On-demand performance and IT resource trade-offs
Using the on demand compute resources via cloud computing has shifted paradigms across many industry sectors, including compute intensive and business critical applications, like big data analytics, and pharmaceutical drug design. Until recently though this trend has not reached the semiconductor industry. OneSpin’s cloud solution is changing that.
OneSpin’s 360-CS™ (Cloud Solution) offers easy adoption and usage, on-demand performance and IT resource trade-offs. It provides a patented highly secure solution, in that the design code is never transferred to the cloud.
Formal in the Cloud
Using multiple prove engines concurrently significantly accelerates the formal verification process. The OneSpin cloud solution makes use of formal prove engines in the cloud and manages communication with the cloud engines, retaining the HDL design code locally and handling the debug of the cloud processing results. The user experience is identical to the complete tool running locally, allowing an easy integration into the user’s existing design automation environment. The cloud model also allows for greater financial control, shifting some of the IT infrastructure expense burden from capital budgets to operational or project money sources on an as-needed basis.
Read more about the benefits of OneSpin 360-CS:
Parallel Proof Acceleration
OneSpin’s 360-CS is used to accelerate the performance of a verification run by leveraging additional engines. The cloud allows for any number of engines to be leveraged across the entire cloud, providing the appearance of an infinite source of compute power. Using the cloud solution, the OneSpin 360 DV products can transparently communicate to multiple concurrent OneSpin 360-Prove@OnCE™ engines in the local network as well as in the cloud. The user adds these additional prove engines to the verification environment on demand.
The cloud solution use-model is straightforward. OneSpin 360-DV needs to be installed on a local machine or in the local network. The user will then follow the provided instructions to set up an account at one of the supported cloud services, start the desired number of Prove@OnCE cloud engines, and attach the engines to the local OneSpin 360-DV client.
The inherent capability of formal verification processes to be linearly accelerated through the use of concurrent engines, and the unique cost and capacity control properties of the cloud make this a highly effective performance accelerator.
Assuring Design Security
The security of design intellectual property is a critical issue when leveraging cloud solutions. In these days of corporate espionage and network hacking it is ever harder to protect the company's crown jewels, and an IP transfer to the cloud could be a source of vulnerability. The answer is simple – Don’t transfer the IP at all!
Design security is assured by the OneSpin patented secure cloud solution, processing the IP in the local office and only transferring a mathematical abstraction of specific verification problems to the cloud. With all design detail stripped out, the design information cannot be recreated from this abstraction. The mathematical representation is further encrypted for transfer between the local office and the cloud. After processing, no data is retained in the cloud. This mechanism assures that no intellectual property subject to verification is at risk.
Lower IT Investment
One of the most significant benefits of the cloud solution is the implication on IT budgeting. Traditional EDA IT infrastructure usually requires a significant upfront investment. This requires the customer to have a good understanding of tool resource requirements, particularly hard to estimate with any verification tool, and for varying design projects. In most hardware verification scenarios, there is a bulge of resource usage that gets larger as more code is completed and checked into the design database. This is constantly subjected to regression testing, usually in the form of overnight tool runs. As the RTL coding nears completion, the demand on verification resources will be extensive, and drop off as the design is synthesized.
As such, a certain, predictable proportion of the IT resources will be used throughout the entire design flow, whereas a somewhat unpredictable number of resources will be applied during the bulge period.
OneSpin 360-CS enables a combined model where some resources may be available locally and setup up front while others are applied on-demand using the cloud. As such, the user company is in control of its IT requirements, leveraging exactly what it needs for the variable components of the verification process and eliminating redundant compute power.
With OneSpin 360-CS, the IT infrastructure cost is the same regardless of the number of engines utilized due to an almost linear concurrency to performance trade-off. For example, one engine taking 10 hours is the same IT cost as 10 engines requiring one hour to execute. As such, the only consideration is to find the number of engines required to maximize performance before the trade-off becomes non-linear, maximizing verification throughput.
This model also allows for greater financial control, shifting some of the IT expense burden from capital budgets to operational or project money sources.