4pbt - adaptive analysis, time consumption details

The user times are measured using code with no compiler optimization, running on Pentium III processor at 800 Mhz, RedHat Linux 6.1,  egcs-2.91.66 compiler.
 

Mesh	nnodes/nelems	steps	dummy mapping, no renumbering user times [sec] (pm/ipm/ipu/mt//tt)	optimized mapping, no renumbering user times [sec] (pm/ipm/ipu/mt//tt)	optimized mapping, renumbering user times [sec] (pm/ipm/ipu/mt//tt)
1	145/234	1	--/--/--/--//0	-/-/-/-//0	-/-/-/-//0
2	293/522	2-5	0/ 0/ 0/ 1//4	0/0/0/0//3	0/0/0/0//2
3	397/722	6-8	0/ 1/ 1/ 3//6	0/0/0/0//4	0/0/0/0//3
4	507/940	9-10	0/ 2/ 2/ 6//8	0/0/0/0//3	0/0/0/0//2
5	961/1820	11-12	1/ 5/11/18//28	0/0/0/1//12	0/0/0/1//8
6	1150/2189	13-14	5/14/16/35//49	0/0/0/1//17	0/0/0/1//11
7	1391/2657	15-16	7/21/23/52//73	0/0/1/2//25	0/0/1/2//14
8	1451/2777	17-18	10/26/26/63//82	0/0/1/2//23	0/0/1/2//14
9	1605/3085	19-21	11/31/32/75//112	0/0/1/2//39	0/0/1/2//24
10	1683/3241	22-25	11/36/35/86//129	0/0/1/3//46	0/0/1/3//29
11	1810/3490	26-32	16/42/41/100//185	0/0/1/3//100	0/0/1/3//51
12	1896/3663	33-42	18/47/46/112//245	0/0/1/3//157	0/0/1/3//79

Legend:

• dummy mapping - mapping procedure based on whole mesh search
• optimized mapping - mapping procedure using tree technique to speed up the spatial localization. Used also to build up the nonlocal influence table at each integration point (explains difference for times needed to update IP state).
• pm - user time needed to map primary unknowns (displacement and their increments) from old mesh to new one.
• ipm - user time needed to map integration points internal variables
• ipu - user time needed to update integration points state
• mt - user time consumed by complete mapping process
• tt - total user time needed to solve the problem, until next remeshing required

This page is part of the OOFEM project documentation (www.oofem.org) (c) 2003 Borek Patzak, e-mail: info@oofem.org