PLM & CAD/CAM Survival through Value-Add Dialogue
Started in CAD/CAM technology in 1966. Spent 31 years of my career with Grumman Aerospace Corporation. Designed, crated and maintained major CAD applications in support of Master Dimensions requirement. I was a key member of the ad-hoc committees that introduced/deployed CADAM and subsequently … More »
August 18th, 2010 by Bill Abramson
Jeffery Katzenberg, American film producer and CEO of DreamWorks Animation was in a recent meeting with Intel executives and suggested a need to “revolutionize the way we work” in order realize the benefits of the technology we have access to today. If I might, let us explore the ramifications of that statement.
Let’s explore some opportunities that might offer significant improvements in your organizations utilization and deployment engineering assets if you choose to revolutionize the way you work.
First recognize your most important asset – your engineer. They are users of technology and they employ it to design, test and modify their ideas. If they exploit all the available compute cycles, they might actually do it faster. That means an organization will be closer to producing an optimal product, rather than adequate product that may use more material, be more expensive to manufacture or potentially compromise on its original product requirements. Optimal products tend to generate higher returns and long term customers both of which a good for the bottom line. Simple goal – design and manufacture products closer to the optimal design without prolonging the development cycle.
Two technologies that play a role in your product development are workstations and HPC resources. However, their definitions have changed; today’s workstation with two Intel® Xeon® processors, almost 200GB of memory and over a tera-byte of storage looks and sound like an HPC solution I used to develop products. Barbara Hutchinson from ANSYS Corporation has suggested that small & medium CAE and CFD simulations might even be more efficiently run on a workstation. She suggested that by eliminating the time to transfer data, queue up a job, process it, and retrieve the results that it might be possible to look at today’s workstation as powerful workbench that designs and computes. The added benefit is your most prized investment, your engineer, can now power through more designs in a day. This will help him or her stay at the top of the innovation curve.
Now what happens if you have an investment in an HPC resource with all these new powerful workstations? After all you just removed a large number of jobs from its queue. As Seymour Cray suggested, HPC systems are designed to make un-tractable problems tractable. They are best used for the large complex problems that really need an answer fast. By removing small and medium jobs from an HPC queue, you effectively remove parasites that steal cycles away from the big hairy problems that need answers fast. You also give your engineer a tool that helps them quickly iterate through more ideas in less time than ever before. They can refine their ideas and when it is time to increase their models fidelity, they are more likely to make better use of the HPC cycles they need to solve their big hairy problem that they were refining on their workstation. You just optimized three critical assets of your product development cycle – your engineer, your workstation and your HPC investment.
Let’s look at technology another way.
His team, along with Intel and HP technologist embarked on a journey to identify and parallelize software in order exploit multi-core technologies. That is the obvious answer to making software work faster on today’s technology. DreamWorks also looked at the way they worked. They asked how we might possibly reinvent ourselves in order to be more productive – to create new innovative 3D movies faster. The idea was obvious – they needed to do what they had done to their software. They need to THINK PARALLEL.
The result was a remarkable increase in compute utilization and an even more impactful set of block buster hits with the most amazing special effects in less time than ever before.
Intel took the idea too; they looked at their engineering core and asked what if we armed our engineers with more powerful workstations – what might happen? The answer was obvious, like DreamWorks, when they revolutionized the way they worked, Intel saw an increase in engineering productivity and they also saw a new way to increase the number of available compute cycles their engineers required to create the next new micro-architecture. Their pilot is showing them a new way to add 400 racks of compute capacity without having to increase investments in data center, power and cooling infrastructure.
As you can see I like revolutions when it comes to changing the way we work and exploiting new technology. The revolution is just starting too.
We are currently working at a number of large aerospace, automotive and digital media companies where together they are piloting what it means to potentially revolutionize the way engineers work. Not just employing parallel software, but thinking in parallel too. We are looking at multi point design optimization strategies and asking how we might be able to exploit workstation technologies to reduce the time between an idea and a product.
Intel, has been very excited about their research and pilot efforts in discovering solutions and innovations, where critical workflows can be truly optimized from the perspectives of the quality, product development time and cost by configuring the technology that specifically benefits those specific critical work processes.
The belief is that the successes realized to-date can be readily extrapolated to other enterprises. The benefits to industry are so great and the timing is so opportunistic that Intel in concert with key hardware developers is investing significantly in these efforts.
We welcome your feedback in any of these areas. Please feel free to contact me at:
July 15th, 2009 by Bill Abramson
I have had a recent discussion with someone who was describing their very seious intraoperability issues. Areas such as flight test and field support experience major problems in accessing product definitions in a form that they can use. Applications used by engineering and manufacturing … among others … do not provide the critical functions these other areas require forcing them to use other non-compatible applications.
What do you think?
June 5th, 2009 by Bill Abramson
These are incredibly challenging times. We are being forced to continue to operate successfully, maintain our competitive position and do it with major reductions in people and resources. In my past experiences, we have helped people in our network address so many of their issues by establishing, informally, direct lines of dialogue between concerned parties. We found that by discussing our lessons learned, we were able to make some sound decisions. To initiate this blog, let’s have a discussion in the following areas: