2Outline of talkOrigins of DYNA3D at LLNLfor LSDYNA4Origins of DYNA3DManual released in August 1976 for public distributionJohn Hallquist was the development teamFUFO bomb cancelledDevelopment of DYN ID: 877859
Download Pdf The PPT/PDF document "The History of LSDYNADavid J BensonDept ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
1 The History of LS-DYNA®David J. BensonDe
The History of LS-DYNA®David J. BensonDept. of Mech. & Aero. Eng.UCSD 2 Outline of talk Origins of DYNA3D at LLNL. for LS-DYNA. 4
2 Origins of DYNA3D Manual released in Aug
Origins of DYNA3D Manual released in August, 1976, for public distributionJohn Hallquist was the development team.FUFO bomb canc
3 elled Development of DYNA2D and NIKE2D,
elled Development of DYNA2D and NIKE2D, NIKE3D started (also with Hallquist as the development team). Request for DYNA3D source co
4 de from France. DYNA3D released into th
de from France. DYNA3D released into the public domain (1978) without restrictions. 6 Origins of DYNA3D The 1979-1981 versions an
5 d their revisions created interest in Ja
d their revisions created interest in Japan and Europe. BCS in London had several large users including Rolls-Royce Jet engines.
6 User seminars started in Japan and Europ
User seminars started in Japan and Europe in 1982 Lab started to get inquires from several companies for permission to commerciali
7 ze the code.by a letter from a lab attor
ze the code.by a letter from a lab attorney (Technically, permission was not needed.)Two companies begin sales and marketing activ
8 ities for DYNA3D based software, creatin
ities for DYNA3D based software, creating even more interest in the free public domain version. 8 Origins of DYNA3D Critical capab
9 ility for buckling in crash. 11 Origins
ility for buckling in crash. 11 Origins of DYNA3D By 1988 approximately 600 tapes containing DYNA3D, DYNA2D, NIKE2D, NIKE3D, TAURU
10 S, and INGRID had been sent to requestor
S, and INGRID had been sent to requestors from LLNL. companies and organizations on the use of DYNA3D.consulting by DOE employees
11 to transfer technology to industry.allow
to transfer technology to industry.allowed to consult with LSTC due to potential conflicts of interest. 13 Origins of DYNA3D By 19
12 89 DYNA3D was the most advanced FEA code
89 DYNA3D was the most advanced FEA code available for transient dynamics. A user base of several hundred companies, which needed
13 support. Hallquist had connections to th
support. Hallquist had connections to the user base due to contacts while at LLNL.This customer base provided a starting point for
14 LSTC.Industry started purchasing superc
LSTC.Industry started purchasing supercomputers. 15 John Hallquist: 1 class in Fortran 66.David Benson: 1 class in Fortran 66. Al
15 l DYNA3D development in Fortran. Develop
l DYNA3D development in Fortran. Developed on Crays. Execution speed was always a concern. Support of 1 computer scientistfor grap
16 hics 17 Cost-Benefit Analysis Roughly 1
hics 17 Cost-Benefit Analysis Roughly 10 prototypes per year= $2.5M Prototype costs up.Computing costs down. 19 Crash Model Size
17 Trends 1986: First model had 3439 elemen
Trends 1986: First model had 3439 elements. 1990: 15-20,000 elements. 1995: 50-100,000 elements. 2000: 100-250,000 elements. 2005:
18 1-1.5x106 Allcurrent simulations perfor
1-1.5x106 Allcurrent simulations performed on 21 Early Crash Calculation ~1994 36000Elements 22 1,500,000 Elements 23 LSTC LS-DY
19 NA Development LSTC developments are con
NA Development LSTC developments are concentrated on three products: LS-PrePost and LS-Opt are part of the LS-Dyna distribution an
20 d do not require license keys. 26 LS-DYN
d do not require license keys. 26 LS-DYNA Development Advantages of the one code strategyImplicit MPP utilizes all prior efforts f
21 or explicit More freedom for developers,
or explicit More freedom for developers, who can work on multiple developments governed by different field equationsLS-PrePost/LS-
22 Opt software development supports one in
Opt software development supports one interface.QA is performed on one codeNo costly add-ons for customers who require multi-physi
23 cs solutions. 28 Development Goals-Imp
cs solutions. 28 Development Goals-Implicit Springback for sheet metal stamping. Static initialization of crash models. Dynamic
24 springback simulation after crash simula
springback simulation after crash simulationReliable measurements between numerical and physical results can be more easily obtain
25 ed. An embedded linear capability to aut
ed. An embedded linear capability to automatically solve modes.Include infinitesimal motions superimposed on rigid bodies for NVH
26 and durability modeling. Eigenvalue anal
and durability modeling. Eigenvalue analysis to check the rigid body modes in the crash models.Identify inadvertent constraints. 3
27 0 LSTCs Vision LS-DYNA specific pre-pro
0 LSTCs Vision LS-DYNA specific pre-processing, post-processing, LS-PrePost, and optimization, LS-OPT, with no added charges. Unr
28 estricted open databases. Focus on larg
estricted open databases. Focus on large distributed memory low-cost clusters running large simulations. software prices per proc
29 essor will proportionally decrease to ke
essor will proportionally decrease to keep simulation costs affordable. Optimization technology will automate engineering design t
30 echnology. 32 Near Future for Explicit M
echnology. 32 Near Future for Explicit Model sizes of 10,000,000 elements. 128-512 processors in overnight runs. Human dummy mode
31 ls, such as THUMS, will increase model s
ls, such as THUMS, will increase model sizes even further. Honeycomb barriers will be modeled by shell elements. Number of process
32 ors will increase 5-10 times. Optimizati
ors will increase 5-10 times. Optimization software use in crash analysis will become widespread. 34 Parallel Computing -2006 the
33 use of explicit codes has From 100% seri
use of explicit codes has From 100% serial and SMP licensed CPUs for crash to 90% MPP with the remaining 10% of CPUs typically r
34 unning smaller models on 1-8 processors.
unning smaller models on 1-8 processors.Today serial and SMP explicit codes are becoming obsolete and will eventually be phased ou
35 t. Requires more expensive hardware so
t. Requires more expensive hardware so 36 Scalability on Large Clusters Cray XD1 with RapidArray intercon 3 Car crash simulation
36 run to completion (750K nodes)64 x 2 x 2
run to completion (750K nodes)64 x 2 x 2 = 256 1696 sec32 x 2 x 2 = 128 241624 x 2 x 2 =962981single core 2.2 GHz16 x 2 x 2 =64384
37 632 x 2 x 1 = 64461912 x 2 x 2 =4852268
632 x 2 x 1 = 64461912 x 2 x 2 =4852268 x 2 x 2 =3275914 x 2 x 2 =16140782 x 2 x 2 =8262304 x 2 x 1 =8 246811 x 2 x 2 =4494602 x 2