Method and apparatus for addressing and improving holds in logic networks
View Patent ↗A method and apparatus for modifying a synchronous logic network so that the hold slack calculated at all pins is greater than or equal to a user-specified threshold, with the condition that the setup slack at any pin does not become negative or smaller than a user-specified margin. The result is an improved design which is less likely to fail due to a hold time violation. The method and apparatus introduce a limited number of logic cells which helps keep power consumption and design size to a minimum.
1. A method on an apparatus for improving hold slack in a logic network comprising:
a processor of the apparatus performing static timing analysis to determine a setup slack and a hold slack for each pin in the logic network;
the processor determining a critical network based on the hold slack for each pin in the logic network;
the processor calculating a critical congestion for each pin in the critical network;
the processor grouping pins into pin classes where pins with a same hold slack are grouped in a same pin class;
the processor ranking each pin in the same pin class based on the critical congestion of the pin, wherein a pin having a highest critical congestion is ranked highest;
the processor tagging the highest ranked pin as an insertion point; and
the processor inserting a delay cell at the insertion point.
2. The method of claim 1 , wherein the critical network is defined by the hold slack for each pin in the logic network that is less than a user-specified threshold.
3. The method of claim 1 , further comprising:
the processor calculating a general congestion for each pin in the logic network.
4. The method claim 3 , wherein the general congestion for each pin is used to break a tie between pins with the same critical congestion.
5. The method of claim 4 , wherein the pin having a lowest general congestion is ranked highest.
6. The method of claim 1 , further comprising:
the processor determining whether each ranked pin has been previously tagged.
7. The method of claim 6 , wherein determining comprises:
the processor determining whether each ranked pin has been previously tagged and for each ranked pin determined to not have been previously tagged, inserting a delay cell.
8. The method of claim 1 , further comprising repeating the steps of performing, determining, calculating, grouping, ranking, tagging and inserting until no further pin is tagged as an insertion point.
9. The method of claim 1 , wherein calculating critical congestion is performed by levelization.
10. The method of claim 1 , wherein calculating critical congestion is performed by topological sorting.
11. The method of claim 1 , wherein calculating critical congestion is performed by depth first search.
12. An apparatus for improving hold slack in a logic network comprising:
a static timing analyzer configured to perform static timing analysis and output a setup slack and a hold slack for each pin in the logic network; and
a processor configured to:
determine a critical network based on the hold slack for each pin in the logic network;
calculate a critical congestion for each pin in the critical network;
group pins into pin classes where pins with a same hold slack are grouped in a same pin class;
rank each pin in the same pin class based on the critical congestion of the pin, wherein a pin having a highest critical congestion is ranked highest;
tag the highest ranked pin as an insertion point; and
insert a delay cell at the insertion point.
13. The apparatus of claim 12 , wherein the critical network is defined by the hold slack for each pin in the logic network that is less than a user-specified threshold.
14. The apparatus of claim 12 , wherein the processor is further configured to calculate a general congestion for each pin in the logic network.
15. The apparatus of claim 14 , wherein the general congestion for each pin is used to break a tie between pins with the same critical congestion.
16. The apparatus of claim 15 , wherein the pin having the lowest general congestion is ranked highest.
17. The apparatus of claim 12 , wherein the processor is further configured to determine whether each pin has been previously tagged.
18. The apparatus of claim 17 , wherein the processor is further configured to insert a delay cell when the ranked pin has not previously been tagged.
19. A non-transitory computer-readable storage medium storing a set of instructions for execution by one or more processors to perform hold slack improvement, the set of instructions comprising:
a static timing analysis segment for performing static timing analysis to determine a setup slack and a hold slack for each pin in the logic network;
a critical network segment for determining a critical network based on the hold slack for each pin in the logic network;
a critical congestion segment for calculating critical congestion for each pin in the critical network;
a grouping segment for grouping pins into pin classes where pins with a same hold slack are grouped in a same pin class;
a ranking segment for ranking each pin in the same pin class based on the critical congestion of the pin, wherein a pin having a highest critical congestion is ranked highest;
a tagging segment for tagging the highest ranked pin as an insertion point; and
an inserting segment for inserting a delay cell at the insertion point.
20. The non-transitory computer-readable storage medium according to claim 19 , wherein the set of instructions are hardware description language (HDL) instructions used for the manufacture of a device.