> What's interesting is that Gary listed NO bughunters, NO statistical timing
> analysis, NO Verilog/VHDL simulation, NO test, NO physical verification, NO
> constraint reduction tools, NO synthesis, and NO analog/SPICE/RF tools this
> year (while I defintely did.)  Over the past 5 years, Gary's had a strong
> preference for "virtual prototyping" tools, so I guess he decided to kill
> off all these other extra tools to pimp SW/ESL even more so this year.
> 
> The major difference in our "must see" lists?  Gary's big into pretending
> about chips being made, while I'm big into chips actually being made.  :)
> 
>     - from http://www.deepchip.com/wiretap/060822.html


From: George Harper <gharper=user domain=bluespec spot mom>

Hey, John,

While I thought your play on "pretending" was a nice touch, I wanted to offer
another perspective.

Push-Me, Pull-You

I believe the differences in your two DAC "Must See" lists are more rooted in
perspective.  My sense is that your opinion leverages heavily on real-world
feedback and public proof of adoption for validation.  Gary, in contrast,
tries to assess which solutions address current & anticipated customer issues.

In your case, John, your perspective is naturally retrospective; tools that
have user-proven track records get more weight in your view.

In Gary's case, he takes a more prospective view given the bets he must take
on which solutions are most compelling.  This requires a willingness to
take public risks (albeit calculated) and lead, something you don't do, John.
Identifying disruptive innovation is difficult if you only look backwards 
listening to current customer opinions and requests.

Because of these perspective differences, it's harder for a newer solution
to get your attention, John, even if it's had compelling, though non-public
customer successes.

Take Bluespec as an example.  While we've made Gary's list since 2004, we
didn't make your 2006 "Cheesy Must See List for DAC" in:

> 10.) The real fight I see in the much ballyhooed ESL niche boils down to
>      Forte SystemC vs. Mentor ANSI C.  At Forte (booth 1428), they'll be
>      showing their Cynthesizer that now supposedly does "TLM synthesis"
>      which they claim is now TLM ANSI C => SystemC => RTL => gates.  Ask
>      for Brett Cline.  Freebie: weird computer monitor sweepers.  At
>      Mentor (booth 928) their new Catapult SL does pure ANSI C++ => RTL
>      and it supposedly has some sort of new hierarchy "channels" in it.
>      Ask for Shawn McCloud or Anil Khanna.  Freebie: one-armed backpack.
>
>          - from http://www.deepchip.com/gadfly/gad072006.html


But you changed your mind at DAC and then Bluespec got recognized at #10 by
you in your own Best of DAC 06 video list in:

             http://www.deepchip.com/wiretap/060810.html

Is ESL only about elevating DSP Algorithms?  Pre-DAC, your assessment was
that ESL boiled down to a match between the two most well-known algorithmic
synthesis solutions: "The real fight I see in the much ballyhooed ESL niche
boils down to Forte SystemC vs. Mentor ANSI C."

There's no doubt that Mentor Catapult C and Forte Cynthesizer each offer
interesting algorithmic synthesis solutions.  But while they may be good for
DSP filters, FFTs, and audio processors, these algorithmic synthesis tools
don't offer a significant advantage over Verilog or VHDL for the bulk of
gates that are shipped today, including microcontrollers, DMA controllers,
cache and memory controllers, bus/switch interconnects, bus interfaces,
network/link layer controllers, sorting/queuing engines, finite state
machines, processors (whether CISC/RISC/DSP/graphics), etc.

This is because when it comes to designs with control logic and complex
datapaths (which I define as those datapaths with control complexity to
them such as bypassing, MUXing, feedback, and stalling) unless you are
using Bluespec, you're still dealing with RTL.

I hope by now, post-DAC (unless you still mistakenly think ESL is meant to
be only for trivially parallelized DSP SW algorithms) for you ESL should
certainly now include high-level synthesis solutions for control logic and
complex datapaths.  Bluespec currently is the only one with that solution.

I don't always agree with Gary Smith, and that should be expected when
someone's trying to call the future, but I do respect much of what he says
and his belief in his convictions.

    - George Harper
      Bluespec, Inc.                             Waltham, MA