================================ The Unofficial History of ANSYS first edited on June 18, 2002 updated Nov 3, 2004 updated Nov 11, 2004 updated Dec 12, 2005 updated Jan 11, 2006 updated Dec 3, 2006 This is a very important update, so I will have to mention a little about the source and history here. Dr. Peter Kohnke contacted me in the middle of 2005, telling me some of the information in this article may not be so accurate. Dr. Kohnke has worked with Dr. Swanson since 1/1/1973, and is still working for ANSYS as of December of 2005. So it's more than 32 years now. He specially pointed out the "sandwiches" story. He doubt that might be wrong. This is what he said "........I very much doubt that story about the sandwich. I never heard of anyone who was hired to answer the phone. What would they say? That John was busy? Rather, his first wife Marcy helped out a lot in many ways. Also, I never knew John to pass up a chance to go to a restaurant.............Indeed, for many years, John invited the entire staff to a restaurant regularly for our staff meetings." And, one thing even more important "There was no garage. The offices were part of his home." So I am going to delete this paragraph. I want to thank Dr. Kohnke again for kindly providing these information. ================================ In 1963, Dr. John Swanson worked at Westinghouse Astronuclear Labs in Pittsburgh, responsible for stress analysis of the components in NERVA nuclear reactor rockets. He used computer codes to model and predict transient stresses and displacements of the reactor system due to thermal and pressure loads. Swanson continued to develop 3-D analysis, plate bending, nonlinear analysis for plasticity and creep, and transient dynamic analysis, in the next several years, using a finite element heat conduction program that was developed by Wilson at Aerojet. The old Westinghouse codes included a 2d/axisymmetric one also, possibly called FEATS (according to Kohnke). John wanted to combine these codes to remove the duplication, like equation solvers and some postprocessing Swanson believed an integrated, general-purpose FEA code could be used to do complex calculations that engineers typically did manually, such as heat transfer analysis. It would save money and time for Westinghouse and other companies. Westinghouse didn't support the idea, and Swanson left the company in 1969. Before he left, he made sure that all code work had been sent to COSMIC, so that he could pick it up again from the outside. Swanson Analysis Systems, Inc was incorporated in the middle of 1970 at Swanson's home. The offices were part of Swanson's home (There was no garage) in Pittsburgh. at the same time, Westinghouse realized that they needed John, so they hired him as a consultant. John said sure, but with the proviso that whatever he put into STASYS, the Westinghouse code, he could also put into ANSYS. Westinghouse had no trouble with this, as they just wanted to solve their problems. So this consulting kept bread on the table for the Swansons, and at the same time brought forth further improvements to ANSYS. He developed his program using a keypunch and a time-shared mainframe at U.S. Steel. The first version of ANSYS was coded by the end of 1970, and the ANSYS program was first leased soon after that. Westinghouse was the first customer, running as a data center. The data center was at the Telecommunications Center on the Parkway East, on the east side of Pittsburgh. According to Dr. Swanson, the name ANSYS was because the copyright lawyers assured Swanson that ANSYS was just a name, and did not stand for anything. Understandable, during that period all programs were "written" on punch card. When installing the program on the customer's computer, it meant carried a relatively big case of punch cards to the customer's place, and fed them into the machine. Dr. Peter Kohnke met John Swanson first about early 1971. Swanson offered Peter a job in the fall of 71, but Peter did not accept. As that time Peter was a brand new father at the time, and Westinghouse looked a lot more secure than SAS. Peter told John he was interested and would accept in the spring of 72, but then he had hired Gabe DeSalvo, and did not have the resources to hire Peter also. But that fall John did, and then hired Peter. Dr. Peter Kohnke's start date was 1/1/73. (Dr. Kohnke still worked at ANSYS as of December of 2005). When Peter started work, he asked what John wanted him to do. John told Peter that he developed code, did technical support, wrote manuals, gave seminars and did systems work so that the program would run on a variety of systems. John said he needed relief, so Peter should pick one or two of them. Ultimately, Peter did all of them, except systems work. In around 1970, users can ran ANSYS 2.x on a CDC 6600 machine over the Cybernet timesharing network. That time only fixed format input was available. The users would work up the input listing off-line, key it onto a tape cassette, log on, submit the run about quitting time for the best computer rates and stop by the CDC data center next morning to find out what went wrong. In 1973, ANSYS ran on three kinds of hardware: CDC, Univac, and IBM. And around 1973 the USS mainframe that they used to develop code was the US Steel CDC 6500. The first minicomputer that ANSYS ran on was a MODCOMP 4 (or IV?) VAX came later. Being a small company, everyone did everything. When a 'mini' computer was delivered, everyone helped wrestle it off of the truck. When printout paper was delivered, everyone helped unload the boxes. In 1975, MITS began to build and sell the first PC ever in human history, the Altair. That, of course, did not have anything to do with ANSYS yet. The so-called PC was just a few switches and lights on the front board, and input had to be done in a binary fashion (no keyboard and monitor, of course). What was worse, was that you have to assemble it by yourself. And, it usually didn't work. Although Altair was rather popular, nobody really knew what to do with this machine. One former customer said that, the most popular activity on Altair, was to figure out what to do with this machine. At the At the same time, Microsoft built the BASIC language for Altair. 1977, Apple I was born. In around 1979, Revision 3.0, ANSYS run on a VAX 11-780 minicomputer. ANSYS evolved from fixed format input to purely command line driven and monocolor (green) on a Tektronix 4010 or 4014 vector graphics monitor. For a descent size model, the hidden lines plots could take 20-30 minutes. All of the nodes and elements were created separately without the benefit of importing CAD geometry. NGEN, EGEN, RPnnn, were used extensively. There was a geometry prepcessor, PREP7. 1980, we had Apple II. In around 1980, John Swanson bought a Radio Shack TRS-80 machine, and planned to build a commercial version on it. However, later John returned the machine because Radio Shack left out (a socket for) a floating point processor. John decided that Finite Element Analysis probably should utilize a floating point processor, so he got his money back for that one. Also around 1980,Rev 4 on an VAX 11-780 system was great, according to some old users. The chasm between batch and interactive running pretty much disappeared and file management was a very easy thing. No more element hard coding, the post processing got hugely better and you could mix batch and interactive running as you saw fit. Big dynamic transient runs or substructuring over night, post-processing and plotting next morning. Emag capabilities were first introduced at Rev 4.1. Also in 1980, Microsoft signed contract with IBM to provide the OS, PC DOS, for its up coming PC. This OS, however, was not created by Microsoft. Microsoft bought it rom an engineer for 50K USD, which was named the QDOS - the Quick and Dirty Operation System. 1981, IBM PC was born. This computer was created using the off the shelf technology, and an open architecture. The original reasons were to push the product to the market ASAP, so that IBM could catch up with the PC market. However, the BIOS was proprietary. Later Compaq reverse-engineer the BIOS and created a fully IBM PC compatible BOIS. This ignited the PC cloning market and war. The booming of PC market directly changed the meaning of computing. PC price dropped 30% at one month. And, it was the booming of cloned IBM PC that really brought money into Microsoft. 1984, the revolutionary Macintonsh was born. Macintosh was far advanced then the IBM PC family at that time. The concept of GUI in the OS level and WYSIWYG was not possible on IBM PC until almost one decade later. However, the market of Macintosh did not pick up very soon, which caused the Steve Job's leave from Apple computer. However, later the sales of Macintonsh began to take off, which proved that Steve Job's vision had all been right. Macintosh saved Apple, and was directly responsible for the phenomena of Apple craze and fans. A PC version of ANSYS was also available at around version 4.0 too in about 1984. It was running on a Intel 286, with interactive command line input and limited graphics on the screens, like elements and nodes. No Motif GUI yet. In the first release on ANSYS on PC's, preprocessing, solution and post processing were performed in separate programs. "Design Optimization" was introduced at Rev 4.2 (1985). This is also the release at which "Macro length is no longer limited to 400 characters." FLOTRAN started as a graduate (PhD) project by Rita J. Schnipke in the University of Virginia circa 1986. After grad school Rita started (or helped start) Compuflo which was later sold to ANSYS in 1992. Rita later started her own shop which is in Charlottesville VA called Blue Ridge Numerics. They make CFDesign, a finite element based CFD code (www.cfdesign.com). 1988 at an ANSYS conference in California, IBM was there pushing their first unix machine, the "RT". It was slow. They asked Dr. Swanson if he would make a comment on it. He said "RT must stand for Real Turkey. SASI first started working with Compuflo (FLOTRAN) in 1989. At ANSYS Rev 5.0 and FLOTRAN V2.1A, SASI had what they called a "seamless interface" between the two programs (1993). FLOTRAN was "fully integrated" into ANSYS at Rev 5.1 (1994). In 1993, Version 5.0 was released. And the version 5.1 later has a Motif GUI, which would remained the similar layout up to 6.0. Swanson Analysis Systems, Inc., was sold to TA Associates in 1994. The new company name, ANSYS, Inc., was announced at AUTOFACT '94 in Detroit. According to many different people in the old SAS, Inc, John Swanson treated the people there pretty well. In contrast to the old "sandwitch" jokes, John never passed up a chance to go to a restaurant. Indeed, for many years, John invited the entire staff to a restaurant regularly for the staff meetings. At one time, Kohnke told Swanson that he ran the company like a benevolent dictatorship, and later Swanson told Kohnke that he liked that characterization. Many people had said that John Swanson had an amazing overall understanding As well as detailed knowledge of the ANSYS code. Kohnke told a small story in an email to the author : " Sometime in the late 70's, a bug came my way. I wrestled with it for maybe half a day without making real progress. Then I went to John's office to ask him if he had any ideas. After I explained the bug, John thought about it for about 3 seconds (literally!) and said: 'Didn't you make a change in XXX about 6 months ago that would have a bearing on this?' In a nutshell, he was correct and I was then able to resolve the bug! John's knowledge and understanding of the code was always amazing to me." 1995, Windows 95 was published. Windows 95 was an important milestone for Microsoft. It bridged between the old DOS OS into the new NT technology. The birth of Windows 95 finally made it more and more acceptable for engineering community to use PC as a heavy duty calculation machine like workstations. In 1996, ANSYS 5.3 was published, with support for LS-DYNA. The feature of ANSYS/LS-DYNA in ANSYS 5.3 was still in the beginning stage. On June 20, 1996, ANSYS Inc. common stock began trading on Nasdaq under ANSS after being 26 years a privately held company. The IPO generated more than $41 million. 1998, ANSYS began to ship ANSYS/ed to university labs and paper reviewers. One of the copy arrived at the Structures Lab of Civil Engineering Department in Arizona State University, and that was the first time the author knew about ANSYS. In the same year, on ANSYS's Annual report, it said "John is retiring from his direct role at ANSYS Inc., but will continue his association as a key consultant, mentoring all of us for many years to come.." On August 31, 2000, ANSYS acquired ICEM CFD. January 2001, ANSYS announced the release of CADfix (International TechneGroup Incorporated) for ANSYS version 5.6.2 and 5.7. CADfix was to address the issue of importing CAD model into ANSYS with automatic geometric data repair. November 2001, ANSYS acquired CADOE S.A, an independent software vendor that specializes in the CAD/CAE market. In the same month, ANSYS announced a strategic OEM partnership with SAS LLC, a provider of NASTRAN simulation software and services. The alliance was focused on the joint development of a new NASTRAN computer-aided engineering solution that will be distributed exclusively by ANSYS Inc. November 2001, ANSYS announced AI*Environment. AI*Environment combines ICEM CFD Engineering's pre- and post-processor technologies. December 2001, ANSYS 6.0 was released. In this version, the Sparse solver was greatly improved. Efficient and reliable large scale model analysis (say, 1M DOF) finally became practical. The graphics screen of ANSYS was also painted blue in 6.0, which came out to be a great disappointment to a lot of users. In April of 2002, ANSYS 6.1 was released. The familiar Motif GUI was replaced by a Tcl/tk developed interface. It runs on 64-bit Intel Itanium architecture with Windows XP. In February 26, 2003, ANSYS acquired CFX. ANSYS also announce that the functionality of Flotran will be "capped" at 8.1. That is, there will be more more development of Flotran after 8.1. Expect for the Multiphysic platform, Flotran will be replaced by CFX. March 2004, ANSYS Announces ParaMesh 2.3 May 2004, ANSYS 8.1 and CFX 5.7 was released. According to the statement from ANSYS, the functionality of Flotran is "capped" since this verision. That is, there will be no more revision of Flotran after 8.1. June 2004, ICEM CFD 5.0 was released. Jan 5, 2005, ANSYS announced that it has acquired Century Dynamics. Century Dynamics. Century Dynamics' main product, AUTODYN, includes computational structural dynamics finite element solvers (FE), finite volume solvers for fluid dynamics (CFD), mesh-free particle solvers for high velocity, large deformation and fragmentation problems (SPH), and multi-solver coupling for multi-physics solutions including coupling between FE, CFD and SPH methods. Feburary 16, 2006, ANSYS Signs Definitive Agreement to Acquire Fluent. =================== What is a data center ? =================== I actually hear nouns "data center" frist time from Dr. Kohnke. For my age, I am more familar with workstation and PC and xbox and psp and so on........... I used VMS when I was in college. And I remember later soon the whole room was occupied by PC and Macs. So I asked Dr. Kohnke what is a data center anyway ? And this is what he told me : .... long ago (in the computer age), computers were relatively rare (and expensive). So, if you invested in one, you wanted to get the maximum use of it. You would run it day and nite. This is unlike our pc's now (like our cars), which frankly just sit there most of the time, waiting for our command. Before it was very much a shared and continually used resource. So, ANSYS got its start in that environment. Users at Westinghouse wanted to run ANSYS, so Westinghouse made a contract with SASI, where SASI would supply the code and support, and Westinghouse would pay to SASI so much per unit of time royalty for the time that the computer was actually being used to run ANSYS. As a result, many users were using the same machine (one after the other - parallel did not exist then). This was called a 'data center'. The next step was external users. Knowing that even as a large company, you might not need your computer full time, so you would have your sales people out there selling time on it, often only at nite, to external users. Other companies sprang up that had no user base of their own, but only a computer(s), sales people, and external customers. The better ones also offered their own very good technical support. These were also called data centers. And of course the whole concept of a data center disappeared as computers got cheaper and faster. The 'lease' jumped to the concept that the clock was not important; this 'new' contract had SASI being paid so much per month, regardless of actual usage. This is closer to how things are now. =================== From The Author =================== I began to collected and wrote this article in the early 2002. One of the reasons was that I signed a contract with a publisher in Taiwan, to write a book about using ANSYS for industrial product reliability analysis. I felt that it was necessary to have a chapter totally contribute to the history of ANSYS. And, the most important, it will be very interesting. The author wants to thank the help from many engineers and scientists in the xansys internet group. Some of the former employees of ANSYS also contribute greatly to this article, and many of them prefer not to be named. I also received emails from differetn people, and I usually tried to verify before I used them. Although I am trying to keep all the statement as accurate as possible, I really can not guarantee the correctness of any information in this article. Many of us, including the author in the xansys group especially want to thank Dr. John Swanson, who invented ANSYS, and changes the life of many engineers forever in certain ways. I have lived in the States for totoally 10 years. I lived there, educated there, married there, had my son there, bought my first new car there, and had my first house there. It totally changes my life, my thinking (and my head) and everything. And I have to say most of my finacial basis was built on the existense of ANSYS and another program LS-DYNA. Anyone is welcome to distribute this article anyway he or she wants, as long as the original article remains unchanged (including "From The Author"). Comments and suggestion should be forwarded to the authors directly. I will be glad to update this file continuously. This article will be always on the web site www.FEA-Optimization.com, util I finally can not afford to pay the registration fee. To avoid spam, I am not going to put my email address here. You should be able to contact me, or find the way to contact me through the web site. PS. I worked as the VP of Engineering Consulting Division of CADMEN since September of 2002 (CADMEN, TADC is an ASD in Taiwan). However I have left CADMEN in August of 2004. If anyone wants to contact me, please do not write to my former email at CADMEN anymore. As of 2005, I am having my own software company, FEA-Opt Technology, which devoted in consulting services and design optimization software. Please give our web site a visit when you have time. Shen-Yeh Chen, Ph.D. www.FEA-Optimization.com