Method to provide highly productive qualitative and quantitative analysis of an integrated circuit with minimal compute resource cost
Publication Date: 2010-Nov-23
The IP.com Prior Art Database
AbstractDisclosed is a method to produce highly productive analysis of an integrated circuit. Specifically, a merger of textual and visual interfaces are merged, providing reasonable data with minimal compute cost overhead.
Page 01 of 4
VLSI Hardware design engineers use several mechanisms which aid in analysis of semiconductor designs. These mechanisms tend to fall into two main categories. First, the methods which are most highly productive tend to be very straining on computing resource. For example, an engineer may wish to analyze the physical placement of a design by use of a tool. He must first load the technology rules and the complete design which can involve a significant amount of physical memory, processing time, I/O transactions, etc. A second method involves very limited computing resources but is less productive. For example, the engineer may wish to analyze timing of a design by viewing previously generated textual or visual reports. While this second approach has a low demand on computing resource, it may not provide as much detail as an interactive environment where all the data is available. Moreover, it does not readily integrate several types of data about any one particular situation of interest even though that data may have been captured during the original run in one or more of the existing static reports. As a result, this low-computing method generally takes longer to analyze since data is limited and the engineer needs to manually process or structurally compare much of the information.
A new paradigm is necessary which allows productive analysis without overwhelming computing resources. This idea relies on data produced by other potentially long-running processes, thereby adding a negligible increase to computing demand but providing enough information for a productive analysis.
This invention provides the necessary information to bridge the gap between the existing low-compute/low-productivity and a high-compute/high-productivity solution. The invention provides a continuum wherein each designer can choose what works best in their situation. In other words, he can choose to land at any point on the spectrum and does not need to choose one of the extremes shown in Figure 1.
(This page contains 00 pictures or other non-text object)
Ȉ ˇ˄ ˙ ˝ ˛ ˚
Similar to the existing low-compute method, this new approach mostly relies upon data which has been previously generated. However, like the high-compute method, this approach makes use of a dynamic and visually-based cockpit, thereby increasing overall productivity. Specific details of the invention and a proposed implementation are described. The invention surpasses conventional approaches such as, but not limited to:
Allow designers to work with the data critical to changes that need to be considered
without incurring the overhead of having large amounts of noncritical data also equally easily available
Page 02 of 4
Allow multiple designers to access the same data but look at it totally independently
instead of having to dedicate a whole resource-intensive job to one designer's interactive use
Eliminate the risk of a designer error in interacting with a live day-to-recreate live...