Track distribution method and system for integrated circuit wiring and medium
By modeling the track allocation problem as a scheduling problem and solving it using mathematical models, the problems of multiple cost factors such as overlapping of line segments and pins in integrated circuit wiring are solved, and a faster and better track allocation solution is achieved, reducing wiring costs , improves the efficiency and cost control of integrated circuit wiring.
Patent Information
- Application Number
- CN202411013095.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-26
- Publication Date
- 2024-11-22
AI Technical Summary
The existing track distributors fail to effectively consider the pin overlap cost of the line segment and other networks in integrated circuit wiring, resulting in insufficient optimization of track distribution effect, and ignore the cost of local pin winding in detailed wiring.
Using a mathematical model solution method based on scheduling problems, multiple cost functions are set, including line length cost, line segment overlap cost, pin overlap cost and obstacle overlap cost. By modeling the track allocation optimization problem as a scheduling problem, using the scheduling algorithm Perform the solution to obtain the optimal track allocation scheme.
It significantly reduces wiring costs, improves the efficiency and cost control of track allocation, and provides a better track allocation solution, which is faster than traditional methods.
Smart Images

Figure CN119005111A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of integrated circuit computer-aided design, and in particular relates to a track allocation method, system and medium for integrated circuit wiring. Background Art
[0002] Routing is a key step in the design of very large scale integrated circuits (VLSI). The routing process mainly includes overall routing and detailed routing. The goal of overall routing is to divide the routing area into global units to give a set of straight line segments passing through several global units. In the overall routing stage, the connection and details between specific wire nets are not considered, but the overall layout and path planning are mainly focused on. The results of overall routing will provide guidance and basis for the subsequent detailed routing stage. Detailed routing is a stage in the physical design of very large scale integrated circuits (VLSI), which is carried out on the basis of overall routing. The task of detailed routing is to realize the specific connection between the pins and straight line segments in each wire net, that is, to determine the final path of the wire net on the chip. The result of routing is related to the production cost and product quality of chip design, so it is crucial to the successful completion of the physical design of very large scale integrated circuits.
[0003] Track allocation is an intermediate step between overall routing and detailed routing. Track allocation allocates the straight line segments obtained by global routing to the grid lines (i.e., tracks) of the global unit before detailed routing. The introduction of track allocation has greatly improved routing efficiency. In track allocation, it is necessary to minimize the overlap between components and the length of the wiring lines. In order to better guide detailed routing, track allocation needs to estimate the impact of the allocation method on the cost of detailed routing and select an allocation method with a lower detailed routing cost. In existing track distributors, the overlap between line segments, the overlap between line segments and obstacles, and the line length are mostly included in the routing cost in track allocation. The overlap between line segments and pins of other network lines is not considered, and the cost of local winding of pins in detailed routing is ignored, making the effect of track allocation less than optimized. Summary of the invention
[0004] The purpose of the present invention is to propose a track allocation scheme for solving the mathematical model of the scheduling problem in integrated circuit wiring. A variety of different cost functions can be set to solve the track allocation problem considering the overlap between line segments, the overlap between line segments and obstacles, the overlap between line segments and pins of other wire networks, and the line length.
[0005] The present invention achieves the above-mentioned purpose by adopting the following technical scheme:
[0006] A track allocation method for integrated circuit wiring comprises the following steps:
[0007] 1) Reading the circuit information after the overall layout of the integrated circuit, including the line segments to be allocated, the positions of obstacles and the positions of pins;
[0008] 2) Define the wiring cost function, take the wire length cost, line segment overlap cost, pin overlap cost and obstacle overlap cost into the wiring cost in track allocation, and perform weighted summation;
[0009] 3) Based on the wiring cost function, the track allocation optimization problem is modeled as a mathematical model of the scheduling problem;
[0010] 4) Solve the scheduling problem of the mathematical model;
[0011] 5) If there is a feasible solution, convert the solution into a track allocation plan.
[0012] Furthermore, in step 1), the line segment to be allocated is represented by the coordinates of the starting point and the end point on the wiring area, and the obstacle position and the pin position are represented by the coordinates on the wiring area.
[0013] Furthermore, the rule for track allocation in step 2) is that each routing area has only one track in the X-axis or Y-axis direction for routing.
[0014] Furthermore, in step 2), the line length cost is the cost generated by the line length from the line segment to the connected pin in the track allocation; the line segment overlap cost is the cost generated by the overlap between two line segments allocated to the same track; the pin overlap cost is the cost generated by the overlap between the line segment and the pins of different wire nets; the obstacle overlap cost is the cost generated by the partial or complete overlap between the line segment allocated to the track and one or more obstacles.
[0015] Furthermore, in step 3), the step of modeling the track allocation optimization problem as a mathematical model of a scheduling problem includes:
[0016] Represent the line segment as a task variable with fixed start and end time, represent the track as a machine variable in the scheduling problem, and transform the track allocation decision into a machine allocation decision variable in the scheduling problem;
[0017] The wiring cost function is set as the objective function, with the goal of minimizing the total cost;
[0018] The relationship between the machine allocation decision variables and the routing cost function is taken as a constraint;
[0019] A mathematical model of the scheduling problem is constructed based on task variables, machine variables, machine allocation decision variables, objective functions and constraints.
[0020] Furthermore, the step of solving the scheduling problem for the mathematical model in step 4) includes:
[0021] The mathematical model is input into the scheduling problem solver and the scheduling algorithm is used to solve the problem within the set time.
[0022] If the current solution is feasible, the current solution is returned; otherwise, the iterative solution is continued.
[0023] A track distribution system for integrated circuit wiring includes a memory and a processor. The memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.
[0024] A computer-readable storage medium stores a computer program, which implements the steps of the above method when executed.
[0025] The beneficial effects achieved by the present invention are as follows:
[0026] The present invention proposes a track allocation method for integrated circuit wiring. By setting a variety of different cost functions and using mathematical models to transform the track allocation optimization problem into a scheduling problem for solution, the track allocation problem taking into account multiple cost factors such as line segment overlap, obstacles and pin overlap is effectively solved. Compared with the traditional method, the method of the present invention not only has a faster solution speed, but also can significantly reduce the wiring cost, thereby providing a better track allocation solution and further optimizing the efficiency and cost control of integrated circuit wiring. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 It is a flow chart of a track allocation method for integrated circuit wiring of the present invention.
[0028] Figure 2 It is a flow chart of a track allocation method for integrated circuit wiring according to an embodiment of the present invention. DETAILED DESCRIPTION
[0029] In order to make the various technical features and advantages or technical effects in the above technical solutions of the present invention more obvious and easy to understand, they are described in detail below with reference to the accompanying drawings.
[0030] The overall flow chart of the track allocation method for integrated circuit wiring proposed by the present invention is as follows: Figure 1 As shown, it mainly includes the following steps:
[0031] Step 1: Read in the circuit information after the overall layout of the integrated circuit, including the line segments to be allocated, obstacle locations, and pin locations.
[0032] Step 2: Define the routing cost function, factor the wire length cost, line segment overlap cost, pin overlap cost, and obstacle overlap cost into the routing cost in track allocation, and perform a weighted sum.
[0033] Step 3: Based on the routing cost function, the track assignment optimization problem is modeled as a mathematical model of the scheduling problem.
[0034] Step 4: Solve the scheduling problem using the mathematical model.
[0035] Step 5: If there is a feasible solution, convert the solution into a track allocation plan.
[0036] The embodiment of the present invention discloses a specific implementation process of a track allocation method based on a scheduling model in wiring, such as Figure 2 The detailed description is as follows:
[0037] Step 1: Read the circuit information after the overall layout of the integrated circuit to obtain the line segments to be allocated, the positions of obstacles and the positions of pins. There are straight lines in the X-axis and Y-axis directions in the wiring area, called tracks. Each wiring area can only have one track in one direction for wiring, and this direction is also determined in the overall layout. The line segments to be allocated are described by the starting point and the end point, and the positions of obstacles and pins are described by the coordinates on the wiring area. The line segments and pins have their own wire nets, which are distinguished by the wire net numbers.
[0038] Step 2: Construct the routing cost function.
[0039] Definition 1: “Wire length cost”: The wire length of the connection from the wire segment in track allocation to the pin to be connected. The cost incurred is the wire length cost.
[0040] Definition 2: “Line segment overlap cost”: The overlap between two line segments assigned to the same track is called line segment overlap, and the cost incurred is called line segment overlap cost.
[0041] Definition 3 “Pin overlap cost”: The overlap between a line segment and pins of different nets is called pin overlap, and the resulting cost is called pin overlap cost.
[0042] Definition 4 “Obstacle overlap cost”: The cost incurred when a line segment assigned to a track partially or completely overlaps with one or more obstacles is called obstacle overlap cost.
[0043] Describe the above costs with mathematical formulas and perform weighted summation to obtain the wiring cost function.
[0044] Step 3: Construct a mathematical model of the scheduling problem.
[0045] Step 3.1: Define variables: abstract the line segment into a task variable with fixed start and end time, abstract the track into a machine variable in the scheduling problem, and transform the track allocation decision into a machine allocation decision variable in the scheduling problem.
[0046] Step 3.2: Define the objective function: Set the routing cost function in step 2 as the objective function, with the goal of minimizing the total cost.
[0047] Step 3.3: Define constraints: Describe the relationship between the machine assignment decision variables and the routing cost function in step 2 as constraints.
[0048] Step 3.4: Establish mathematical model: Based on the defined variables, objective function and constraints, a mathematical model of the scheduling problem can be established to describe the track allocation optimization problem.
[0049] The following is a specific example:
[0050] In this example, there are n line segments to be assigned, and the endpoint position of the i-th line segment is [a i ,b i ], m tracks, pin set Pin, forbidden area set Block. Then the mathematical model of scheduling problem is established as follows.
[0051] (1) Variables:
[0052] Task 1 …Task n ∈[1…n]: Define n tasks, representing n line segments, and the start and end time of the task represent the two endpoints of the line segment.
[0053] Machine[1…n]∈[1…m]: Treat the track as a machine and define which machine each task is on, that is, the track of segment allocation.
[0054] ST[1…n]: The start time of the task, corresponding to the starting point of the line segment.
[0055] ED[1…n]: The end time of the task, corresponding to the end point of the line segment.
[0056] (2) Expression:
[0057] Define the overlap length function overlap(Task i ,Task j )=|max(ST[Task i ],ST[Task j ])-min(ED[Task i ],ED[Task j ])|, represents the overlapping length of tasks i and j. When i and j are placed on the same machine, the overlapping length cost of i and j will be incurred.
[0058] Define the cost function costp(Task i ,Machine[Task i]) represents the pin overlap cost of task i on the corresponding machine.
[0059] costb(Task i ,Machine[Task i ]) represents the forbidden area overlap cost of task i on the corresponding machine.
[0060] costd(Task i ,Machine[Task i ]) represents the line length cost of task i on the corresponding machine.
[0061] (3) Constraints:
[0062] Forall i,ST[Task i ] = a i Indicates the start time of the task, corresponding to the starting point of the line segment.
[0063] Forall i,ED[Task i ]=b i Indicates the end time of the task, corresponding to the end point of the line segment.
[0064] Forall i,Machine[Task i ]∈[1…m] means that each task must be placed on one of the machines 1…m.
[0065] (4) Objective: Minimize the total cost. The function is as follows:
[0066]
[0067] Therefore, the above-mentioned original track allocation optimization problem is transformed into a scheduling problem, which can be solved using a variety of scheduling algorithms.
[0068] Step 4: There are many algorithms that can be used to solve the mathematical model. The following uses a heuristic based on linear relaxation as an example:
[0069] Step 4.1: Solve the linear relaxation of the established mathematical model and obtain the lower bound of the problem.
[0070] Step 4.2: Fix the variables whose values are 0 in the optimal solution of linear relaxation to 0, and re-solve the model. If the target value of the solution obtained is equal to the optimal value of linear relaxation, return the current solution and end the solution.
[0071] Step 4.3: Each time during the iteration, some of the fixed variables are unfixed and the original model is re-solved. If the solution obtained is feasible, the current solution is returned, otherwise the iteration continues.
[0072] Step 5: Output the solution results.
[0073] Step 5.1: If there is a feasible solution, convert the solution of the mathematical model into an allocation plan and output it;
[0074] Step 5.2: If there is no feasible solution, output infeasible.
[0075] Through the above steps, the track allocation problem can be transformed into a mathematical model of the scheduling problem, and the optimal track allocation solution can be obtained through the solution algorithm to meet the design needs and optimization requirements.
[0076] The present invention uses 4 actual industrial circuits as input to calculate the track allocation scheme, and compares the results with those directly solved by the best commercial solver Gurobi, see Table 1.
[0077] Table 1
[0078]
[0079] As shown in Table 1, the solution speed of the method of the present invention on the large-scale instance input_1_17 is 20 times that of Gurobi, and the solution is better than Gurobi, and the wiring cost is reduced by 0.10247%.
[0080] Although the present invention has been disclosed as above by way of embodiments, it is not intended to limit the present invention. Appropriate modifications or equivalent substitutions of the technical solutions of the present invention made by ordinary technicians in the field should all be included in the protection scope of the present invention. The protection scope of the present invention shall be based on what is defined in the claims.
Claims
1. A track allocation method for integrated circuit wiring, characterized in that: The following steps are involved: 1) Reading the circuit information after the overall layout of the integrated circuit, including the line segments to be allocated, the positions of obstacles and the positions of pins; 2) Define the wiring cost function, take the wire length cost, line segment overlap cost, pin overlap cost and obstacle overlap cost into the wiring cost in track allocation, and perform weighted summation; 3) Based on the wiring cost function, the track allocation optimization problem is modeled as a mathematical model of the scheduling problem; 4) Solve the scheduling problem of the mathematical model; 5) If there is a feasible solution, convert the solution into a track allocation plan.
2. The track allocation method according to claim 1, characterized in that: In step 1), the line segment to be allocated is represented by the coordinates of the starting point and the end point on the wiring area, and the obstacle position and the pin position are represented by the coordinates on the wiring area.
3. The track allocation method according to claim 1, characterized in that: The rule for track allocation in step 2) is that each routing area has only one track in the X-axis or Y-axis direction for routing.
4. The track allocation method according to claim 1, characterized in that: In step 2), the line length cost is the cost generated by the line length from the line segment to the connected pin in the track allocation; the line segment overlap cost is the cost generated by the overlap between two line segments allocated to the same track; the pin overlap cost is the cost generated by the overlap between a line segment and pins of different wire nets; the obstacle overlap cost is the cost generated by the partial or complete overlap between a line segment allocated to the track and one or more obstacles.
5. The track allocation method according to claim 1, characterized in that: The step of modeling the track allocation optimization problem as a mathematical model of the scheduling problem in step 3) includes: Represent the line segment as a task variable with fixed start and end time, represent the track as a machine variable in the scheduling problem, and transform the track allocation decision into a machine allocation decision variable in the scheduling problem; The wiring cost function is set as the objective function, with the goal of minimizing the total cost; The relationship between the machine allocation decision variables and the routing cost function is taken as a constraint; A mathematical model of the scheduling problem is constructed based on task variables, machine variables, machine allocation decision variables, objective functions and constraints.
6. The track allocation method according to claim 1, characterized in that: The steps of solving the scheduling problem of the mathematical model in step 4) include: The mathematical model is input into the scheduling problem solver and the scheduling algorithm is used to solve the problem within the set time. If the current solution is feasible, the current solution is returned; otherwise, the iterative solution is continued.
7. A track distribution system for integrated circuit wiring, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method according to any one of claims 1 to 6 when executing the computer program.
8. A computer-readable storage medium, characterized in that: A computer program is stored, and when the computer program is executed, the steps of the method according to any one of claims 1 to 6 are implemented.
Citation Information
Cited By
Track allocation method based on GPU (Graphics Processing Unit) parallel acceleration
CN119917268A
