## Friday, February 18, 2011

### Autodesk to Acquire Blue Ridge Numerics, CFdesign

## Thursday, February 10, 2011

### Opportunity to learn ACIS Geometric Kernel and Hoops 3D Graphics system in 3 day workshop

I am attending workshop organized by Spatial Corporation from February 21^{st} to 24^{th}, 2011 at IISc, Bangalore. ACIS- Geometric Kernel, 3D Interop – CAD translation library and Hoops – 3D graphics system by Techsoft3D would be covered in 4 days. It is a good opportunity to have quick hands on important tools of CAx software development. You can register yourself free at http://www.spatial.com/community/2011-spatial-seminar-india. Make sure your registration is confirmed as there are limited seats.

This is an excellent proactive major by Spatial to reach to world fastest moveing economy, India. Developer and researcher will be provided hands on and expert talk on their components, technologies and more important how to use them. I have come across few CAD developers or manager who does not what is geometric kernel or what is Scene Graph or have heard just name. Also many does not have idea how it is beneficial to their 3D product development. I do not consider it as totally their fault. Very less information is available about Geometric Kernel, Constrain Solvers and advance graphics system in public domain.

ACIS is considered to be acronym **A**lan, **C**harles, **I**an's **S**ystem name of three founders. ACIS empowers lot many CAx software, latest from famous one is SpaceClaim. ACIS as Geometric Kernel have been developed for more than 25 year. (I guess this year they would be completing Silver jubilee). ACIS provides C++ API interface for accessing more than 300 unique functionalities. Model Representation, Creation, Editing, Rendering and Application Support can be considered as sub set of modules. You can access complete documentation (yes completely open) of ACIS at: http://doc.spatial.com/index.php/3D_ACIS_Modeler.

Interestingly ACIS has little different topological structure as shown in above structure. Each of topological entity can be defined in following manner:

- BODY - The highest level of model objects, and is composed of lumps.
- LUMP - A 1D, 2D, or 3D set of points in space that is disjoint with all other lumps. It is bounded by shells.
- SHELL - A set of connected faces and wires, and can bound the outside of a solid or an internal void (hollow).
- Subshells form a further decomposition of shells for internal efficiency purposes.
- FACE - A connected portion of a surface bounded by a set of loops.
- LOOP - A connected series of co edges. Generally, loops are closed, having no actual start or end point.
- WIRE - A connected series of co edges that are not attached to a face.
- COEDGE - Represents the use of an edge by a face. It may also represent the use of an edge by a wire.
- EDGE - A curve bounded by vertices
- VERTEX - Location of a point

Model structure of ACIS is shown in below figure. Various API are provided to access all the relations. Figure also show how geometry is mapped to topological entity to define complete BREP model.

I would write details article on ACIS to explain the system in comprehensive manner. Mean time do attend Workshop and also a session on ACIS at GeomTech.

In Bangalore workshop, 3D Interop is second component in focus. 3D Interop is another famous CAx component developed by Spatial Corp. I think most of its development happens at 3D PLM Software at Pune. As matter of fact our CCTech’s five DACAD student are the developer in this division J. 3D Interop provides strong CAD connection to more than 15 CAD systems in native format. I think it is one of its kind components as there is hardly any competitor for the bundle of feature it offers.

Last component that would be covered is Hoops 3D by TechSoft 3D. Hoops 3D is advance graphics system which is internally uses OpenGL or DirectX. One of the question often asked is if OpenGL can do all the graphics why another system ? Anyone who had written code in OpenGL would have realize, OpenGL API are pretty low level. To draw the circle you have to 5 line function. For doing hidden line operation, surface rendering one really need to get into fundaments of graphics. Last but not the least OpenGL’s most sophisticated way of saving the data is by mean of display list.

HOOPS3D scores in all of these three points. HOOPS3D is retain mode graphics whereas OpenGL is immediate mode graphics. HOOPS3D provide API to manipulate SceneGraph and high end functionality to create geometric primitives; second it provides simple API to control your rendering pipe line and third has lot application level functionalities. No doubt it accelerates your CAx product development.

Above picture give big picture of HOOPS3D ! As one can see play role of controller and saves you from the pain of low level bug fixing.

Rajesh Bharatiya, Member of Advisory Committee GeomTech and CEO, ProtoTech Solution would be conducting the few HOOPS 3D sessions in Bangalore. Rajesh and his company has been association with TechSoft3D from long time. Mostly we will also have session in GeomTech on HOOPS3D to see understand it better.

Once again it is heartening to see development forums and events in India gaining the momentum. It is not far to that many indigenous commercial CAD/CAM/CFD software from India starts to flood the world market!!!

PS:

CGM -Convergence Geometric Modeler would be also introduce in the seminar. To know more about CGM Visit:http://www.spatial.com/downloads/datasheet_cgm10.pdf

## Tuesday, January 25, 2011

### Brief History of Manifold Topology

**Euclid**300 BC, also known as Euclid of Alexandria, was a Greek mathematician and is often referred to as the Father of Geometry. His work “Elements” is the most successful textbook in the history of mathematics. In “Elements”, the principles of what is now called

**Euclidean geometry**were deduced from a small set of axioms. The geometrical system described in the “Elements” was long known simply as geometry, and was considered to be the only geometry possible. Today, however, that system is often referred to as

**Euclidean geometry**to distinguish it from other so-called

**Non-Euclidean geometries**that mathematicians discovered in the 19th century.

**two- and three-dimensional**

**Euclidean geometry**. An essential property of a Euclidean space is its

**flatness.**Important point to note is other spaces exist in geometry that are not Euclidean. For example, the surface of a sphere is not; a triangle on a sphere (suitably defined) will have angles that sum to something greater than 180 degrees.

**Leonhard Pa**

**u**

**l**

**Euler**(15 April 1707 – 18 September 1783). In 1736, Euler solved the problem known as the Seven Bridges of Königsberg. The city of Königsberg, Prussia was set on the Pregel River, and included two large islands which were connected to each other and the mainland by seven bridges. The problem is to decide whether it is possible to follow a path that crosses each bridge exactly once and returns to the starting point. It is not: there is no Eulerian circuit. This solution is considered to be the first theorem of graph theory, specifically of planar graph theory.

**Euler**also discovered the formula

**V − E + F = 2**relating the number of vertices, edges, and faces of a convex polyhedron, and hence of a

**planar graph**. The constant in this formula is now known as the

**Euler characteristic**for the graph (or other mathematical object), and is related to the genus of the object. The study and generalization of this formula, specifically by Cauchy and L'Huillier, is at the origin of topology.

**Georg Friedrich**

**Bernhard Riemann**(September 17, 1826 – July 20, 1866) an extremely influential German mathematician who made important

**Non-Euclidean geometries**. This discovery was a major paradigm shift in mathematics, as it freed mathematicians from the mistaken belief that Euclid's axioms were the only way to make geometry consistent and non-contradictory. Research on these geometries led to, among other things,

**Einstein's theory of general relativity**, which describes the

**universe as non-Euclidean**. The subject founded by his work is

**Riemannian geometry**. Riemann found the correct way to extend into n dimensions the differential geometry of surfaces. The fundamental object is called the

**Riemann curvature tensor**. For the surface case, this can be reduced to a number (scalar), positive, negative or zero; the non-zero and constant cases being models of the known

**non-Euclidean geometries**. contributions to analysis and differential geometry. He was first one to discover

**Riemannian geometry**is the branch of differential geometry that studies

**Riemannian manifolds**,

**smooth manifolds**with a

**Riemannian metric**, i.e. with an inner product on the tangent space at each point which varies smoothly from point to point. This gives in particular local notions of angle, length of curves, surface area, and volume. From those some other global quantities can be derived by integrating local contributions.

**Riemannian geometry**deals with a broad range of geometries categorized into two standard types of

**Non-Euclidean geometry**,

**spherical geometry**and

**hyperbolic geometry**, as well as

**Euclidean geometry**itself.

**Non-Euclidean geometry**describes

**hyperbolic and elliptic geometry**, which are contrasted with

**Euclidean geometry**. The essential difference between Euclidean and non-Euclidean geometry is the

**nature of parallel lines**. Euclid's fifth postulate, the parallel postulate, is equivalent to Playfair's postulate, which states that, within a two-dimensional plane, for any given line l and a point A, which is not on l, there is exactly one line through A that does not intersect l. In hyperbolic geometry, by contrast, there are infinitely many lines through A not intersecting l, while in elliptic geometry, any line through A intersects l.

**Eule**

**r**was first one publish paper on topology in seven bridge problem demonstrating that it was impossible to find a route through the town of Königsberg that would cross each of its seven bridges exactly once. This result did not depend on the lengths of the bridges, nor on their distance from one another, but only on connectivity properties: which bridges are connected to which islands or riverbanks. In other words to solve many of geometric problems we do not need to know spatial information but what is requiring to be known is neighborhood or connectivity information termed as

**topology**.

**Jules Henri Poincaré**(29 April 1854 – 17 July 1912) was a French mathematician and theoretical physicist, and a philosopher of science.

**He was a founder of topology, also known as “rubber-sheet geometry,” for its focus on the intrinsic properties of spaces.**

**“manifold”**to describe such an abstract topological space.

**Euler- Poincaré equation**

**Euler’s**polyhedron formula was applicable to only simple polyhedron. This new equation from Poincaré provides relationship between topological elements for any single two-manifold body.

**Classification of manifolds**

**0-manifold**is just a discrete space. Eg Point in Cartesian space corresponds to vertex in topological space. Point or set of points are zero dimensional manifolds

**1-manifold**is curve in Cartesian space. Eg circle, line, parabole b-spline curve (non intersecting)

**2-manfold**is surface Cartesian space. Sphere (empty inside), torus, plane, cylinder, b-spline surface (if closed empty inside, orientable and non self-intersecting ) , circular disc

**3-manfold**is 3 dimensional manifold. Eg Solid Objects, Solid Sphere, Solid Cylinder, Our universe J

**Poincaré proposed that all closed, simply connected, three-dimensional manifolds—those which lack holes and are of finite extent—were spheres.**The conjecture was potentially important for scientists studying the largest known three-dimensional manifold: the universe. Proving it mathematically, however, was far from easy. Most attempts were merely embarrassing, but some led to important mathematical discoveries, including proofs of Dehn’s Lemma, the Sphere Theorem, and the Loop Theorem, which are now fundamental concepts in topology.

**Poincaré’s conjecture**had been proved in all dimensions

**except the third**. In 2000, the Clay Mathematics Institute, a private foundation that promotes mathematical research, named the

**Poincaré one of the seven most important outstanding problems**in mathematics and offered a million dollars to anyone who could prove it.

**Grigori Perelman**sketched a proof of the conjecture in a series of papers made available in 2002 and 2003. The proof followed the program of Richard Hamilton. Several high-profile teams of mathematicians have since verified the correctness of Perelman's proof.

**Fields Medal**, which he declined. The Poincaré conjecture remains the only solved Millennium problem.

*This article was written with the help of various sources of web sources, mainly Wikipedia. Intention is to explain interesting subject such as manifold which carries paramount importance in CAD development in less mathematical and elaborative manner. Research in Manifolds is as recent as 2006 and is one of the most studied area in mathematics.*

### Any Midrange CAD software at $99/month and High-end CAD software at $199/month?

Bigger question in future **CAx industry would be Service Industry Vs Product?** Cloud computing model of Software As A Service (Saas) has raised the hope for many. There are many who says CAD has become commodity product my question is whether it will be available at commodity services cost?

Last month when my company was making decision on buying perpetual licenses of Ansys Fluent and Ansys CFX we had big brainstorming session. To buy a product of $50 K for 36 year, with $4K AMC for every year, does that make a sense? What if these products become available on cloud. They have to come at much cheaper rate. One may buy cloud license for 1 month for $300 for the project of $4000, it can be still economical as cost is covered in the project. It like buying home in metro city is always costly than renting the place.

If software vendors really goes for cloud and drop their prices won’t the existing customer who have made enormous amount of investment would feel cheated. Thereby common logic prevails on cloud software won’t cost less than their desktop/workstation counterpart (I have my own reservation on this). This would bring software vendors into catch22 situation. As all new entrepreneurs in CAx industry would like to develop and offer all the future CAx applications in cloud. Reason being they are easy to market, fast to deploy and upgrade. To run the company at zero profit with minimum investment for few years is possible in cloud environment. Similar to what we see in Google, Facebook or Amazon where capturing the market is important. To do this they will need to come up with products which are equally good as existing and at negligible or no cost.

Lower cost or going to cloud is not going to make established big OEM less profitable (eg Google would have not crossed $200 billion mark). There might be small hiccup for year or two. But in long run cloud going to make software far more reachable, easy to deploy, upgrade. This is a completely different business model.

From my prospective cloud computing will bring lot more innovation in CAx space. Industry specific application catering niche category will be spring up in cloud environment. Currently very good example of industry specific could be Homestyler free online home design software (It’s in flash but quality is very good) or project neon photorealistic rendering services. Right now picture for cloud is that it would predominantly serve OEM enterprises, but my logic says it would equally serve new entrepreneurs.

If you have a great idea for CAx application, participate in GeomTech2011 Innovation Contest. Submit your proposal, who know you would win the prizes and peer recognition!!! Join GeomTech2011, Inspiring Innovation. Do write your comments.

## Thursday, January 20, 2011

### Hybrid Computing in CAx products – Next disruptive technology

From designer’s point of desktop or cloud does not make big difference (his employer would have lot!). Rather what he wants is next generation technology CAx applications. CAD, CAM, CFD or CAE, every product today is using same old age technology, same methodology with small difference in user interface and workflow. Amount of innovation you are able to see in silicon chip is less evident our CAx applications. As a matter of fact 95% of CAx software products are not able to use more than one or two core of your Quad core processors. It is sheer wastage of computing resource, but things are going to change with new computing paradigm.

Hybrid computing is aiming to use multiprocessor environment weather it multi core CPU, multi core GPU or any computing processer attach to your desktop. War on high performance is heating up between Intel vs NVIDIA vs AMD. NVidia’s claim of 100x GPU vs CPU was debunked by intel publishing technical report claiming GPUS ARE ONLY UP TO 14 TIMES FASTER THAN CPUS. This paper silently proved NVIDIA’s point GPU is many fold faster than CPU for floating point calculation. **CUDA** (**C**ompute **U**nified **D**evice **A**rchitecture) by NVIDIA and StreamSDK of AMD/ATI has already started making inroads in scientific computing. Both are derived from **OpenCL** (**Open C**omputing **L**anguage) developed Khronos Group. Following graph shows wide gap in computing power between CPU and GPU. Tesla card by NVIDIA has 128 to 1792 processers and thereby no wonder Tesla currently power the fastest supercomputer in the world, Tianhe-1A in Tianjin, China.

It does not mean Intel has given up. Intel has already come up with 48 core processer called “single-chip cloud computer”. 100 core processed is on the way to be made available by other CPU maker.

This heterogeneous computing environment with mammoth computing resource opens Pandora of algorithms that can be used in CAx applications. Hepatic user interfaces, intelligent CAx system can now be realities. No need to wait for Quantum computing, smart programming in hybrid computing can provide required computing. Real time simulations as you define the problem in your CAD/CAM/CFD/CAE application are not distant dream. One would also realize corporations who treat engineering design as secrete would go for hybrid option than cloud for many reasons.

Moving toward hybrid computing will involved re-engineering many of the application or rewriting complete new applications, as highly serial code cannot be parallelize so easily. Computing intensive search structures would be required to be kept on GPU where as main algorithm would required utilizing multi core CPU. Developer in me is exciting to design and program these new age applications on new platforms, Are you excited? Do join us in GeomTech2011 and don’t forget to give your comments.