版塊導(dǎo)航: 正在加載中...

登錄注冊

應(yīng)《網(wǎng)絡(luò)安全法》要求，自2017年10月1日起，未進行實名認(rèn)證將不得使用互聯(lián)網(wǎng)跟帖服務(wù)。為保障您的帳號能夠正常使用，請盡快對帳號進行手機號驗證，感謝您的理解與支持！

24小時熱門版塊排行榜

返回列表

sic029

鐵蟲 (初入文壇)

應(yīng)助: 0 (幼兒園)
金幣: 70.6
帖子: 7
在線: 5.4小時
蟲號: 1960088
注冊: 2012-08-28
性別: GG
專業(yè): 結(jié)構(gòu)陶瓷

[求助] qsub提交并行siesta不成功，求助

大家好，交流下集群程序使用遇到的問題，多謝。
[node21:10714] *** An error occurred in MPI_Comm_rank
[node21:10714] *** on communicator MPI_COMM_WORLD
[node21:10714] *** MPI_ERR_COMM: invalid communicator
[node21:10714] *** MPI_ERRORS_ARE_FATAL: your MPI job will now abort
--------------------------------------------------------------------------
mpirun has exited due to process rank 3 with PID 10711 on
node node21 exiting improperly. There are two reasons this could occur:

1. this process did not call "init" before exiting, but others in
the job did. This can cause a job to hang indefinitely while it waits
for all processes to call "init". By rule, if one process calls "init",
then ALL processes must call "init" prior to termination.

2. this process called "init", but exited without calling "finalize".
By rule, all processes that call "init" MUST call "finalize" prior to
exiting or it will be considered an "abnormal termination"

This may have caused other processes in the application to be
terminated by signals sent by mpirun (as reported here).
--------------------------------------------------------------------------
[node21:10707] 7 more processes have sent help message help-mpi-errors.txt / mpi_errors_are_fatal
[node21:10707] Set MCA parameter "orte_base_help_aggregate" to 0 to see all help / error messages

新編譯的計算程序siesta，用qsub job提交上去很快結(jié)束提示的信息，能否幫忙診斷一下情況。在另外一個集群上編譯后直接用mpirun -np 4 siesta可以順利執(zhí)行的，不知道為何在新集群用qsub出現(xiàn)這個問題，這個新集群不讓進入到子節(jié)點，所以必須要解決這個問題才行，多謝了。

不知道是哪里的問題，之前在該環(huán)境并行編譯的lammps和vasp都使用很順利，就是siesta用qsub提交作業(yè)總是無法正常計算，但是并行編譯的siesta在另外環(huán)境下的子節(jié)點用mpirun -np 4 siesta執(zhí)行很順利，糾結(jié)了。

哦，登錄節(jié)點上mpirun我試過的，請幫忙看看，感覺被管理員設(shè)置了也無法用：
mpirun -np 4 siesta
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
This will severely limit memory registrations.
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
This will severely limit memory registrations.
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
This will severely limit memory registrations.
libibverbs: Warning: RLIMIT_MEMLOCK is 32768 bytes.
This will severely limit memory registrations.
--------------------------------------------------------------------------
The OpenFabrics (openib) BTL failed to initialize while trying to
allocate some locked memory.  This typically can indicate that the
memlock limits are set too low.  For most HPC installations, the
memlock limits should be set to "unlimited".  The failure occured
here:

  Local host: manage1
  OMPI source: btl_openib_component.c:1115
  Function:    ompi_free_list_init_ex_new()
  Device:       mlx4_0
  Memlock limit: 32768

You may need to consult with your system administrator to get this
problem fixed.  This FAQ entry on the Open MPI web site may also be
helpful:

http://www.open-mpi.org/faq/?category=openfabrics#ib-locked-pages
--------------------------------------------------------------------------
--------------------------------------------------------------------------
WARNING: There was an error initializing an OpenFabrics device.

  Local host: manage1
  Local device: mlx4_0
--------------------------------------------------------------------------
[manage1:16214] *** An error occurred in MPI_Comm_rank
[manage1:16214] *** on communicator MPI_COMM_WORLD
[manage1:16214] *** MPI_ERR_COMM: invalid communicator
[manage1:16214] *** MPI_ERRORS_ARE_FATAL: your MPI job will now abort
--------------------------------------------------------------------------
mpirun has exited due to process rank 0 with PID 16212 on
node manage1 exiting improperly. There are two reasons this could occur:

1. this process did not call "init" before exiting, but others in
the job did. This can cause a job to hang indefinitely while it waits
for all processes to call "init". By rule, if one process calls "init",
then ALL processes must call "init" prior to termination.

2. this process called "init", but exited without calling "finalize".
By rule, all processes that call "init" MUST call "finalize" prior to
exiting or it will be considered an "abnormal termination"

This may have caused other processes in the application to be
terminated by signals sent by mpirun (as reported here).
--------------------------------------------------------------------------
[manage1:16211] 3 more processes have sent help message help-mpi-btl-openib.txt / init-fail-no-mem
[manage1:16211] Set MCA parameter "orte_base_help_aggregate" to 0 to see all help / error messages
[manage1:16211] 3 more processes have sent help message help-mpi-btl-openib.txt / error in device init
[manage1:16211] 3 more processes have sent help message help-mpi-errors.txt / mpi_errors_are_fatal

計算子節(jié)點無法進入，被限制死了的。

這邊用的pbs作業(yè)管理gridview，我用的提交腳本是：
====================================
#PBS -N test
#PBS -l nodes=1:ppn=8
#PBS -j oe
#PBS -l walltime=24:00:00

cd $PBS_O_WORKDIR
NP=`cat $PBS_NODEFILE|wc -l`
source /public/software/mpi/openmpi1.5.4-intel.sh
mpirun  -machinefile $PBS_NODEFILE -np $NP  \
/home/sw/siesta/siesta-3.1/Obj/siesta < fe.fdf | tee output
=====================================

感謝蟲友幫忙，多謝。

回復(fù)此樓

» 猜你喜歡

求調(diào)劑已經(jīng)有4人回復(fù)
一志愿北京化工大學(xué)070300 學(xué)碩336求調(diào)劑已經(jīng)有6人回復(fù)
考研調(diào)劑已經(jīng)有3人回復(fù)
能源材料化學(xué)課題組招收碩士研究生8-10名已經(jīng)有16人回復(fù)
318求調(diào)劑已經(jīng)有5人回復(fù)
285求調(diào)劑已經(jīng)有5人回復(fù)
306求調(diào)劑已經(jīng)有3人回復(fù)
354求調(diào)劑已經(jīng)有10人回復(fù)
303求調(diào)劑已經(jīng)有5人回復(fù)
環(huán)境學(xué)碩288求調(diào)劑已經(jīng)有4人回復(fù)

» 本主題相關(guān)價值貼推薦，對您同樣有幫助:

【求助】siesta 安裝出錯已經(jīng)有18人回復(fù)
【求助】求助siesta的Ordern方法不能收斂的問題已經(jīng)有4人回復(fù)
【求助】siesta3.0并行成功安裝后不能并行的問題已經(jīng)有15人回復(fù)

research

1樓 2012-09-07 16:09:01

已閱回復(fù)此樓關(guān)注TA 給TA發(fā)消息送TA紅花 TA的回帖

redsnowolf

銀蟲 (小有名氣)

應(yīng)助: 7 (幼兒園)
金幣: 416.5
散金: 36
紅花: 5
帖子: 213
在線: 689.8小時
蟲號: 1332218
注冊: 2011-06-27
性別: GG
專業(yè): 半導(dǎo)體材料

【答案】應(yīng)助回帖

★
liliangfang: 金幣+1, 謝謝交流 2012-09-15 15:12:35

我前兩天用vasp也出現(xiàn)類似問題，剛剛解決～

The OpenFabrics (openib) BTL failed to initialize while trying to
allocate some locked memory.  This typically can indicate that the
memlock limits are set too low.  For most HPC installations, the
memlock limits should be set to "unlimited".  The failure occured
here:

  Local host: node21
  OMPI source: btl_openib_component.c:1055
  Function:    ompi_free_list_init_ex_new()
  Device:       mlx4_0
  Memlock limit: 65536

You may need to consult with your system administrator to get this
problem fixed.  This FAQ entry on the Open MPI web site may also be
helpful:

http://www.open-mpi.org/faq/?category=openfabrics#ib-locked-pages
--------------------------------------------------------------------------
--------------------------------------------------------------------------
WARNING: There was an error initializing an OpenFabrics device.
上面那個網(wǎng)址里15、16、17說的挺清楚的，我的情況是在每個節(jié)點ulimit -a顯示locked memory都正常，可就是出錯說內(nèi)存分配不正常，那個網(wǎng)址里說可能是登錄時沒有正常執(zhí)行系統(tǒng)所設(shè)的locked memory，或者作業(yè)調(diào)度系統(tǒng)沒有分配給應(yīng)用程序足夠大的內(nèi)存，最后重啟了一下每個節(jié)點的pbs調(diào)度系統(tǒng)的守護進程，問題解決了～
或者你可以在mpirun前邊兒加上ulimit -l unlimited，用qsub提交下試試
希望以上信息對樓主有用～

贊一下