We hear a lot from vendors about the new Portable Stimulus Standard, but how does their messaging help me and ultimately you?
I’m pretty upbeat when it comes to portable stimulus. Or maybe it’d be better to say I’m pretty upbeat on the idea of portable stimulus. While doing my best to brush aside the usual EDA propaganda (propaganda I’ve found to be a bit haphazard, but more on that in a minute), I’ve put a lot of thought into how portable stimulus could fit into verification flows, the purpose of using it and the potential value it delivers. I’ve identified entry points with an estimated value and potential for success. From all that, the idea of portable stimulus feels pretty solid.
Then there’s the reality of portable stimulus. And with tool support for PSS-1.0, its a reality set for dramatic change toward an outcome that heavily depends on the role PSS-1.0 plays in portable stimulus adoption.
The Bridge
Visualize PSS-1.0 as a bridge, seamlessly connecting various islands of portable stimulus early adopters to each other and to the mainland of mainstream users. It spans technical impediments related to vendor compatibility thereby unleashing an otherwise prepared and cohesive community of users.
That’d be great, but PSS-1.0 is no bridge. A bridge suggests the existence of a fractured user base ready to unite as part of a portable stimulus free-market. I don’t see that being the case. The early adopter group for portable stimulus is still quite small even though it’s been around a while. It’s difficult to find teams that are using current technology, nevermind challenging it. And I realize vendor compatibility is a necessary part of the roadmap, but I sincerely doubt it’ll be the deciding factor that encourages a meaningful uptick in adoption rates.
Engaging users through vendor compatibility won’t become an effective strategy for growing the market until there are more users to actually engage.
The Chainsaw
With it being too early to bank on vendor compatibility for a portable stimulus inflection point, the bridge analogy is out. That’s fine because a more viable strategy in my view is to present PSS-1.0 as a tool that provides a productivity advantage; something practical, like a chainsaw. When you’re used to cutting down trees with a hand saw all day/every day, a chainsaw is a welcome productivity upgrade. The risk with the chainsaw, however, is that it’ll go through limbs even faster than it’ll go through trees; a catastrophic downgrade in productivity to be aware of before you get started.
Acknowledging the ideological advantages of PSS-1.0, there is a legitimate possibility it becomes a technical liability for portable stimulus. A new language will unearth the usual hesitation. I expect its declarative style will be a hurdle for most of us (I deliberately avoid calling it a declarative language because people much smarter than me have suggested the PSS-1.0 is really a procedural language in declarative clothing). The committee has also admitted that the standard is incomplete and has room to grow. I’ve seen that spun as a positive though it’s tempting to see it as more of a procedural deadlock that ended short of expectations.
All said, if I were a tool vendor I’d be nervous about PSS-1.0 killing more users than it creates. Figuratively, or course. Not literally.
The Mousetrap
This is an easy analogy with a lot of the technical messaging I’ve seen running along these lines; claims that were used with constrained-random and UVM recycled and reapplied to portable stimulus. In particular, I find ‘reuse’ – or better reuse as it were – to be a weak selling feature that sounds like I’m being sold a better mousetrap.
Straight to the point: PSS-1.0 can’t just be a different way to achieve the results we’re already achieving. New technology is fun, but it has to come with discernable value that takes us beyond the mousetraps we already have.
A Tiblosherater
This will sound a tad harsh, but I strongly believe tool vendors are overdue for a wakeup call when it comes to positioning. Up to now, the potential of portable stimulus has been framed as being so broad and so valuable that the usage model has been effectively obfuscated to the point of being non-existent. That’s about as nice as I can say it; that’s from someone who wants to see the tools succeed. Which brings us back to haphazard messaging and the last possibility for PSS-1.0; that it becomes a portable stimulus tiblosherater.
For those new to tiblosherating, a tiblosherater is a fictional tool created to revolutionize its target market while having no clearly stated or understood purpose. In all seriousness, the more people I talk to about portable stimulus the stronger I feel this is where we’re at. We have something that could fundamentally change the way we verify hardware, but no one knows what it does or how to use it.
The Fork In The Road
There’s a lot riding on PSS-1.0. Through standardized language and tooling, vendors are asserting an expectation that portable stimulus is ready for the mainstream. Mission accomplished if mainstream usage ends up being the outcome. If not, it’s a serious blow to its legitimacy and a clear indication that portable stimulus may never be ready for the mainstream.
In my opinion, the path to mainstream starts by narrowing down and nailing down a usage model that provides legitimate benefit to users. That usage model must be practical on a scale that is manageable. We need this desperately. The usage model doesn’t need to be coherent across vendors. But independently it does need to be clear, from all of them. The next step is to turn PSS-1.0 into the productivity chainsaw it’s meant to be through a demonstration of how it facilitates implementation of that usage model and ultimately makes our lives easier. Simple in theory; difficult in practice.
There are no other viable options. A better mousetrap buys time but not success. If we keep tiblosherating I can’t see how portable stimulus lasts much longer. Likewise if we’re not careful with the chainsaw.
 |
 |
 |
 |
 |
 |
 |
 |
 |  |
Great article Neil. All good points. Let me take the side of marketing for a moment, which, like all jobs, has conflicting requirements. On one side marketing needs to provide a vision, a high-level view of what a new technology can deliver. Nobody, neither EDA companies, nor users, want to invest in new technology to solve a small problem for a limited time window. However, I do agree on the need to provide immediate value on targeted use cases. If I look at the development of formal, in the beginning there was little emphasis on targeted use cases. Until Apps arrived. Now formal enjoys widespread adoption and also a variety of use models, including formal signoff and safety verification. But it took quite a few years since SVA and PSL standardization. With the exception of Specman, which took off rapidly despite being a proprietary language, the hardware world hardly see quick changes. UVM has spread quickly but it is a derivative methodology, not radically new. PSS makes sense to me, at least for SoC level verification of use cases at this point. It will likely take some years to see where it goes.
sergio, I think the formal analogy is a good one, you could even go so far as to directly compare sva/psl to pss. along those lines, I guess my hope would be to learn from formal tech adoption by prioritizing the solution over showcasing the technology. And I’d certainly never suggest giving up on big picture potential, but the sooner we synthesize that big picture into real applications/solutions, the better.