Closed hfp closed 4 years ago
Hi, I have installed cp2k-7.1 successfully using xconfigure when I implied __NO_IPI_DRIVER in my arch file, but I get one test failed when i do "make ARCH=Linux-x86-64-intelx VERSION=popt test":
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx /home/liuly/opt/cp2k-7.1/regtesting/Linux-x86-64-intelx/popt/TEST-Linux-x86-64-intelx-popt-2020-09-28_20-34-08/QS/regtest-ri-mp2/opt_basis_O_auto_gen.inp.out eps_taylor : 0.10000E-15 max_taylor : 4 ----------------------------------- OT ---------------------------------------
1 OT CG 0.80E-01 0.3 0.02365643 -15.6427980750 -1.56E+01
2 OT LS 0.32E+00 0.2 -15.6460782008
3 OT CG 0.32E+00 0.2 0.01262754 -15.6533694192 -1.06E-02
4 OT LS 0.79E+00 0.1 -15.6565454346
5 OT CG 0.79E+00 0.2 0.00467350 -15.6582921032 -4.92E-03
6 OT LS 0.81E+00 0.1 -15.6589842819
7 OT CG 0.81E+00 0.2 0.00110702 -15.6589847767 -6.93E-04
8 OT LS 0.72E+00 0.1 -15.6590183185
9 OT CG 0.72E+00 0.2 0.00029918 -15.6590189647 -3.42E-05
10 OT LS 0.91E+00 0.1 -15.6590219971
11 OT CG 0.91E+00 0.2 0.00009876 -15.6590221427 -3.18E-06
12 OT LS 0.73E+00 0.1 -15.6590224017
13 OT CG 0.73E+00 0.2 0.00003168 -15.6590224193 -2.77E-07
14 OT LS 0.73E+00 0.1 -15.6590224481
15 OT CG 0.73E+00 0.2 0.00000778 -15.6590224481 -2.88E-08
16 OT LS 0.78E+00 0.1 -15.6590224499
17 OT CG 0.78E+00 0.2 0.00000125 -15.6590224499 -1.84E-09
18 OT LS 0.98E+00 0.1 -15.6590224500
19 OT CG 0.98E+00 0.2 0.00000065 -15.6590224500 -5.97E-11
SCF run converged in 19 steps
Electronic density on regular grids: -5.9999976402 0.0000023598 Core density on regular grids: 6.0000000001 0.0000000001 Total charge density on r-space grids: 0.0000023598 Total charge density g-space grids: 0.0000023598
Overlap energy of the core charge distribution: 0.00000000000000 Self energy of the core charge distribution: -41.19899268745614 Core Hamiltonian energy: 11.60018684694388 Hartree energy: 17.13897389160868 Exchange-correlation energy: 0.00000000000000 Hartree-Fock Exchange energy: -3.19919050105590
Total energy: -15.65902244995949
outer SCF iter = 1 RMS gradient = 0.65E-06 energy = -15.6590224500 outer SCF loop converged in 1 iterations or 19 steps
Integrated absolute spin density : 2.0211945195 Ideal and single determinant S**2 : 2.000000 2.006475
Used number of processes per group: 1 Maximum allowed memory usage per MPI processes: 1024.00 MiB Available memory per MPI process for MP2: 853.00 MiB
Optimization of the auxiliary RI-MP2 basis
regtesting location summary file: /home/liuly/opt/cp2k-7.1/regtesting/Linux-x86-64-intelx/popt/TEST-Linux-x86-64-intelx-popt-2020-09-28_20-34-08/summary.txt regtesting location error_summary file: /home/liuly/opt/cp2k-7.1/regtesting/Linux-x86-64-intelx/popt/TEST-Linux-x86-64-intelx-popt-2020-09-28_20-34-08/error_summary regtesting location memory_summary file: /home/liuly/opt/cp2k-7.1/regtesting/Linux-x86-64-intelx/popt/TEST-Linux-x86-64-intelx-popt-2020-09-28_20-34-08/memory_summary regtesting location output dir: /home/liuly/opt/cp2k-7.1/regtesting/Linux-x86-64-intelx/popt/TEST-Linux-x86-64-intelx-popt-2020-09-28_20-34-08 regtesting location last dir: /home/liuly/opt/cp2k-7.1/regtesting/Linux-x86-64-intelx/popt/LAST-Linux-x86-64-intelx-popt --------------------------------- Timings -------------------------------- Plot: name="timings", title="Timing Distribution", ylabel="time [s]" PlotPoint: name="100th_percentile", plot="timings", label="100th %ile", y=88.69, yerr=0.0 PlotPoint: name="99th_percentile", plot="timings", label="99th %ile", y=21.518, yerr=0.0 PlotPoint: name="98th_percentile", plot="timings", label="98th %ile", y=15.5952, yerr=0.0 PlotPoint: name="95th_percentile", plot="timings", label="95th %ile", y=9.035, yerr=0.0 PlotPoint: name="90th_percentile", plot="timings", label="90th %ile", y=5.672, yerr=0.0 PlotPoint: name="80th_percentile", plot="timings", label="80th %ile", y=3.09, yerr=0.0 --------------------------------- Summary -------------------------------- Number of FAILED tests 1 Number of WRONG tests 0 Number of CORRECT tests 3266 Number of NEW tests 8 Total number of tests 3275 GREPME 1 0 3266 8 3275 X
Summary: correct: 3266 / 3275; new: 8; failed: 1; 9min Status: FAILED
Mon Sep 28 20:43:26 CST 2020 *** testing ended ****
does it mean there still some errors in my compilation? By the way, I also want to ask if I can compile cp2k-7.1 with plumed when using xconfigure and intel compiler? if it can be done, what should I do? I have tried to just add the following lines to "Linux-x86-64-intelx.arch" produced by xconfigure, "include /path/to/your/plumed2.0/installation/lib/plumed/src/lib/Plumed.inc EXTERNAL_OBJECTS=$(PLUMED_STATIC_DEPENDENCIES) DFLAGS += -D__PLUMED2 LIBS += -lz -ldl -lstdc++"
but it failed in the compilation process:
/home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In function
PLMD::isdb::RDC::do_svd()': (.text+0x21029a): undefined reference to
gsl_vector_alloc' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x2102b6): undefined reference to
gsl_vector_alloc' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x2102c8): undefined reference to
gsl_vector_alloc' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x2102d7): undefined reference to
gsl_vector_alloc' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x2102e6): undefined reference to
gsl_vector_alloc' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210307): undefined reference to
gsl_matrix_alloc' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210326): undefined reference to
gsl_matrix_alloc' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x21033a): undefined reference to
gsl_matrix_alloc' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210347): undefined reference to
gsl_matrix_set_zero' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x21034f): undefined reference to
gsl_vector_set_zero' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210474): undefined reference to
gsl_vector_set' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210481): undefined reference to
gsl_matrix_get' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x2104b8): undefined reference to
gsl_matrix_set' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x2104c8): undefined reference to
gsl_matrix_get' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x2104f3): undefined reference to
gsl_matrix_set' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210503): undefined reference to
gsl_matrix_get' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210534): undefined reference to
gsl_matrix_set' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210544): undefined reference to
gsl_matrix_get' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210564): undefined reference to
gsl_matrix_set' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210574): undefined reference to
gsl_matrix_get' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x21059b): undefined reference to
gsl_matrix_set' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210649): undefined reference to
gsl_matrix_memcpy' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x21066b): undefined reference to
gsl_linalg_SV_decomp' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210683): undefined reference to
gsl_linalg_SV_solve' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x2106d9): undefined reference to
gsl_vector_get' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x21073b): undefined reference to
gsl_vector_get' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210817): undefined reference to
gsl_vector_get' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210873): undefined reference to
gsl_vector_get' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x2108d7): undefined reference to
gsl_vector_get' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x21090d): undefined reference to
gsl_blas_dgemv' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210963): undefined reference to
gsl_vector_get' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210a24): undefined reference to
gsl_matrix_free' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210a2e): undefined reference to
gsl_matrix_free' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210a38): undefined reference to
gsl_matrix_free' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210a42): undefined reference to
gsl_vector_free' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210a4c): undefined reference to
gsl_vector_free' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210a56): undefined reference to
gsl_vector_free' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210a60): undefined reference to
gsl_vector_free' /home/liuly/opt/plumed/lib/plumed/obj/kernel.o: In functionPLMD::isdb::RDC::do_svd()': (.text+0x210a6a): undefined reference to
gsl_vector_free' make[3]: [/home/liuly/opt/cp2k-7.1/exe/Linux-x86-64-intelx/graph.popt] Error 1 make[3]: Waiting for unfinished jobs.... ^Z [2]+ Stopped make ARCH=Linux-x86-64-intelx VERSION=popt AVX=3 LIBINTROOT=/home/liuly/opt/libint/intel-skx LIBXCROOT=/home/liuly/opt/libxc/intel-skx ELPAROOT=/home/liuly/opt/elpa/intel-skx-omp -j 24
Could you give me some suggestions about the above questions? Thanks!
The updated ARCH files fixed the Regtest error(s). I am now getting 0 errors for PSMP and POPT when running CP2K 7.1 Regtests (Intel Compiler 2020u2). The issue was about relaxed/increased opt-level for more recent Intel Fortran Compiler, however, it is generally more stable to rely on -O2
(regardless of the compiler version). Please redo ./configure-get.sh cp2k
within your CP2K-7.1 root-directory.
Regarding a well-performing executable, I recommend GCC, LIBXSMM, and MKL since performance-critical code is covered by libraries (at least wrt CP2K 7.1). I further recommend PSMP over POPT at least for scale-out. Regarding the latter, you may want to have a look at https://xconfigure.readthedocs.io/cp2k/plan/ which contains perhaps useful information regardless of whether your workload is scale-out or emphasizes the "square-number property".
Regarding Plumed, I have to add recipes for this library. Are you depending on Plumed functionality in CP2K?
CP2K 7.1 was the first release with Plumed (if I remember correctly).
Thank you for your kind reply! I have passed all the tests successfully this time!
Yes, I want to do constrained molecular dynamics calculations depending on Plumed, but I don't how to install it using xconfigure, so you mean Plumed has been integrated to CP2K7.1 and I don't need to install/compile it by myself?
Thanks!
so you mean Plumed has been integrated to CP2K7.1 and I don't need to install/compile it by myself?
Just ignore my previous comment. I have now added Plumed2 support to XCONFIGURE and updated the step-by-step guide for building CP2K. I am currently testing the Plumed support for both Intel Compiler and GNU Compiler as well as for CP2K/master, CP2K 6.1, and CP2K 7.1. Tests are still in progress, and I will close this issue when they are successfully completed. You need to build Plumed using the new recipe as well as updating the ARCH files (./configure-get.sh cp2k
) and then rebuild CP2K (rm -rf exe lib obj
, etc).
Hold off testing it yet, there are problems with the current recipe (missing support for MPI, FFTW, etc.).
OK, I see! By the way, when I installed plumed-2.6.1 following your step-by-step guide for building CP2K, I got the following errors:
creating build/lib.linux-x86_64-3.7 g++ -pthread -shared -B /opt/intel/2020u2/intelpython3/compiler_compat -L/opt/intel/2020u2/intelpy thon3/lib -Wl,-rpath=/opt/intel/2020u2/intelpython3/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/t emp.linux-x86_64-3.7/plumed.o -o build/lib.linux-x86_64-3.7/plumed.cpython-37m-x86_64-linux-gnu.so /bin/ld: this linker was not configured to use sysroots collect2: error: ld returned 1 exit status error: command 'g++' failed with exit status 1 make[1]: [all] Error 1 make[1]: Leaving directory `/home/liuly/opt/plumed2-2.6.1/python' make: [all] Error 2
the intel compiler 2020u2 has been used in the installation, do you know the reason about this error?
Thanks very much!
As said, the recipes were not ready. Your g++
(above log) also indicates this was not the Intel variant of the configure wrapper script.
Now, the current status at GitHub (from an hour ago) seems to be good. I fixed potentially all missing features and CP2K/master regtests with Plumed passed (GNU Compiler). I am currently building CP2K 7.1 (built with Intel Compiler) and soon/hopefully starting the regtests soon.
CP2K 7.1 built with Intel Compiler and Plumed passed all regtests. I believe the whole thing is now ready.
I have tried to install Plumed again, however, I got the same error :
creating build/lib.linux-x86_64-3.7 g++ -pthread -shared -B /opt/intel/2020u2/intelpython3/compiler_compat -L/opt/intel/2020u2/intelpython3/lib -Wl,-rpath=/opt/intel/2020u2/intelpython3/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/plumed.o -o build/lib.linux-x86_64-3.7/plumed.cpython-37m-x86_64-linux-gnu.so /bin/ld: this linker was not configured to use sysroots collect2: error: ld returned 1 exit status error: command 'g++' failed with exit status 1 make[1]: [all] Error 1 make[1]: Leaving directory `/home/liuly/opt/plumed2-2.6.1/python' make: [all] Error 2
it is very strange, could you give me some suggestions about this error?
Do you made Intel Compiler available on your command line? When did this happen, I mean which step? Was it when you built Plumed2?
Yes, I made Intel Compiler available on my command line,. When i do"make -j" during Plumed compilation , I got this error.
I cannot help much with this other than pointing out the steps again:
make distclean
./configure-get.sh plumed
./configure-plumed-skx.sh
make -j; make install
You may use "hsw" (or even "snb") rather than "skx" depending on your target system/hardware. The non-gnu variants of the above script attempt to use Intel Compiler, given your log ([...] g++ -pthread -shared -B /opt/intel/2020u2/intelpython3/compiler_compat -L/opt/intel/2020u2/intelpython3/lib -Wl,-rpath=/opt/intel/2020u2/intelpython3/lib -Wl,--no-as-needed -Wl,--sysroot=/ build/temp.linux-x86_64-3.7/plumed.o -o build/lib.linux-x86_64-3.7/plumed.cpython-37m-x86_64-linux-gnu.so [...]
), it looks like using the GNU Compiler (g++
) and also attempts to build the shared library variant of Plumed2. The latter is disabled in my current configure-wrapper scripts. So, your log/error looks suspicious to me. As a reminder, you should have pulled the latest configure-wrapper scripts before you build Plumed2 (./configure-get.sh plumed
).
When I change the intel compiler from 2020u2 to 2018u4, the Plumed-2.6.1 can be installed/compiled successfully with xconfigure, however, the installation of libint-v2.6.0-cp2k-lmax-6 fails finally:
make: *** [src/HRRPart0bra0ket0hg001.lo] Error 1 icpc: error #10106: Fatal error in /home/liuly/intel/compilers_and_libraries_2018.5.274/linux/bin/intel64/mcpcom, terminated by kill signal compilation aborted for src/HRRPart0bra0ket0if.cc (code 1) make: *** Waiting for unfinished jobs.... icpc: error #10106: Fatal error in /home/liuly/intel/compilers_and_libraries_2018.5.274/linux/bin/intel64/mcpcom, terminated by kill signal compilation aborted for src/HRRPart0bra0ket0if010.cc (code 1) make: *** [src/HRRPart0bra0ket0if001.lo] Error 1 make: *** [src/HRRPart0bra0ket0md.lo] Error 1 icpc: error #10106: Fatal error in /home/liuly/intel/compilers_and_libraries_2018.5.274/linux/bin/intel64/mcpcom, terminated by kill signal compilation aborted for src/HRRPart0bra0ket0i100f.cc (code 1) icpc: error #10106: Fatal error in /home/liuly/intel/compilers_and_libraries_2018.5.274/linux/bin/intel64/mcpcom, terminated by kill signal compilation aborted for src/HRRPart0bra0ket0hg100.cc (code 1) make: *** [src/HRRPart0bra0ket0hg010.lo] Error 1 icpc: error #10106: Fatal error in /home/liuly/intel/compilers_and_libraries_2018.5.274/linux/bin/intel64/mcpcom, terminated by kill signal compilation aborted for src/HRRPart0ket0bra0md.cc (code 1) make: *** [src/HRRPart0bra0ket0h010g.lo] Error 1 make: *** [src/HRRPart0bra0ket0h100g.lo] Error 1 make: *** [src/HRRPart0bra0ket0i010f.lo] Error 1 make: *** [src/HRRPart0bra0ket0i100f.lo] Error 1 make: *** [src/HRRPart0bra0ket0ld100.lo] Error 1 make: *** [src/HRRPart0bra0ket0lf.lo] Error 1
sorry to bother you again!
thanks very much!
Hi, I have installed the CP2K-7.1 with Plumed following your guide successfully in another new device/system, Maybe there are some unknown problem in my current system, I will continue to pay attention to this problem and if I find the solutions, I will share it immediately!
Thank you for all your hard work and kind help these days!
Best, Leyu Liu
@Liuly98 Thank you for letting me know! Feel free to log any issue you run into!
Thank you also for requesting Plumed2 support in XCONFIGUREd CP2K. This is certainly useful for others as well.
See https://github.com/cp2k/cp2k/issues/1106