First Version Of Russian Geometric Kernel Announced

According to this press release on isicad, the first version of the Russian Geometric Kernel (RGK) will be shown at COFES Russia 2013 to be held at St. Petersburg at the end of May. Top Systems and Ledas, two Russian CAD software development companies have been working as contractors for STANKIN Moscow State Technlogical University developing this kernel for some time now.

The press release came with this screenshot of what appears to be a test application using the kernel.

One may question the need for another modeling kernel. Top Systems CTO Sergey Kozlov made this very interesting observation in the press release:

“The most significant practical tasks carried out by our company was the development of the general architecture of the 3D kernel enabled by modern multi-core platforms and graphical computing devices. The result will be the ability to develop end-user applications that run on many platforms and operating systems.”

Top Systems has its own MCAD product called T-Flex, which uses Parasolid as its modeling kernel. So I’m guessing they could be one of the first companies to dump Parasolid and use RGK instead.

And speaking of Parasolid, SolidWorks CEO Bertrand Sicot wrote in a blog post today that there was no end-of-life plans for the Parasolid based version of SolidWorks. For a long time now SolidWorks management has been maintaining that customers don’t care about which modeling kernel is being used by their CAD system. I’m glad they have started to think otherwise.

  • “So I’m guessing they could be one of the first companies to dump Parasolid and use RGK instead.”

    No! Sergey Kozlov yesterday said that they have no plans to change kernel. “This is not advisible”.

    If developer of a kernel thinks that using it is not advisible, what other developers can think? 🙂

    • That is very interesting. Did he elaborate on why it wasn’t advisable?

      • It was something about labour-intensiveness…

        • Absolutely makes sense. I’ve had several interview conversations with Dr. Biplab Sarkar who was with PTC and has been CTO at Nemetschek Vectorworks for years now and the labor explained in moving to a new kernel can be significant.

    • Владимир Малюх


      If DS SW developers do believe that it’s better to leave original SW with Parasolid, rather to switch to DS CGM, or also DS owned ACIS- what shall we think? 🙂

      In fact, switching the kernel means helluwa lot of troubles for existing customers, who own heap of legacy data. How they must deal with their beloved data after such switch?

      I strongly believe that true concept is “new kernel is for new applications” but no for replacing kernels in the software with long history.

      • Absolutely. Changing kernels can be big work for the developer and there is serious financial reasons also. Not just about legacy data. Nemetschek Vectorworks switched kernels (went to Parasolid) about 4 versions back and they have just now finished up converting all the elements of their application to the new kernel. They spread out the work over many years so they could reserve developer resources for new features.

        • Jon Stevenson

          two products that Biplab worked on in the past switched kernels: SolidEdge from Intergraph, Designwave/Pro/Desktop from Computervision/PTC. Although I believe both occurred after Biplab had moved on.

      • ralphg

        Or, they can do what IronCAD does: run both ParaSolid and ACIS kernals in one MCAD program.

  • CAD Guy

    The decision of whether or not it is immediately profitable to replace an existing modeling kernel would mainly depend on:

    1) Near term product road map.
    2) How well the CAD application is engineered/designed from the beginning so as the geometric modelling code is modularized.

    Having said that, the long term benefits are obvious. The dispute between DS and ADSK over kernel is one example. Vectorworks adopting Parasolid is another example.

    There are only a handful of modeling kernels currently in the industry and they are all owned/controlled by companies that have their own competing products/interests. Moreover, all these kernels CGM, Parasolid, ACIS, Shape Manager, Granite et al were originally developed for the previous generation of Design Products. They don’t take advantage of advances in hardware and software technologies.

    Just as now developers don’t have to write assembly code when developing a website, lots of functionality that was originally done at the application level can now be done instead in the geometric kernel much efficiently.

    Hence, a case for a new modeling kernel is completely justified.

    What would be even better is an open source geometric kernel or a geometric kernel from a neutral vendor. That could unleash the innovation in the CAD/Design industry !

    • murray

      There is an open source kernel, OpenCascade, nowhere near as developed as proprietry kernels, and the open-source CAD apps that use it are understandably underdone compared to commercial equivalents. There’s an extensible/educational CAD app backed by the Indian government education and departments that uses it, though.

    • ralphg

      Autodesk did a major rewrite of ShapeManager over the last three years, so that it distributes work between the desktop and server CPUs.