Comments on: EDA Is Only “Mostly Dead” http://theasicguy.com/2009/03/04/eda-is-only-mostly-dead/ sharing insights into the people side of ASIC design Sat, 11 Feb 2012 19:55:00 +0000 http://wordpress.org/?v=2.3.2 By: Sarita Shekar http://theasicguy.com/2009/03/04/eda-is-only-mostly-dead/#comment-894 Sarita Shekar Mon, 18 May 2009 07:21:03 +0000 http://theasicguy.com/2009/03/04/eda-is-only-mostly-dead/#comment-894 Harry, you cannot be more right. I've always been very against fine-graining and channelising EDA for semiconductors alone. EDA besides being the simplest means to product development, is also innately broad-spectrum for any industry involving automation of any kind. Global recession or not, EDA would then inevitably assert its stance as a very living entity, had we to understand its genome and its birth in the very first place! Harry, you cannot be more right. I’ve always been very against fine-graining and channelising EDA for semiconductors alone. EDA besides being the simplest means to product development, is also innately broad-spectrum for any industry involving automation of any kind. Global recession or not, EDA would then inevitably assert its stance as a very living entity, had we to understand its genome and its birth in the very first place!

]]> By: Frank Schirrmeister http://theasicguy.com/2009/03/04/eda-is-only-mostly-dead/#comment-506 Frank Schirrmeister Fri, 06 Mar 2009 05:09:45 +0000 http://theasicguy.com/2009/03/04/eda-is-only-mostly-dead/#comment-506 Nice post. I very much agree with the aspects on the importance of software. In exchange there are huge opportunities for EDA and Embedded software. Case in point, probably for the first time there is now an EDA track at the Embedded Systems Conference in late March in San Jose. It is called "Improve Productivity at the HW/SW Interface", see https://www.cmpevents.com/ESCw09/a.asp?option=C&V=1&SB=5 for the tracks. BTW, as I am the track chair I can get interested engineers 25% off the full conference. Just use promo code CTEDA15 to get the discount when registering for ESC. Back on point, there also will be a panel on the topic called "Who’s Taking over Whom – Is EDA moving into Embedded or Embedded into EDA?" on Wednesday, April 1st from 3:00pm – 4:00pm in the ESC Theatre. Nice post. I very much agree with the aspects on the importance of software. In exchange there are huge opportunities for EDA and Embedded software. Case in point, probably for the first time there is now an EDA track at the Embedded Systems Conference in late March in San Jose. It is called “Improve Productivity at the HW/SW Interface”, see https://www.cmpevents.com/ESCw09/a.asp?option=C&V=1&SB=5 for the tracks.

BTW, as I am the track chair I can get interested engineers 25% off the full conference. Just use promo code CTEDA15 to get the discount when registering for ESC.

Back on point, there also will be a panel on the topic called “Who’s Taking over Whom – Is EDA moving into Embedded or Embedded into EDA?” on Wednesday, April 1st from 3:00pm – 4:00pm in the ESC Theatre.

]]> By: Tom Borgstrom http://theasicguy.com/2009/03/04/eda-is-only-mostly-dead/#comment-504 Tom Borgstrom Wed, 04 Mar 2009 21:00:44 +0000 http://theasicguy.com/2009/03/04/eda-is-only-mostly-dead/#comment-504 Great writeup. I think the key point is not that hardware or EDA is dead, but that the relative importance of software as a semiconductor deliverable is growing, and is set to keep growing. Both HW and SW are needed for differentiated semiconductor products. Technologies like system prototypes, combining virtual & rapid prototypes with HW & SW IP, will play an increasingly important role in bringing HW+SW semiconductor products to market faster and more economically. Great writeup. I think the key point is not that hardware or EDA is dead, but that the relative importance of software as a semiconductor deliverable is growing, and is set to keep growing. Both HW and SW are needed for differentiated semiconductor products. Technologies like system prototypes, combining virtual & rapid prototypes with HW & SW IP, will play an increasingly important role in bringing HW+SW semiconductor products to market faster and more economically.

]]> By: Mike Demler http://theasicguy.com/2009/03/04/eda-is-only-mostly-dead/#comment-503 Mike Demler Wed, 04 Mar 2009 17:55:18 +0000 http://theasicguy.com/2009/03/04/eda-is-only-mostly-dead/#comment-503 Great post. and thanks for the plug! My perspective is, as always, a bit (ironic pun intended) different. Software is the 'next' big thing? I disagree. Yes, it's an area in need of better integration into SoC design flows, but that's still only a small part of the overall EDA landscape. Look at my favorite topic - how analog meets digital in design and verification. That is huge, and affects more than just SoCs. There is more analog content and more analog behavior (e.g. power management, leakage, variability,etc.) to account for than ever. But Synopsys has digital DNA, and the analog-digital schism is more difficult for anyone in EDA to wrap their brains around. Great post. and thanks for the plug! My perspective is, as always, a bit (ironic pun intended) different. Software is the ‘next’ big thing? I disagree. Yes, it’s an area in need of better integration into SoC design flows, but that’s still only a small part of the overall EDA landscape. Look at
my favorite topic - how analog meets digital in design and verification. That is huge, and affects more than just SoCs. There is more analog content and more analog behavior (e.g. power management, leakage, variability,etc.) to account for than ever. But Synopsys has digital DNA, and the analog-digital schism is more difficult for anyone in EDA to wrap their brains around.

]]> By: James Colgan http://theasicguy.com/2009/03/04/eda-is-only-mostly-dead/#comment-502 James Colgan Wed, 04 Mar 2009 15:30:13 +0000 http://theasicguy.com/2009/03/04/eda-is-only-mostly-dead/#comment-502 I almost agree Harry ;) I imagine that we'll see an oscillation between s/w and h/w centric system development over the long term. This will be driven by the performance/power/heat-dissipation trade-offs. The market will continue to demand greater functionality (software), but it will also drive battery life (power) and be very demanding of the form-factor (heat dissipation) of their devices. So I imagine that we'll see functionality move from software, to hardware, and back to software over the course of system generations. System developers do as much as they possibly can with the hardware they've got. They make trade-offs and innovate around the limitations of the platform. Then once they've reached a point of "exhaustion" they'll push back on the hardware to deliver more. The job of the semiconductor vendor is to anticipate these demand cycles. And this is why it is imperative for EDA to push further into system level design - provide the tools that a) enable the trade-offs to be made through simulation, and b) automate as much as possible the design, verification, and delivery of the ultimate system...regardless of where the partitions reside. I almost agree Harry ;)
I imagine that we’ll see an oscillation between s/w and h/w centric system development over the long term. This will be driven by the performance/power/heat-dissipation trade-offs.
The market will continue to demand greater functionality (software), but it will also drive battery life (power) and be very demanding of the form-factor (heat dissipation) of their devices.
So I imagine that we’ll see functionality move from software, to hardware, and back to software over the course of system generations.
System developers do as much as they possibly can with the hardware they’ve got. They make trade-offs and innovate around the limitations of the platform. Then once they’ve reached a point of “exhaustion” they’ll push back on the hardware to deliver more.
The job of the semiconductor vendor is to anticipate these demand cycles. And this is why it is imperative for EDA to push further into system level design - provide the tools that a) enable the trade-offs to be made through simulation, and b) automate as much as possible the design, verification, and delivery of the ultimate system…regardless of where the partitions reside.

]]>