Thoughts on IC development, EDA, and hardware design languages.

Industry Insights: Why is engineering productivity not a top-down goal?

Posted by: Brad Quinton | Posted on: January 27th, 2015 | 2 Comments

Share on LinkedIn31Share on Facebook0Tweet about this on Twitter0Share on Google+0Email this to someone

Over the last 15 years I’ve been on both sides of the EDA/Semiconductor fence. I’ve used EDA tools to develop ICs, and I’ve developed and sold EDA tools. I’ve worked for companies big and small, even started my a few of my own, and one constant I have observed is that the demand for productivity enhancements flows from the engineers on the ground up into the “business side”, not the other way around. On the surface this is surprising. High quality engineering talent is a rare commodity, you would expect that companies would be scrambling to improve productivity to get the most out of these rare (and relatively expensive) resources.

You’d think that spending a few thousand dollars for 50% increase in productivity for an existing engineer would be a “no brainer” — a much better bet than delaying a project, hiring more resources, or just producing fewer designs. You’d think that teams of accountants would descend on the CAD groups in the semiconductor organizations, demanding to know what innovative new products they were considering to improve the output of their R&D teams. But, in my experience, it doesn’t happen like that. I’m curious to know why.

Don’t get me wrong. Engineers in our industry do get more productive by using better tools and processes (people buy our Invio product today with that in mind, so I am banking on it!). In my experience this happens because the engineers on the ground pull along the “business side”, usually kicking and screaming, not the other way around. Today’s IC engineer designs (or verifies) far more transistors than ever before (see p.3, in these interesting slides from Harry Foster), but it is the engineer on the ground that demands (or begs, or pleads, or cajoles) for faster emulation, access to UVM simulators, more CPUs, more licenses, custom internal tools, formal verification, etc. Given that greater productivity means there is actually less need for engineers it really begs the question as to why directives for better/faster tools and flows are not coming top-down from management? Why are the engineers the ones fighting for better tools?

Maybe my sample set is not big enough? Maybe I’ve just not interacted with the right companies? Or maybe this behaviour makes sense, and smart MBAs will tell me it is always the job of corporate finance to hold tight to the purse strings, keeping the engineers from foolishly spending money? I would love to know of any stories from people out there where someone in charge of the budget (accountant, VP, GM, whatever), came to them saying “What we need to do is spend more money on better tools to get more out of our current teams!”. If you have such a story, please share. Or, if you have a theory why this scenario exists (or even why it makes sense) I’d love to hear that too!


Share on LinkedIn31Share on Facebook0Tweet about this on Twitter0Share on Google+0Email this to someone

Comments (2)

  1. Amir Nahir - Reply
    January 30, 2015

    The opinion expressed below is mine, and not related to my current or past employers.

    Productivity will always be primarily driven by the engineers, for two reasons:
    (1) The desire to be more productive originates from a deep human desire to accomplish more. Sometimes it’s just the desire to go home earlier and spend more time with the family, and (more often), it’s the desire to have a stronger confidence in the quality of one’s work, and the ability to accomplish more in less time (and not having the feeling that a person is wasting his/her time on something that a tool/machine can do).
    (2) It is those engineers that best know which of their tasks can be done better/faster via automation.

    So, in light of that, I’d like to rephrase your question to “Is engineering productivity sufficiently supported by senior management?” (and propose the answer: NO!)
    The way I see it, the insufficient support for productivity improvements originates, primarily, from the focus on accomplishing short-term goals: (1) Productivity aids never solve today’s problem today. Almost any tool I’ve seen (excluding some IDEs) require significant effort today (modelling some parts of the design, modifying RTL to support some tool…) and will only bear fruit later; (2) From an organisational perspective, in case the people developing in-house productivity tools are in the same team as the execution team, it is easy to have them assist in solving the current crisis then to justify why they are working on something that doesn’t support the main goal *today*.

    When I was managing a group in IBM Research, one of the goals I gave to one of my team leads was to build his tool to give “above linear ROI” – that is, the moment the person installs the tools and puts in a minimal effort to operate it, that person will see some value back. This requirement translates to some technical requirements, demanding that the tool adheres to synthesis outputs in their current forms, etc. I can’t really say more, as I don’t know what is already published about this tool. I can suggest to EDA tool developer to consider this approach (where it can be done) to ease the transition to using their tools.

    • Brad Quinton - Reply
      January 30, 2015

      A great answer, Amir. Thanks!

      I think your analysis that engineers are driven to be more productive by their very nature is very insightful. I think that answers the question why they are willing to work so hard, and take some much career risk to increase prductivity with so little support from management. That is something I am going to spend more time thinking about. The desire of engineers to just do their job better is something that I personally have seen time and again. The question is how to nurture and enable them…

      And, your re-statement of my question is bang-on, that is the question I was going for: “Is engineering productivity sufficiently supported by senior management?”

      And I agree with you assessment, as well. I even have to admit, when I have been in project management roles, the drive to solve “todays issue” is very strong.

      Thanks again for the great feedback. Lots of “food for thought” here.

Leave a Comment