdcompilation/makefailed-mkl_vml_service_threade_z_1i_c_1o
> To: "'AMBER Mailing List'" <amber.ambermd.org>
> Date: Wednesday, 8 April, 2009, 10:53 PM
> Done, sorry about this. I had it correct for everything else
> (em64t / opteron64 / ia32) I believe. I couldn't find an
> IA64 machine with anything later than MKL9.0 to test on
> unfortunately and this is largely where the problems lie.
> The degrees of freedom are now so huge :-(.
>
> All the best
> Ross
>
> -----Original Message-----
> From: amber-bounces.ambermd.org
> [mailto:amber-bounces.ambermd.org] On Behalf Of Robert Duke
> Sent: Wednesday, April 08, 2009 6:36 AM
> To: AMBER Mailing List
> Subject: Re: [AMBER]
> pmemdcompilation/makefailed-mkl_vml_service_threader_z_1i_c_1o
>
> Okay, I figured it out. This is all basically due to a
> typo in the
> configure script for amber 11 pmemd, which was never
> checked out. Oh, oops.
> Vijay, you simply need to change the reference to
> libmkl_intel.a (there is
> no such thing) to libmkl_intel_lp64.a. Ross, can you
> please fix this in the
> checked-in amber11 source and wherever else this configure
> script might have
> gotten propagated to. The change applies only to itanium
> 2, not the other
> intel architectures, where we got this right. Vijay, sorry
> it took us so
> long to figure out what is going on here. Too many
> machines, combined with
> too many version-dependent linkage options for mkl.
> Best Regards - Bob Duke
>
> ----- Original Message -----
> From: "Vijay Manickam Achari"
> <vjrajamany.yahoo.com>
> To: "AMBER Mailing List"
> <amber.ambermd.org>
> Sent: Wednesday, April 08, 2009 2:27 AM
> Subject: Re: [AMBER]
> pmemdcompilation/makefailed-mkl_vml_service_threader_z_1i_c_1o
>
>
> Robert,
> The same errors again.
> Attached is the updated config.h script.
>
> ifort -c -auto -IPF_fma -ftz -ip -O3 pmemd.f90
> /lib/cpp -traditional -P -DPUBFFT -DMPI
> -DSLOW_INDIRECTVEC -DMKL -DGB_IA64_ACCEL
> erfcfun.fpp erfcfun.f90
> ifort -c -auto -IPF_fma -ftz -ip -O3 erfcfun.f90
> ifort -i-static -o pmemd gbl_constants.o gbl_datatypes.o
> state_info.o
> file_io_dat.o parallel_dat.o mdin_ctrl_dat.o
> mdin_ewald_dat.o prmtop_dat.o
> inpcrd_dat.o dynamics_dat.o img.o parallel.o pme_direct.o
> pme_recip.o
> pme_fft.o fft1d.o bspline.o pme_force.o pbc.o nb_pairlist.o
> cit.o dynamics.o
> bonds.o angles.o dihedrals.o runmd.o loadbal.o shake.o
> runmin.o
> constraints.o axis_optimize.o gb_ene.o veclib.o gb_force.o
> timers.o
> pmemd_lib.o runfiles.o file_io.o bintraj.o pmemd_clib.o
> pmemd.o random.o
> degcnt.o erfcfun.o nmr_calls..o nmr_lib.o get_cmdline.o
> master_setup.o
> alltasks_setup.o pme_setup.o ene_frc_splines.o
> xtprmtop_section.o -L/opt/intel/mkl/10.0.010/lib/64/
> -Wl,--start-group
> /opt/intel/mkl/10.0.010/lib/64/libmkl_intel.a
> /opt/intel/mkl/10.0.010/lib/64/libmkl_sequential.a
> /opt/intel/mkl/10.0.010/lib/64/libmkl_core.a
> -Wl,--end-group -lpthread -lmpi
>
> -Wl,-rpath=/opt/intel/ipp/5.3/itanium/sharedlib:/opt/intel/mkl/10.0.010/lib/64:/opt/intel/fc/10.1.008/lib:/opt/intel/cc/10.1.008/lib:/opt/intel/fc/10.1.008/lib:/opt/intel/cc/10.1.008/lib:/usr/lib:/usr/local/libgb_ene.o:
> In function
> `gb_ene_mod_mp_gb_ene_':gb_ene.f90:(.text+0x2152):
> undefined reference to
> `vdinv_'gb_ene.f90:(.text+0x22f2): undefined reference
> to `vdexp_'gb_ene.f90:(.text+0x24e2): undefined
> reference to `vdinvsqrt_'gb_ene.f90:(.text+0x2542):
> undefined reference to
> `vdinv_'gb_ene.f90:(.text+0x2672): undefined reference
> to `vdexp_'gb_ene.f90:(.text+0x2692): undefined
> reference to `vdinvsqrt_'gb_ene.f90:(.text+0x4e72):
> undefined reference to
> `vdinvsqrt_'gb_ene.f90:(.text+0x5612): undefined
> reference to `vdinv_'gb_ene.f90:(.text+0x5632):
> undefined reference to
> `vdinv_'gb_ene.f90:(.text+0x5802): undefined reference
> to `vdln_'gb_ene.o: In function
> `gb_ene_mod_mp_calc_born_radii_':gb_ene.f90:(.text+0xe152):
> undefined reference to
> `vdinvsqrt_'gb_ene.f90:(.text+0xefd2): undefined
> reference to `vdinv_'gb_ene.f90:(.text+0xeff2):
> undefined reference to
> `vdinv_'gb_ene.f90:(.text+0x105d2): undefined reference
> to `vdln_'gb_ene.f90:(.text+0x105f2): undefined
> reference to `vdln_'make[1]: *** [pmemd] Error 1make[1]:
> Leaving directory `/usr/local/amber9/src/pmemd/src'make:
> *** [all] Error 2ptmlxsmp:/usr/local/amber9/src/pmemd #Vijay
> Manickam Achari(Phd Student c/o Prof Rauzah Hashim)Chemistry
> Department,University of Malaya,Malaysia
> vjramana.gmail.com--- On Tue, 7/4/09, Robert Duke
> <rduke.email.unc.edu> wrote:> From: Robert Duke
> <rduke.email.unc.edu>> Subject: Re: [AMBER]
> pmemdcompilation/makefailed-mkl_vml_service_threader_z_1i_c_1o>
> To: "AMBER Mailing List"
> <amber.ambermd.org>> Date: Tuesday, 7 April, 2009,
> 8:55 PM> Oops. Looking at the script again, you would
> actually want> to set $MKL_HOME to
> /opt/intel/mkl/10.0.010 as the lib/64/> part is already
> in the library names you list.> Regards - Bob> -----
> Original Message ----- From: "Robert Duke">
> <rduke.email.unc.edu>> To:
> <vjrajamany.yahoo.com>; "AMBER Mailing>
> List" <amber.ambermd.org>> Sent: Tuesday,
> April 07, 2009 8:43 AM> Subject: Re: [AMBER] pmemd>
> compilation/makefailed-mkl_vml_service_threader_z_1i_c_1o>>>
> > Looks to me like you did not substitute the actual>
> location of the math kernel libraries for $MKL_HOME in
> the> config.h script. You must plug in this value. You
> could> also place MKL_HOME =
> /opt/intel/mkl/10.0.010/lib/64 in a> line above the line
> defining MATH_LIBS, and then change all> references to
> $MKL_HOME to $(MKL_HOME).> > Regards - Bob Duke>
> >> > ----- Original Message ----- From:
> "Vijay> Manickam Achari"
> <vjrajamany.yahoo.com>> > To: "AMBER
> Mailing List"> <amber.ambermd.org>> >
> Sent: Tuesday, April 07, 2009 12:25 AM> > Subject: Re:
> [AMBER] pmemd compilation/make>
> failed-mkl_vml_service_threader_z_1i_c_1o> >>
> >> >> > Robert,> > I have altered the
> MATH_LIBS param. the config.h with> your given suggestion
> as below:> > MATH_LIBS = -L$MKL_HOME/lib/64/
> -Wl,--start-group> $MKL_HOME/lib/64/libmkl_intel.a>
> $MKL_HOME/lib/64/libmkl_sequential.a>
> $MKL_HOME/lib/64/libmkl_core.a -Wl,--end-group -lpthread>
> >> > But still errors occured as below:> >
> ifort -c -auto -IPF_fma -ftz -ip -O3 erfcfun.f90> >
> ifort -i-static -o pmemd gbl_constants.o> gbl_datatypes.o
> state_info.o file_io_dat.o parallel_dat.o>
> mdin_ctrl_dat.o mdin_ewald_dat.o prmtop_dat.o
> inpcrd_dat.o> dynamics_dat.o img.o parallel.o
> pme_direct.o pme_recip.o> pme_fft.o fft1d.o bspline.o
> pme_force.o pbc.o nb_pairlist.o> cit.o dynamics.o bonds.o
> angles.o dihedrals.o runmd.o> loadbal.o shake.o runmin.o
> constraints.o axis_optimize.o> gb_ene.o veclib.o
> gb_force.o timers.o pmemd_lib.o runfiles.o> file_io.o
> bintraj.o pmemd_clib.o pmemd.o random.o degcnt.o>
> erfcfun.o nmr_calls..o nmr_lib.o get_cmdline.o>
> master_setup.o alltasks_setup.o pme_setup.o>
> ene_frc_splines.o prmtop_section.o -LKL_HOME/lib/64/>
> -Wl,--start-group KL_HOME/lib/64/libmkl_intel.a>
> KL_HOME/lib/64/libmkl_sequential.a>
> KL_HOME/lib/64/libmkl_core.a -Wl,--end-group -lpthread
> -lmpi> >>
> -Wl,-rpath=/opt/intel/ipp/5.3/itanium/sharedlib:/opt/intel/mkl/10.0.010/lib/64:/opt/intel/fc/10.1.008/lib:/opt/intel/cc/10.1.008/lib:/opt/intel/fc/10.1.008/lib:/opt/intel/cc/10.1.008/lib:/usr/lib:/usr/local/libgb_ene.o:>
> In function>
> `gb_ene_mod_mp_gb_ene_':gb_ene.f90:(.text+0x2152):>
> undefined reference to>
> `vdinv_'gb_ene.f90:(.text+0x22f2): undefined
> reference> to `vdexp_'gb_ene.f90:(.text+0x24e2):
> undefined> reference to
> `vdinvsqrt_'gb_ene.f90:(.text+0x2542):> undefined
> reference to> `vdinv_'gb_ene.f90:(.text+0x2672):
> undefined reference> to
> `vdexp_'gb_ene.f90:(.text+0x2692): undefined>
> reference to `vdinvsqrt_'gb_ene.f90:(.text+0x4e72):>
> undefined reference to>
> `vdinvsqrt_'gb_ene.f90:(.text+0x5612): undefined>
> reference to `vdinv_'gb_ene.f90:(.text+0x5632):>
> undefined reference to>
> `vdinv_'gb_ene.f90:(.text+0x5802): undefined
> reference> to `vdln_'gb_ene.o: In function>
> `gb_ene_mod_mp_calc_born_radii_':gb_ene.f90:(.text+0xe152):>
> undefined reference to>
> `vdinvsqrt_'gb_ene.f90:(.text+0xefd2): undefined>
> reference to `vdinv_'gb_ene.f90:(.text+0xeff2):>
> undefined reference to>
> `vdinv_'gb_ene.f90:(.text+0x105d2): undefined
> reference> to `vdln_'gb_ene.f90:(.text+0x105f2):
> undefined> reference to `vdln_'make[1]: *** [pmemd]
> Error 1make[1]:> Leaving directory
> `/usr/local/amber9/src/pmemd/src'make:> *** [install]
> Error 2Thanks for your assistance.Vijay> Manickam
> Achari(Phd Student c/o Prof Rauzah Hashim)Chemistry>
> Department,University of Malaya,Malaysia>
> vjramana.gmail.com--- On Mon, 6/4/09, Robert Duke>
> <rduke.email.unc.edu> wrote:> From: Robert Duke>
> <rduke.email.unc.edu>> Subject: Re: [AMBER]
> pmemd> compilation/makefailed>
> -mkl_vml_service_threader_z_1i_c_1o> To: "AMBER>
> Mailing List" <amber.ambermd.org>> Date:>
> Monday, 6 April, 2009, 10:25 PM> Okay, I don't
> have> one of these machines handy, but> there are two
> problems> here, at least:> 1) you specified the use of
> fftw, but> you don't have it> (or at least not the
> necessary> info to find the libraries)> 2) you
> specified MKL - not a> problem in itself - but with>
> MKL 10, the library layout> has!> > been co>
> > mpletely whacked by> our dear friends at intel,
> and> we need to change the library> linkage
> info.>> So> first of all, fftw really does not help
> much, if at all,>> especially on more than one
> processor, so just configure> for> my public
> fft's:>> ./configure sgi_altix> ifort mpi
> pubfft>> Secondly, we need to change the mkl>
> linkage info (you could> also just opt to not use
> mkl,> and as long as you are not> running generalized
> Born> simulations, it won't matter> much - you
> decide not> to use mkl as part of the>
> "configure"> dialog):>> Anyway, best
> current guess for mkl linkage> info, assuming> you
> want MKL, is:>> MATH_LIBS => -L$MKL_HOME/lib/64/
> -Wl,--start-group>>
> $MKL_HOME/lib/64/libmkl_intel.a>>
> $MKL_HOME/lib/64/libmkl_sequential.a>>
> $MKL_HOME/lib/64/libmkl_core.a -Wl,--end-group>
> -lpthread>> (so you need to substitute in this
> line> for the $MKL_HOME> variable, and stick the
> result in> config.h in place of the> current
> definition of> MATH_LIBS)>> I have no reason to
> believe that the 3> lines you added at> the end of
> config.h are doing you any> good.>> Regards - Bob
> Duke>> ----- Original> Message ----- From:
> "Vijay Manickam> Achari">
> <vjrajamany.yahoo.com>> To: "Amber mailing>
> List"> <AMBER.ambermd.org>> Sent:
> Monday,> April 06, 2009 3:27 AM> Subject: [AMBER]
> pmemd> compilation/make failed>>
> -mkl_vml_service_threader_z_1i_c_1o>>> Hi,> I
> am> configuring pmemd in SMP machine SGI Altix 4700>
> itanium> 128 cores (SLES 10.2 IA64 Intel compilers &
> MKL>> version 10.0.010) as below:>
> /usr/local/amber9/src/pmemd> # ./configure sgi_altix
> ifort> mpi fftw>> But the> compilation failed,
> as errors below:>>> fft1d.f90:(.text+0x8c2):
> undefined reference to>> `dfftw_destroy_plan_'>
> fft1d.o: In function>
> `fft1d_mod_mp_fft1d_create_':>>
> fft1d.f90:(.text+0xbc2): undefined reference to>>
> `dfftw_plan_dft_1d_'> fft1d.f90:(.text+0xbf2):>
> undefined reference
> to>>`dfftw_plan_dft_1d_'>/opt/intel/mkl/10.0.010/lib/64/libmkl_intel_thread.a(vml_threading_templates.o):>>
> In function
> `mkl_vml_service_threader_z_1i_c_1o':>>
> _tmp/treading_linux64_intel_thread/vml_threading_templ!>
> > ates.c:(> > .text+0x92):> undefined reference
> to>`__kmpc_global_thread_num'>/opt/intel/mkl/10.0.010/lib/64/libmkl_intel_thread.a(vml_threading_templates.o):>>
> In
> function>>`L_mkl_vml_service_threader_z_1i_c_1o_2325__par_loop0_2$0':>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0x442):>>
> undefined reference
> to>`__kmpc_ok_to_fork'>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0x472):>>
> undefined reference
> to>`__kmpc_push_num_threads'>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0x4e2):>>
> undefined reference
> to>`__kmpc_fork_call'>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0x502):>>
> undefined reference
> to>`__kmpc_serialized_parallel'>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0x582):>>
> undefined reference
> to>`__kmpc_end_serialized_parallel'>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0x6f2):>>
> undefined reference
> to>`__kmpc_for_static_init_4'>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0x852):>>
> undefined reference
> to>`__kmpc_for_static_fini'>/opt/intel/mkl/10.0.010/lib/64/libmkl_intel_thread.a(vml_threading_templates.o):>>
> In
> function>`mkl_vml_service_threader_z_1i_2o':>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0x9c2):>>
> undefined reference
> to>`__kmpc_global_thread_num'>/opt/intel/mkl/10.0.010/lib/64/libmkl_intel_thread.a(vml_threading_templates.o):>>
> In
> function>>`L_mkl_vml_service_threader_z_1i_2o_2299__par_loop1_2$1':>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0xd52):>>
> undefined reference
> to>`__kmpc_ok_to_fork'>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0xd82):>>
> undefined reference
> to>`__kmpc_push_num_threads'>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0xdf2):>>
> undefined reference
> to>`__kmpc_fork_call'>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0xe12):>>
> undefined reference to>
> `__kmpc_serialized_parallel'>_tmp/treading_linux64_intel_thread/vm!>
> > l_thread> > ing_templates.c:(.text+0xe92):>
> undefined reference>
> to>`__kmpc_end_serialized_parallel'>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0xff2):>>
> undefined reference
> to>`__kmpc_for_static_init_4'>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0x1162):>>
> undefined reference
> to>`__kmpc_for_static_fini'>/opt/intel/mkl/10.0.010/lib/64/libmkl_intel_thread.a(vml_threading_templates.o):>>
> In
> function>`mkl_vml_service_threader_z_2i_1o':>_tmp/treading_linux64_intel_thread/vml_threading_templates.c:(.text+0x1282):>>
> undefined reference
> to>`__kmpc_global_thread_num'>/opt/intel/mkl/10.0.010/lib/64/libmkl_intel_thread.a(vml_threading_templates.o):>>
> In function>>
> `L_mkl_vml_service_threader_z_2i_1o_2273__par_loop2_2$2':>>>
> Attached is the config.h file for your review.
> Please>> assist me in this. Thanks.>> Vijay
> Manickam Achari>> (Phd Student c/o Prof Rauzah
> Hashim)> Chemistry> Department,> University of
> Malaya,> Malaysia>>
> vjramana.gmail.com>>>>>>> >>
> _______________________________________________> >>
> AMBER mailing list> > AMBER.ambermd.org> >>
> http://lists.ambermd.org/mailman/listinfo/amber>>
> >>>>>
> _______________________________________________>
> AMBER> mailing list>
> AMBER.ambermd.org>>http://lists.ambermd.org/mailman/listinfo/amber_______________________________________________AMBER>
> mailing>
> listAMBER.ambermd.orghttp://lists.ambermd.org/mailman/listinfo/amber>
> >> >
> _______________________________________________> >
> AMBER mailing list> > AMBER.ambermd.org> >
> http://lists.ambermd.org/mailman/listinfo/amber>
> >>>>
> _______________________________________________> AMBER
> mailing list> AMBER.ambermd.org>
> http://lists.ambermd.org/mailman/listinfo/amber-------------------------------------------------------------------------------->
> _______________________________________________> AMBER
> mailing list> AMBER.ambermd.org>
> http://lists.ambermd.org/mailman/listinfo/amber>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Apr 10 2009 - 01:14:19 PDT
