| ²é¿´: 336 | »Ø¸´: 1 | ||
shulaiÌú³æ (³õÈëÎÄ̳)
|
[ÇóÖú]
vasp±àÒëÎÊÌâÇë½Ì
|
|
±àÒëvasp³öÏÖÈçϾ¯¸æÐÅÏ¢£¬¸ÃÈçºÎ½â¾ö°¡£¿Ð»Ð» fortcom: Warning: nonl.F, line 814: The statement following this DEC loop optimization directive must be an iterative do-stmt, a vector assignment, an OMP pdo-directive, or an OMP parallel-do-directive. !DIR$ IVDEP ------^ fortcom: Warning: hamil.F, line 1652: The statement following this DEC loop optimization directive must be an iterative do-stmt, a vector assignment, an OMP pdo-directive, or an OMP parallel-do-directive. !DIR$ IVDEP ------^ Space exceeded in Data Dependence Test in xi_mp_calculate_xi_ Subdivide routine into smaller ones to avoid optimization loss |
» ²ÂÄãϲ»¶
353Çóµ÷¼Á
ÒѾÓÐ7È˻ظ´
Çóµ÷¼Á
ÒѾÓÐ7È˻ظ´
²ÄÁÏ»¯¹¤306·ÖÕÒºÏÊʵ÷¼Á
ÒѾÓÐ11È˻ظ´
µ÷¼Á
ÒѾÓÐ8È˻ظ´
¹¤¿ÆÇóµ÷¼Á
ÒѾÓÐ13È˻ظ´
¸´ÊÔµ÷¼Á
ÒѾÓÐ5È˻ظ´
Ò»Ö¾Ô¸±±¾©»¯¹¤085600 310·ÖÇóµ÷¼Á
ÒѾÓÐ11È˻ظ´
Ò»Ö¾Ô¸±±¾©2£¬²ÄÁÏÓ뻯¹¤308Çóµ÷¼Á
ÒѾÓÐ10È˻ظ´
0817»¯Ñ§¹¤³ÌÓë¼¼ÊõÇóµ÷¼Á£¬Ò»Ö¾Ô¸Öк£Ñó319
ÒѾÓÐ9È˻ظ´
Ò»Ö¾Ô¸µç×ӿƼ¼´óѧ085600²ÄÁÏÓ뻯¹¤ 329·ÖÇóµ÷¼Á
ÒѾÓÐ6È˻ظ´
» ±¾Ö÷ÌâÏà¹Ø¼ÛÖµÌùÍÆ¼ö£¬¶ÔÄúͬÑùÓаïÖú:
vasp5.2 openmpi+intel+mkl±àÒ룬ÄÜÁ¿ÕýÖµÎÊÌâ
ÒѾÓÐ7È˻ظ´
vsap 5.2 »·¾³ºÍ±àÒëÎÊÌâ
ÒѾÓÐ14È˻ظ´
vasp5.2µÄ±àÒëÎÊÌ⠷dz£¸Ðл £¡£¡
ÒѾÓÐ14È˻ظ´
¡¾ÇóÖú¡¿vasp±àÒëÎÊÌ⣬Çë×Ô¼º¶¯ÊÖ°²×°³É¹¦µÄÈ˻ظ´£¡
ÒѾÓÐ9È˻ظ´
¡¾ÇóÖú¡¿vasp±àÒëÎÊÌâ¡£
ÒѾÓÐ6È˻ظ´
¡¾ÇóÖú¡¿VASP±àÒë³ö´í
ÒѾÓÐ5È˻ظ´
¡¾ÇóÖú¡¿vasp ¶Î´íÎó
ÒѾÓÐ6È˻ظ´
¡¾ÇóÖú¡¿vasp±àÒëʱ£¬±¨´í£ºcannot find -lmkl_em64t
ÒѾÓÐ11È˻ظ´
¡¾ÇóÖú¡¿vasp5.2Òì³£Í˳öµÄÎÊÌâ
ÒѾÓÐ16È˻ظ´
¡¾ÆäËû¡¿ÔÚÏß´ðÒÉ:VASPµÄÎÊÌâ
ÒѾÓÐ314È˻ظ´
shulai
Ìú³æ (³õÈëÎÄ̳)
- Ó¦Öú: 3 (Ó×¶ùÔ°)
- ½ð±Ò: 99.8
- ºì»¨: 1
- Ìû×Ó: 39
- ÔÚÏß: 4.9Сʱ
- ³æºÅ: 833501
- ×¢²á: 2009-08-24
- רҵ: ÐÂÐÍÐÅÏ¢Æ÷¼þ
|
±àÒëvaspµÄʱºòÓöµ½ÈçÏÂÕ⼸¸ö¹Ø¼ü×Ö£¬Ã»¿´Ã÷°×ÊÇʲôÒâ˼£¬Çë´ó¼ÒÖ¸½Ì£¬Ð»Ð»£¡ 1. avoidalloc If -Davoidalloc is set in the makefile, ALLOCATE and DEALLOCATE sequencies are avoided in some performance sensitive areas. Notably under LINUX ALLOCATE and DEALLOCATE is slow, and hence avoiding it improves the performance of some routines by roughly 10%. 2. MPI_BLOCK Presently VASP breaks up immediate MPI send (MPI_isend) and MPI receive (MPI_irecv) calls using large data blocks into smaller ones. We found that large blocks cause a dramatic bandwidth reduction on LINUX clusters linked by a 100 Mbit and/or Gbit Ethernet (all Kernels, all mpi versions including 2.6.X Linux kernels, lam.7.1.1). MPI_BLOCK determines the block size. If use_collective is used, MPI_BLOCK is used only for the fast global sum routine (search for M_sumf_d in mpi.F). |
2Â¥2012-02-09 15:26:52














»Ø¸´´ËÂ¥