Blog Review: May 23
Reckless driving; whizz-wheels; standards; stats spats; cowbells; subsystems, UVM.
"Indeed, adding an accelerator block to the system may make the worst case power terrible when the block is activated. But..." - MikeB

Deep Insights for Chip Architects and Engineers
Reckless driving; whizz-wheels; standards; stats spats; cowbells; subsystems, UVM.
Monolithic semiconductors for lower-dose x-rays.
Cadence boosts verification speed, rolls out first subsystem; Mentor wins deals for PCB design and UI; Synopsys wins VIP deal with AMD; MIPS squares off against ARM.
Virtualization of more parts of a chip gains attention from large chipmakers, but coherency remains a challenge.
With the issue of system-level model reusability still hotly debated, the use case scenario for models is evolving with increasing design complexity.
Wirebonding is not keeping up with advanced technologies; sophisticated packaging issues must be considered at the very earliest stages of design.
Physics and statistics are making life difficult for digital designers. Simulation tools can help, but Berkeley Design’s CEO explains why statistical modeling and analysis platforms are the real answers.
It’s just a matter of time until EDA is ‘cloudified,’ gradually moving from an IaaS model to a SaaS model.
A quick look at two different business models in the cloud and how they could apply to EDA.
The role of SystemC in verification; interaction with UVM; the need for better coverage; multi-language interoperability; religious wars between designers.
Last of three parts: Cost of models; synchronization problems; growing sense of optimism about co-design; differences between hardware and software; who’s in charge, and who’s being blamed for power problems.
Vendors and chipmakers agree there are gaps between automation tools, but they don’t necessarily agree on what needs to be fixed.
While EDA tools have advanced to enable an impressive level of design complexity, users say usability hasn’t always kept up with functionality.
First of three parts: Exploding complexity; bridging the language and engineering cultural gap; what’s really driving this market; who needs it and who doesn’t.
Second of three parts: Disjointed design schedules; approximately timed modeling; the limits and benefits of RTL, emulation and FPGA prototypes; general-purpose vs. highly specific processors in SoCs.
Last of three parts: Cost of models; synchronization problems; growing sense of optimism about co-design; differences between hardware and software; who’s in charge, and who’s being blamed for power problems.
First of three parts: Stacked die; the impact of software and integration; the changing role of EDA; who will be left standing and what they will make.
Second of three parts: Changes in the supply chain; what happens to little companies; rethinking business models; disruptions caused by stacked die.
Last of three parts: FPGA co-processors; virtualization; designs at 14nm; risks and guarantees for IP and subsystems; the value proposition for stacking die.
First of three parts: What’s behind system-level tool adoption; Japan and Europe lead; complexity surpasses human capabilities; ESL sub-flows around TLM and HLS; the growing challenges of verification in context.
Second of three parts: Stacked die and power; design exploration; standard models and interfaces; the cost of developing models; creating an interoperable ecosystem.
Last of three parts: Addressing software and power earlier; changing economics in design and new competitors; new roles for FPGAs; impacts of stacked die; what standards are needed.
First of three parts: Hardware-software co-design; raising the level of abstraction; pathfinding; finding commonality between designs and increasing re-use; modeling languages.
Changes in the ecosystem and new requirements of SoCs are forcing companies to rethink coherency.
The push toward better performance in mobile devices is changing needs in SoCs designs.
A look at where the problems are in co-design and how much progress we’ve made.
Complexity, time-to-market, Moore’s Law and stacked die are all converging on the market for system-level design tools; growth begins to ramp.
What’s working and what isn’t at advanced nodes for stacked die configurations.
The move to both fully-depleted SOI and FinFETs may be inevitable over the next couple process nodes, regardless of which one comes first.
Sonics CEO Grant Pierce looks at what needs to change in SoC design, what’s driving those changes, and how all of this will be affected by 2.5D and 3D stacking.
What are the design challenges of the leading GPU company and how are they changing?
Bernard Murphy digs down into what’s changing in semiconductor design and how 3D stacking will affect it.
A look at the next big thing in semiconductors and technology and what it’s going to take to get there.