摘要:
A method for efficient multithreaded analysis of a timing graph is described. The method is applicable to multithreaded common path pessimism removal, critical path traversing for timing report generation, and other types of analysis requiring traversal of sub-graphs of timing graph. In order to achieve high efficiency and scalability for parallel multithreaded execution, the number of access locks is minimized. One parent computation thread and multiple child threads are employed. The parent computational thread identifies the tasks for analysis and distributes them among child threads. Each child thread identifies a sub-graph to be analyzed, creates a thread-specific replica of the identified sub-graph, and performs the analysis required. After completing the analysis, the child thread transfers the results back to the main timing graph and waits for next task. As all data structures of each child thread are accessed only by the child thread owing them, no access locks are required for construction and processing of thread specific graph replica of the timing sub-graph. The construction of each thread specific graph replica is performed by the child thread without locking the main timing graph data structures. Access locks are used only for transferring results of the analysis back to the main timing graph where the results computed by all child threads are combined together.
摘要:
A method for efficient multithreaded analysis of a timing graph is described. The method is applicable to multithreaded common path pessimism removal, critical path traversing for timing report generation, and other types of analysis requiring traversal of sub-graphs of timing graph. In order to achieve high efficiency and scalability for parallel multithreaded execution, the number of access locks is minimized. One parent computation thread and multiple child threads are employed. The parent computational thread identifies the tasks for analysis and distributes them among child threads. Each child thread identifies a sub-graph to be analyzed, creates a thread-specific replica of the identified sub-graph, and performs the analysis required. After completing the analysis, the child thread transfers the results back to the main timing graph and waits for next task. As all data structures of each child thread are accessed only by the child thread owing them, no access locks are required for construction and processing of thread specific graph replica of the timing sub-graph. The construction of each thread specific graph replica is performed by the child thread without locking the main timing graph data structures. Access locks are used only for transferring results of the analysis back to the main timing graph where the results computed by all child threads are combined together.
摘要:
A method of applying common path credit in a static timing analysis in the presence of correlations between asserted arrival times, comprising the steps of using a computer, identifying one or more pairs of asserted arrival times for which one or more correlations exist; propagating to each of the one or more pairs of asserted arrival times a timing value dependent on the one or more correlations; and performing a subsequent common path pessimism removal analysis for at least one test during which a timing value dependent on the one or more correlations between asserted arrival times is used to compute an adjusted test slack.
摘要:
A method of applying common path credit in a static timing analysis in the presence of correlations between asserted arrival times, comprising the steps of using a computer, identifying one or more pairs of asserted arrival times for which one or more correlations exist; propagating to each of the one or more pairs of asserted arrival times a timing value dependent on the one or more correlations; and performing a subsequent common path pessimism removal analysis for at least one test during which a timing value dependent on the one or more correlations between asserted arrival times is used to compute an adjusted test slack.
摘要:
Computing accurately and effectively the impact of clock skew on statistical slack in the presence of statistically variable timing quantities that accounts for both common path credit in the common portion of the clock tree, and RSS credit in the non-common of the clock tree. The clock skew is measured on a per launch and capture path-pair basis as a function of on the post-CPPR path-specific slack (including RSS credit), total mean value of latch-to-latch delay, RSS value of random latch-to-latch delay, test guard time and test adjust. The method includes: performing an initial block-based SSTA including CPPR analysis; selecting at least one launch and capture path-pair for skew analysis; for the at least one path pair, recording post CPPR slack, total mean value of latch-to-latch delay, RSS value of latch to latch delay, test guard time and test adjust; and quantifying the impact of clock skew on statistical slack thereof.
摘要:
Computing accurately and effectively the impact of clock skew on statistical slack in the presence of statistically variable timing quantities that accounts for both common path credit in the common portion of the clock tree, and RSS credit in the non-common of the clock tree. The clock skew is measured on a per launch and capture path-pair basis as a function of on the post-CPPR path-specific slack (including RSS credit), total mean value of latch-to-latch delay, RSS value of random latch-to-latch delay, test guard time and test adjust. The method includes: performing an initial block-based SSTA including CPPR analysis; selecting at least one launch and capture path-pair for skew analysis; for the at least one path pair, recording post CPPR slack, total mean value of latch-to-latch delay, RSS value of latch to latch delay, test guard time and test adjust; and quantifying the impact of clock skew on statistical slack thereof.
摘要:
Systems and methods for accommodating correlated parameters in SSTA are provided. The method includes determining a correlation between at least two parameters. The method further includes calculating a new parameter or a new parameter set based on the correlation between the at least two parameters. The method further includes performing the SSTA such that the new parameter or the new parameter set is propagated into the SSTA. The method further includes projecting slack using the correlation between the at least two parameters and using a processor.
摘要:
An approach that represents and propagates a variational voltage waveform in statistical static timing analysis of digital circuits is described. In one embodiment, there is a statistical static timing analysis tool for analyzing digital circuit designs. The statistical static timing analysis tool includes a variational waveform modeling component that is configured to generate a variational waveform model that approximate arbitrary waveform transformations of waveforms at nodes of a digital circuit. The variational waveform model transforms a nominal waveform into a perturbed waveform in accordance with a plurality of waveform transformation operators that account for variations that occur between the nominal waveform and the perturbed waveform. A variational waveform propagating component is configured to propagate variational waveforms through a timing arc from at least one input to at least one output of the digital circuit in accordance with the variational waveform model.
摘要:
An approach that represents and propagates a variational voltage waveform in statistical static timing analysis of digital circuits is described. In one embodiment, there is a statistical static timing analysis tool for analyzing digital circuit designs. The statistical static timing analysis tool includes a variational waveform modeling component that is configured to generate a variational waveform model that approximate arbitrary waveform transformations of waveforms at nodes of a digital circuit. The variational waveform model transforms a nominal waveform into a perturbed waveform in accordance with a plurality of waveform transformation operators that account for variations that occur between the nominal waveform and the perturbed waveform. A variational waveform propagating component is configured to propagate variational waveforms through a timing arc from at least one input to at least one output of the digital circuit in accordance with the variational waveform model.
摘要:
Techniques for improving parametric chip yield of manufactured chips are provided. In one aspect, a method for optimizing parametric chip yield is provided. The method includes the following steps. Parametric chip yield is computed based on performance and power consumption of a plurality of manufactured chips subject to a given voltage binning scheme. It is then determined whether the parametric chip yield computed is optimal. If the parametric chip yield is not optimal, the voltage binning scheme is altered and the compute and determine steps are repeated. Otherwise the binning scheme is left unaltered.