US20200311224A1 - Multi-instantiation time budgeting for integrated circuit design and manufacturing - Google Patents
Multi-instantiation time budgeting for integrated circuit design and manufacturing Download PDFInfo
- Publication number
- US20200311224A1 US20200311224A1 US16/370,008 US201916370008A US2020311224A1 US 20200311224 A1 US20200311224 A1 US 20200311224A1 US 201916370008 A US201916370008 A US 201916370008A US 2020311224 A1 US2020311224 A1 US 2020311224A1
- Authority
- US
- United States
- Prior art keywords
- blocks
- design
- delay values
- block
- optimization
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000013461 design Methods 0.000 title claims abstract description 186
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 23
- 238000005457 optimization Methods 0.000 claims abstract description 47
- 238000000034 method Methods 0.000 claims abstract description 46
- 238000003860 storage Methods 0.000 claims description 7
- 230000004044 response Effects 0.000 claims 1
- 238000002620 method output Methods 0.000 abstract 1
- 230000001934 delay Effects 0.000 description 45
- 238000004422 calculation algorithm Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 7
- 238000007796 conventional method Methods 0.000 description 6
- 238000012938 design process Methods 0.000 description 6
- 238000005192 partition Methods 0.000 description 4
- 230000001902 propagating effect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000009826 distribution Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 239000000758 substrate Substances 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000013515 script Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G06F17/5081—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/39—Circuit design at the physical level
- G06F30/392—Floor-planning or layout, e.g. partitioning or placement
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/901—Indexing; Data structures therefor; Storage structures
- G06F16/9024—Graphs; Linked lists
-
- G06F17/5072—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/32—Circuit design at the digital level
- G06F30/33—Design verification, e.g. functional simulation or model checking
- G06F30/3308—Design verification, e.g. functional simulation or model checking using simulation
- G06F30/3312—Timing analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/39—Circuit design at the physical level
- G06F30/398—Design verification or optimisation, e.g. using design rule check [DRC], layout versus schematics [LVS] or finite element methods [FEM]
Definitions
- the exemplary embodiments generally relate to multi-instantiation time budgeting for components of a semiconductor integrated circuit (IC) in the design and manufacturing of the IC.
- IC semiconductor integrated circuit
- An IC includes a large number of electronic components that are manufactured on a substrate by forming layers of different materials and of different geometric shapes.
- the design of ICs continues to evolve, which increases the complexity and time required for design and manufacturing of ICs. Further, manufacturing technology continues to decrease the size of components, which allows a larger number of functions to be incorporated on a single IC.
- IC designers are capable of creating systems of components on an IC that are increasingly more complex, robust, and require a high-level design language in order to construct, verify, and test the IC.
- EDA Electronic design automation
- a design module is a functional partition of a full IC design that describes a function of the block according to the electronic components installed in the block.
- an EDA system may receive descriptions and parameters of the IC device defining interconnections of nodes and components on the IC, and includes multiple types of physical information of circuit components. Time budgeting derives input and output delay constraints along with path exceptions for the timing closure of the blocks or partitions, enabling the timing closure of all the blocks simultaneously.
- An integrated circuit designer may use the EDA system to create a physical integrated circuit design layout from a logical circuit design of the IC.
- the EDA system uses geometric shapes of different materials to design a layout of the various electrical components on a substrate.
- a designer Upon creating an initial integrated circuit layout, a designer then analyzes and optimizes the IC layout using the EDA system.
- One function of an EDA system is to determine whether the IC design meets timing requirements along signal paths in the design. Failure of a design to adequately meet timing requirements could result in the IC failing during use and/or not properly functioning for its intended purpose. Thus, for design purposes, it is important for designers to be able to accurately confirm that an IC design will satisfy its timing requirements.
- Time budgets are defined for each data path at each port of a block on the IC.
- a time budget derives a timing constraint of a block from the timing constraint of the full IC design.
- Time budgeting is a necessary step for large IC design, which uses algorithms to distribute timing period requirements to each input port and output port of each block of the IC.
- a first example of prior methods for performing time budgeting was to convert the full chip design into a directed acyclic graph (DAG), and use algorithms to determine a timing specification for the input and output ports of each block in the IC design.
- DAG directed acyclic graph
- the timing specifications of the blocks are different from each other, and thus require the implementation of each timing specification for each block to be different from each other.
- a second example of prior methods for performing time budgeting was to specify the time budget in a bottom-up method.
- the designer assigned a timing specification depending on their expert knowledge or known timing specifications of a particular design module, instead of using the full IC design's timing specification.
- the designer resolves these issues by manually re-allocating the block timing constraints.
- the exemplary embodiments provide a method of time budgeting of a multi-instantiation IC that is divided into multiple blocks, where at least two of the blocks include the same design module of components.
- the exemplary embodiments distribute an equal amount of time budgeting to blocks of the same design modules, while still achieving the full IC timing specifications. Therefore, time budgeting for multiple blocks of an IC with the same design module is performed together for the multiple blocks with the same design module, which reduces the engineer workloads and computational resource machine (computer) requirements, increases the speed and efficiency of the design process, and results in a final design of the IC chip where blocks with identical design modules have the same final time budgeted design.
- the same time budgeting of blocks also results in the same physical design of the components of the blocks when manufactured.
- the design process of the exemplary embodiments also changes the physical implementation of the blocks when the IC is manufactured.
- the time budgeting of each block was different from each other despite multiple blocks having the same design module.
- This prior method performed time budgeting for each block individually from each other. Because time budgeting affects the physical layout of a manufactured IC, when a different time budget is set for identical design modules and blocks, the physical implementation requires different optimization resulting in a different final design of the elements that connect to the input and output ports of the IC.
- the timing specification is calculated based on each design module, instead of the full IC timing specification.
- the timing specification only needs to be performed for one implementation for each design module, the implementation considers the design module only.
- the conventional technique calculates the time budget for each block separately, which results in different values of internal and external delays for different blocks. Therefore, conventional designers have to create physical implementations of each block individually, which could require tremendous engineering work.
- An alternative conventional technique a designer could pick one of the delay values to implement to identical design modules, which may result bad timing of the timing paths in the other block. This would then result in the designers needing to fix the timing paths iteratively with several rounds of design and fix.
- the workload is too high, the designer could use prior experiences to set the delay values, which could also cause bad designs, and/or burdensome iterations to fix adjacent blocks.
- the exemplary embodiments distribute equal time budgeting to blocks of the same design modules, while still achieving the full IC timing specifications. For example, the exemplary embodiments set identical time budgeting to all blocks of the same design, and then determine a time budgeting for the set blocks of the same design module together, instead of individually as in the conventional techniques.
- Time budgeting methods distribute timing delays relative to the different input and output ports of each block.
- IC design progresses and becomes more complex in current electronic devices, there could be several blocks in the full IC that have the same design to each other, i.e., blocks of the same design module.
- these blocks have separate time budgeting distribution according to different algorithms, and engineers need to optimize these blocks one-by-one.
- the exemplary embodiments improve the conventional methods by determining equal time budgeting to blocks of the same design modules, while still achieving the full IC timing specifications.
- the exemplary embodiments provide improved time budgeting by performing time budgeting for multiple blocks with the same design module together, instead of individually, such that the time budgeting for the multiple blocks having the same design module are determined to be identical to each other.
- This improved time budgeting reduces the engineer and computational resource (computer) requirements for physical designs, increases the speed and efficiency of the design process, and results in a final design of the IC chip where blocks with identical design modules have the same final time budgeted design (and manufactured layout).
- FIG. 1 illustrates a full IC having four blocks where two of the blocks have the same design module according to an exemplary embodiment.
- FIG. 2 illustrates an example of a block of the IC with internal and external delays of ports of the IC according to an exemplary embodiment.
- FIG. 3 illustrates internal and external delays of ports of each block of the full IC of FIG. 1 according to an exemplary embodiment.
- FIG. 4 illustrates timing relationships along timing paths between the blocks of the full IC according to an exemplary embodiment.
- FIG. 5 illustrates a flowchart of exemplary processes for multi-instantiation time budgeting according to an exemplary embodiment.
- FIG. 6 illustrates a block diagram of an exemplary time budgeting device for multi-instantiation time budgeting of an IC according to an exemplary embodiment.
- FIG. 1 is a block diagram illustrating a full integrated circuit (IC) 100 that is partitioned into, for example, four functional blocks 101 of electronic components (not shown). The blocks 101 partitioning the IC 100 into different regions perform functions according to the electronic components within each block 101 . Each block 101 is a physical instantiation of a design module 102 in the full IC 100 .
- IC integrated circuit
- the exemplary embodiments are directed to budgeting of clock signal timing of IC design for time budgeting of the blocks 101 of the IC 100 in order to satisfy a predetermined time budget.
- Time budgeting divides the time for a data signal to propagate along the data path extending between different blocks 101 of the IC 100 .
- Clock budget timing may be based on timing models for the different blocks 101 and/or the full IC 100 .
- the exemplary embodiments provide improved time budgeting by performing time budgeting for multiple blocks 101 with the same design module 102 together, instead of individually, such that the time budgeting for the multiple blocks 101 having the same design module 102 are determined to be identical to each other.
- this improved time budgeting reduces computational requirements of the EDA, increases the speed and efficiency of the design process, and results in a final design of the IC chip where blocks with identical design modules have the same final time budgeted design (and manufactured layout).
- Time budgeting is a part of the block implementation in the design of the IC 100 that allows for generating of timing and physical constraints for each partition (e.g., design module, block, etc.).
- IC 100 may be depicted in multiple different manners, such as a flattened chip or a tree hierarchy of different levels of circuits.
- FIG. 1 shows the IC 100 as a flattened chip for illustrative purposes.
- FIG. 1 is a block diagram illustrating the IC 100 that is partitioned into the four functional blocks 101 of electronic components.
- Each block 101 is a physical instantiation of a design module 102 in the full IC 100 .
- the design module 102 is a functional partition of the IC 100 .
- the design module defines the function behavior of the electronic components within the block 101 .
- the design module 102 may have one or multiple physical realizations in the IC. In other words, the full IC 100 may have multiple design modules 102 that are identical.
- Design Module A defines the functional behavior of Block 1 and Block 3 , which is different than the functional behavior defined by Design Module B and Design Module C.
- Design Module B and Design Module C respectively define Block 2 and Block 4 .
- Each block 101 is defined by one or more ports 103 (P 11 -P 19 , P 21 -P 29 , P 210 , P 211 , P 31-39 , and P 41 -P 46 ), as shown in FIGS. 2 and 3 .
- the ports 103 are located on a first side as input ports to receive incoming data signals to the block 101 , and on a second side as output ports where the outgoing data signals are transmitted from each block 101 .
- the ports 103 have connected elements 107 that are initially located at specified positions relative to the ports 3 of each block 101 .
- the position of one or more of the connected elements 107 may be moved relative to a respective one of the ports 103 depending on a change in the time budgeting.
- the location of each port 103 and connected element 107 in a final design of the IC 100 identifies the physical location of each port 103 and connected element 107 on the IC 100 when manufactured using an integrated circuit manufacturing and/or fabrication tool(s).
- a total clock delay of the IC 100 at a top level of hierarchy is divided between the blocks 101 in parallel with the data time budgeting of the data signal paths.
- the total clock delay of the IC 100 is specified as either an external clock delay 104 or an internal clock delay 105 of each block 101 .
- external delays 104 (e 11 , e 1824 , e 2833 , e 3843 , and e 45 ) are set for delays outside of each block 101
- internal delays 105 (d 11 -d 19 , d 21 -d 25 , d 210 , d 211 , d 31 -d 35 , d 38 , d 39 , d 41 -d 43 , and d 46 ) are set for delays for components within each of the blocks 101 .
- external delays 104 are the segments outside of all blocks 103 , which are values that are given from the “top” design.
- Internal delays 105 are the segments “inside” each block 101 , and are values that are optimized by the optimization of the time budgeting.
- a timing path 106 connects between different blocks 101 of the IC 100 .
- the timing paths 106 indicate the data signals that propagate between the ports of the different blocks 101 .
- the timing paths 106 may be configured to connect between varieties of different blocks 101 depending on the configuration of electronic components on the IC 100 .
- a first timing path 106 extending from Block 1 to Block 2 to Block 3 indicates data signals propagating from within Block 1 , accounting for internal delay d 18 , port P 18 , external delay e 1824 , port P 24 internal delay d 24 , port P 28 , external delay e 2833 , port P 33 , and internal delay d 33 .
- a second timing path 106 extends from Block 3 to Block 4 , which indicates data signals propagating from within Block 3 , accounting for internal delay d 38 , port P 38 , e 3843 , port P 43 , internal delay d 43 , port P 45 , and e 45 .
- a clock cycle of the IC 100 must be set to be equal to or larger than all of the timing paths 106 in order for the IC 100 to function properly, where the data signals have sufficient time to propagate through the IC 100 .
- a clock cycle for each block 101 must be set equal to or larger than the time necessary for a data signal to propagate along a timing path 106 extending through the respective block 101 .
- FIG. 5 illustrates a flowchart explaining exemplary steps of a method of time budgeting IC 100 .
- the method may include fewer or additional steps than the exemplary embodiments set forth herein.
- step S 200 a graph data structure of an IC design for IC 100 is acquired based on an initial physical database of the electronic components of the IC 100 .
- FIG. 3 illustrates an example of the graph data structure.
- the graph data structure divides electronic components of the IC design into the plurality of blocks 101 .
- Each of the plurality of blocks 101 corresponding to one of the design modules 102 .
- At least two blocks 101 (e.g., Block 1 and Block 3 ) correspond to an identical design module 102 .
- the graph data structure indicates an initial layout of ports 103 (P 11 -P 19 , P 21 -P 29 , P 210 , P 211 , P 31-39 , and P 41 -P 46 ) of each block 101 (Block 1 , Block 2 , Block 3 , and Block 4 ).
- the graph data structure also indicates external delays 104 and internal delays 105 for each port 103 .
- step S 210 values of the internal delay 105 and values of the external delay 104 of the plurality of ports 103 of each design module 103 are acquired and initially set corresponding to the plurality of blocks 101 .
- the values of the internal delay 105 and values of the external delay 104 are time delays that are caused by a data signal propagating between ports 103 .
- step S 210 through step S 230 optimization of the values of the internal delay 105 and the values of the external delay 104 is performed to satisfy a predetermined clock cycle of the IC 100 .
- Optimization refers to a quantitative computer algorithm or method that is applied to find a feasible solution to time budgeting.
- the algorithms may be linear optimization which may also be referred to as linear programming. Since the relationships between the internal delays 105 and the externals delays 104 , and the clock period of the clock cycle is a linear relationship, it is practical to use linear optimization or linear programming to calculate a solution for the variables.
- Time budgeting algorithms distribute timing requirements for the external delays 104 and internal delays 105 relative to each port 103 of the IC 100 .
- the timing requirement could be in various formats, such as one clock period, half of a clock period, or multiple clock periods.
- these blocks 101 that have the same design module 102 are set as equivalent to each other such that when optimization of the external delays 104 and the internal delays 105 is performed in step S 230 , as discussed below, the calculated time budgeting of the internal delays 105 are identical for each block 101 having the same design module 102 in the IC 100 .
- the method implements the multi-instantiation of these blocks 101 having the same design 102 during the optimization to reduce the computational requirements of the optimization of the full IC 100 and maintains the time budgeting as identical for these blocks 101 having the same design module 102 , as an improvement over the conventional techniques.
- Blocks 1 - 4 of FIG. 1 would have been assigned separate time budgeting distribution according to different algorithms, and engineers were required to optimize Blocks 1 - 4 one-by-one.
- the exemplary embodiments distribute an equal amount of time budgeting to blocks 101 of the same design modules 102 (i.e., Block 1 and Block 3 ), while still achieving the clock cycle requirements of the full IC 100 . Therefore, there is a need to only optimize identical blocks (Block 1 and Block 3 ) once, instead of separately as inefficiently performed in the conventional art.
- step S 230 optimization of the values of the internal delay 105 and the values of the external delay 104 is performed to determine a solution that satisfies the predetermined clock cycle of the IC 100 .
- the optimization is performed on each of the blocks 101 having a different design module 102 . Meaning, in the example shown in FIGS. 3 and 4 , the optimization is performed on Block 1 and Block 3 together because both Block 1 and Block 3 have the same design module 102 (Design Module A), as shown in FIG. 1 .
- the optimization is also performed on each of Block 2 and Block 4 in addition to the optimization performed on the combination of Block 1 and Block 4 .
- the optimization modifies a physical location of at least one of the connected elements 107 of ports 103 in order to satisfy the acquired clock cycle requirements.
- step S 220 requires setting the equivalence between the parameters of the blocks 101 of the same design module 102 . Because Block 1 and Block 3 each have the same design module 102 (Design Module A), the following equivalence requirements are set to represent the relationship between blocks of the same design module 102 :
- step S 240 a solution for the time budget of all of the internal delays 105 and the external delays 104 is obtained as a result of the optimization performed on the IC 100 .
- the solution at least indicates values of each d ij value and e ij value that result in a time budget equal to or less than the clock period requirement along each timing path 106 .
- the values of the internal delays 105 for Block 1 and Block 3 are identical due to the multi-instantiation of Block 1 and Block 3 , which have the same design module 102 , by setting the parameters of Block 1 and Block 3 as identical to each other and performing the optimization of Block 1 and Block 3 together.
- step S 250 a Quality of Results (QoR) determination is performed.
- QoR Quality of Results
- step S 200 When the QoR do not satisfy the predetermined clock cycle requirements of the IC 100 , then the method returns to step S 200 and iteratively performs steps S 200 to S 250 until the QoR in step S 250 satisfies all design and manufacturing requirements of the IC 100 .
- the parameters of all Blocks 1 - 4 are set based on the solution obtained in step S 240 . Then the graph data structure acquired in step S 200 will be acquired based on the newly set parameters of the IC 100 .
- Step S 260 the method continues to Step S 260 .
- step S 260 a final design of the IC 100 is generated.
- the final design indicates all parameters of the blocks 101 , including the internal delays 105 , the external delays 104 , the timing paths 106 , and the physical location of the connected elements 107 of the ports 103 for the IC 100 .
- step S 260 the final design is output for manufacturing of the IC 100 based on the final design.
- the outputting may be performed through a variety of methods, such as transmitting to a manufacturing tool, displaying on a display screen, notifying to a user, etc.
- step S 270 the IC 100 is manufactured according to the outputted final design.
- the values of the internal delays 105 , the values of the external delays 104 , the timing paths 106 , and the physical location of the connected elements 107 of the ports 103 of the at least two identical blocks (Block 1 and Block 3 ) are identical.
- the blocks with identical design modules of the conventional art were physical different when manufactured due to performing time budgeting individually on all of the blocks of an IC.
- Steps S 200 through step S 270 are merely exemplary steps of the exemplary embodiments and may include additional steps or sub-steps.
- timing delay variables i.e. d ij , e ij
- these variables guide a physical implementation process.
- an initial version of a physical database is generated. Different EDA tools may be used to check different aspects of this physical database.
- the QoR are evaluated to determine whether the physical database satisfies all design and manufacturing requirements, including the clock cycle requirements.
- step S 200 With each iteration of step S 200 through step S 250 , the values of the external delays 104 and the values of the internal delays 105 are gradually improved to consider various physical factors, until the design converges to a physical database that satisfies all requirements.
- the time budgeting device 110 may be configured as one or more computers including one or more processors 120 , one or more storage devices 130 , and an input/output (I/O) interface 140 .
- the one or more processors 120 may be any type of programmed computational device, including central processing units (CPU), microprocessors, microcontrollers, networked computer systems, etc., or a specialized processor for performing integrated circuit design.
- the one or more storage devices 130 may be a computer readable storage medium that includes memory devices, storage media readable by a removable media drive, and/or a hard disk drive, such as random access memory (RAM), read-only memory (ROM), magnetic hard disks, optical storage discs, etc., for storing one or more software modules of instructions that control the processor 120 to perform various operations.
- RAM random access memory
- ROM read-only memory
- magnetic hard disks magnetic hard disks
- optical storage discs etc.
- the I/O interface 140 allows a user to input to and receive data from the processor 120 .
- the I/O interface 140 also allows control of the various operations performed by the processor 102 .
- the I/O interface 140 may comprise one or more input devices, such as a keyboard, a pointing device (e.g., a mouse, a track ball), a touch-sensitive display, microphone, etc.
- the I/O interface 140 may also comprise one or more output devices, such as a display (including a touch-sensitive display).
- the one or more processors 120 and the one or more storage devices 130 of the time budgeting device 110 perform at least the same functions as the above-mentioned method.
- the above description of the method also applies to the following exemplary embodiments of the time budgeting device 110 , which will also be discussed in detail below.
- the one or more processors 120 acquire a graph data structure of an IC design for IC 100 .
- FIG. 3 illustrates an example of the graph data structure that divides electronic components of the IC design into the plurality of blocks 101 .
- At least two blocks 101 e.g., Block 1 and Block 3
- the graph data structure indicates an initial layout of connected elements 107 of ports 103 of each block 101 , and external delays 104 and internal delays 105 for each port 103 .
- the one or more processors 120 acquire and initially set values of the internal delay 105 and values of the external delay 104 of the plurality of ports 103 .
- the values of the internal delay 105 and values of the external delay 104 are periods of time delay that are caused by a data signal propagating between ports 103 .
- the one or more processors 120 perform optimization of the values of the internal delay 105 and the values of the external delay 104 to determine whether the parameters, such as a predetermined clock cycle, of the IC 100 is satisfied.
- the blocks 101 that have the same design module 102 are set as equivalent to each other such that when optimization of the external delays 104 and the internal delays 105 is performed in step S 230 , as discussed herein, the time budgeting of the external delays 104 and the internal delays 105 are identical for each block 101 having the same design module 102 in the IC 100 .
- the method implements the multi-instantiation of these blocks 101 having the same design 102 during the optimization to reduce the computational requirements of the optimization of the full IC 100 and maintains the time budgeting as identical for these blocks 101 having the same design module 102 , as an improvement over the conventional techniques.
- the one or more processors 120 perform optimization of the parameters, such as values of the internal delay 105 and the values of the external delay 104 , to determine a solution that satisfies the requirements of the IC 100 .
- the optimization is performed on Block 1 and Block 3 together because both Block 1 and Block 3 have the same design module 102 (Design Module A), as shown in FIG. 1 .
- the optimization is also performed on each of Block 2 and Block 4 in addition to the optimization performed on the combination of Block 1 and Block 4 .
- timing paths 106 in FIG. 4 assume for this example that the timing paths 106 are all under one clock cycle.
- the following relationships for the internal and external delays are set as requirements:
- the equivalence between the parameters of the blocks 101 of the same design module 102 is set because Block 1 and Block 3 each have the same design module 102 (Design Module A).
- the following equivalence requirements are set to represent the relationship between blocks of the same design module 102 :
- the one or more processors 120 obtain a solution for the time budget of all of the internal delays 105 and the external delays 104 as a result of the optimization performed on the IC 100 .
- the solution at least indicates values of each d u value and e 1 value that result in a time budget equal to or less than the requirements along each timing path 106 .
- the values of the external delays 104 and the internal delays 105 for Block 1 and Block 3 are identical due to the multi-instantiation of Block 1 and Block 3 , which have the same design module 102 , by setting the parameters of Block 1 and Block 3 as identical to each other and performing the optimization of Block 1 and Block 3 together in combination.
- the one or more processors 120 determine QoR of the solution that indicates whether the parameters, such as the values of the internal delay 105 and the values of the external delay 104 , along each timing path 106 satisfy predetermined clock cycle requirements for the IC 100 .
- the one or more processors 120 iteratively re-perform acquiring the graph data structure, setting the parameters of Block 1 equal to Block 3 , and performing optimization, obtaining a solution based on the optimization, as discussed in detail above, until the QoR satisfies the predetermined requirements of the IC 100 .
- the parameters of all Blocks 1 - 4 Prior to acquiring a new graph data structure at the start of each iteration, the parameters of all Blocks 1 - 4 are set based on the obtained solution of the prior iteration. Then the graph data structure acquired will be acquired based on the newly set parameters of the IC 100 .
- the one or more processors 120 When the QoR satisfy the predetermined requirements of the IC 100 , the one or more processors 120 generate a final design of the IC 100 .
- the final design indicates all parameters of the blocks 101 , including the internal delays 105 , the external delays 104 , the timing paths 106 , and the physical location of the connected elements 107 of the ports 103 for the IC 100 .
- the one or more processors 120 output the final design for manufacturing of the IC 100 based on the final design.
- the outputting may be performed through a variety of manners, such as transmitting to a manufacturing tool, displaying on the I/O interface 140 , etc.
- the one or more processors 120 may control the manufacturing of the IC 100 according to the outputted final design.
- the values of the internal delays 105 , the values of the external delays 104 , the timing paths 106 , and the physical location of the connected elements 107 of the ports 103 of the at least two identical blocks (Block 1 and Block 3 ) are identical.
- the blocks with identical design modules of the conventional art were physical different when manufactured due to performing time budgeting individually on all of the blocks of an IC.
- the above-mentioned embodiments provide blocks of an IC with the same design module and perform time budgeting for the identical blocks together, which reduces computational requirements of the EDA, increases the speed and efficiency of the design process, and results in a final design of the IC chip where blocks with identical design modules have the same final time budgeted design.
- the same time budgeting of identical blocks also results in the same physical design of the components of the blocks when manufactured.
- the design process of the exemplary embodiments also changes the physical design of the blocks when the IC is manufactured to be identical, which improves the efficiency of the manufacturing process.
- a time budget for d 18 may be “Clock period ⁇ e 1824 ⁇ d 24 ⁇ e 2833 ⁇ d 33 ” , i.e., with an output delay constraint of “e 1824 +d 24 +e 2833 +d 33 .” While a time budget for d 38 , the budget is “Clock period ⁇ e 3843 ⁇ d 43 ⁇ e 45 ”, which is the output delay constraint of “e 3843 +d 43 +e 45 .” Meaning, using the conventional process the values of d 18 and d 38 are different. On the contrary, the method and device of the exemplary embodiments performs time budgeting of Block 1 and Block 3 together, which results in the values of d 18 and d 38 being identical.
- the block implementation moves the physical location of the connected elements of port P 38 of Block 3 , so that the time budget can be met with a smaller value.
- the exemplary method and device also changes the physical location of the connected elements of port P 18 to be identical. This results in a more efficient design and optimization process that reduce the computational requirements for the optimization, and also results in manufacturing ICs where at two blocks 101 with identical design modules 102 are physically identical, contrary to the conventional art.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- Databases & Information Systems (AREA)
- Architecture (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Design And Manufacture Of Integrated Circuits (AREA)
Abstract
Description
- The exemplary embodiments generally relate to multi-instantiation time budgeting for components of a semiconductor integrated circuit (IC) in the design and manufacturing of the IC.
- An IC includes a large number of electronic components that are manufactured on a substrate by forming layers of different materials and of different geometric shapes. The design of ICs continues to evolve, which increases the complexity and time required for design and manufacturing of ICs. Further, manufacturing technology continues to decrease the size of components, which allows a larger number of functions to be incorporated on a single IC. IC designers are capable of creating systems of components on an IC that are increasingly more complex, robust, and require a high-level design language in order to construct, verify, and test the IC.
- Electronic design automation (EDA) systems are software used to aid a designer in design of an IC. For complex ICs, the physical layout and design of an IC is divided into multiple functional blocks with a plurality of levels of hierarchy, and each of the blocks may correspond to the same design module or different design modules. A design module is a functional partition of a full IC design that describes a function of the block according to the electronic components installed in the block. For example, an EDA system may receive descriptions and parameters of the IC device defining interconnections of nodes and components on the IC, and includes multiple types of physical information of circuit components. Time budgeting derives input and output delay constraints along with path exceptions for the timing closure of the blocks or partitions, enabling the timing closure of all the blocks simultaneously.
- An integrated circuit designer may use the EDA system to create a physical integrated circuit design layout from a logical circuit design of the IC. The EDA system uses geometric shapes of different materials to design a layout of the various electrical components on a substrate. Upon creating an initial integrated circuit layout, a designer then analyzes and optimizes the IC layout using the EDA system. One function of an EDA system is to determine whether the IC design meets timing requirements along signal paths in the design. Failure of a design to adequately meet timing requirements could result in the IC failing during use and/or not properly functioning for its intended purpose. Thus, for design purposes, it is important for designers to be able to accurately confirm that an IC design will satisfy its timing requirements.
- Time budgets are defined for each data path at each port of a block on the IC. A time budget derives a timing constraint of a block from the timing constraint of the full IC design. Time budgeting is a necessary step for large IC design, which uses algorithms to distribute timing period requirements to each input port and output port of each block of the IC.
- There are prior methods for performing time budgeting. A first example of prior methods for performing time budgeting was to convert the full chip design into a directed acyclic graph (DAG), and use algorithms to determine a timing specification for the input and output ports of each block in the IC design. The timing specifications of the blocks are different from each other, and thus require the implementation of each timing specification for each block to be different from each other.
- A second example of prior methods for performing time budgeting was to specify the time budget in a bottom-up method. The designer assigned a timing specification depending on their expert knowledge or known timing specifications of a particular design module, instead of using the full IC design's timing specification. When the physical layout of the blocks is unable to meet their block timing constraints in a full IC, the designer resolves these issues by manually re-allocating the block timing constraints.
- However, as the complexity of designing integrated circuits continues to increase, there is the need to improve the efficiency and accuracy of the time budgeting of the design and manufacturing of integrated circuits.
- The exemplary embodiments provide a method of time budgeting of a multi-instantiation IC that is divided into multiple blocks, where at least two of the blocks include the same design module of components. The exemplary embodiments distribute an equal amount of time budgeting to blocks of the same design modules, while still achieving the full IC timing specifications. Therefore, time budgeting for multiple blocks of an IC with the same design module is performed together for the multiple blocks with the same design module, which reduces the engineer workloads and computational resource machine (computer) requirements, increases the speed and efficiency of the design process, and results in a final design of the IC chip where blocks with identical design modules have the same final time budgeted design. The same time budgeting of blocks also results in the same physical design of the components of the blocks when manufactured. Thus, the design process of the exemplary embodiments also changes the physical implementation of the blocks when the IC is manufactured.
- In the first example of the prior methods, the time budgeting of each block was different from each other despite multiple blocks having the same design module. This prior method performed time budgeting for each block individually from each other. Because time budgeting affects the physical layout of a manufactured IC, when a different time budget is set for identical design modules and blocks, the physical implementation requires different optimization resulting in a different final design of the elements that connect to the input and output ports of the IC.
- According to this prior method, a design module that appears “n” times in the full IC design will need “n” different physical implementations. This time-consuming process may have been previously permitted for smaller and less complex IC, when there was no extensive use of the same design module. However, this time-consuming approach is now burdensome and a waste of engineering and computational resources for current and future ICs having a design module that could be reused from several or hundreds of times in a single IC in modern and future designs.
- In the second example of prior methods, the timing specification is calculated based on each design module, instead of the full IC timing specification. Although the timing specification only needs to be performed for one implementation for each design module, the implementation considers the design module only.
- As a similar example, the conventional technique calculates the time budget for each block separately, which results in different values of internal and external delays for different blocks. Therefore, conventional designers have to create physical implementations of each block individually, which could require tremendous engineering work. An alternative conventional technique, a designer could pick one of the delay values to implement to identical design modules, which may result bad timing of the timing paths in the other block. This would then result in the designers needing to fix the timing paths iteratively with several rounds of design and fix. As another example, if the workload is too high, the designer could use prior experiences to set the delay values, which could also cause bad designs, and/or burdensome iterations to fix adjacent blocks.
- However, these prior methods required an excessive amount of time and processing resources, especially as the number of components and features increases in modern electronic devices.
- In order to solve these problems in the prior methods, the exemplary embodiments distribute equal time budgeting to blocks of the same design modules, while still achieving the full IC timing specifications. For example, the exemplary embodiments set identical time budgeting to all blocks of the same design, and then determine a time budgeting for the set blocks of the same design module together, instead of individually as in the conventional techniques.
- Time budgeting methods distribute timing delays relative to the different input and output ports of each block. As IC design progresses and becomes more complex in current electronic devices, there could be several blocks in the full IC that have the same design to each other, i.e., blocks of the same design module. Under the conventional budgeting methods, these blocks have separate time budgeting distribution according to different algorithms, and engineers need to optimize these blocks one-by-one. The exemplary embodiments improve the conventional methods by determining equal time budgeting to blocks of the same design modules, while still achieving the full IC timing specifications.
- The exemplary embodiments provide improved time budgeting by performing time budgeting for multiple blocks with the same design module together, instead of individually, such that the time budgeting for the multiple blocks having the same design module are determined to be identical to each other. This improved time budgeting reduces the engineer and computational resource (computer) requirements for physical designs, increases the speed and efficiency of the design process, and results in a final design of the IC chip where blocks with identical design modules have the same final time budgeted design (and manufactured layout).
-
FIG. 1 illustrates a full IC having four blocks where two of the blocks have the same design module according to an exemplary embodiment. -
FIG. 2 illustrates an example of a block of the IC with internal and external delays of ports of the IC according to an exemplary embodiment. -
FIG. 3 illustrates internal and external delays of ports of each block of the full IC ofFIG. 1 according to an exemplary embodiment. -
FIG. 4 illustrates timing relationships along timing paths between the blocks of the full IC according to an exemplary embodiment. -
FIG. 5 illustrates a flowchart of exemplary processes for multi-instantiation time budgeting according to an exemplary embodiment. -
FIG. 6 illustrates a block diagram of an exemplary time budgeting device for multi-instantiation time budgeting of an IC according to an exemplary embodiment. - The exemplary embodiments relate to time budgeting of clock signal timing of IC design.
FIG. 1 is a block diagram illustrating a full integrated circuit (IC) 100 that is partitioned into, for example, fourfunctional blocks 101 of electronic components (not shown). Theblocks 101 partitioning theIC 100 into different regions perform functions according to the electronic components within eachblock 101. Eachblock 101 is a physical instantiation of adesign module 102 in thefull IC 100. - The exemplary embodiments are directed to budgeting of clock signal timing of IC design for time budgeting of the
blocks 101 of theIC 100 in order to satisfy a predetermined time budget. Time budgeting divides the time for a data signal to propagate along the data path extending betweendifferent blocks 101 of theIC 100. Clock budget timing may be based on timing models for thedifferent blocks 101 and/or thefull IC 100. The exemplary embodiments provide improved time budgeting by performing time budgeting formultiple blocks 101 with thesame design module 102 together, instead of individually, such that the time budgeting for themultiple blocks 101 having thesame design module 102 are determined to be identical to each other. As discussed in further detail herein, this improved time budgeting reduces computational requirements of the EDA, increases the speed and efficiency of the design process, and results in a final design of the IC chip where blocks with identical design modules have the same final time budgeted design (and manufactured layout). - Time budgeting is a part of the block implementation in the design of the
IC 100 that allows for generating of timing and physical constraints for each partition (e.g., design module, block, etc.). - The exemplary embodiments are used to design an integrated circuit, such as
IC 100.IC 100 may be depicted in multiple different manners, such as a flattened chip or a tree hierarchy of different levels of circuits. For example,FIG. 1 shows theIC 100 as a flattened chip for illustrative purposes. -
FIG. 1 is a block diagram illustrating theIC 100 that is partitioned into the fourfunctional blocks 101 of electronic components. Eachblock 101 is a physical instantiation of adesign module 102 in thefull IC 100. Thedesign module 102 is a functional partition of theIC 100. The design module defines the function behavior of the electronic components within theblock 101. Thedesign module 102 may have one or multiple physical realizations in the IC. In other words, thefull IC 100 may havemultiple design modules 102 that are identical. - For example, as illustrated in
FIG. 1 , two of the blocks 101 (Block 1 and Block 3) have the same design module (Design Module A). Design Module A defines the functional behavior ofBlock 1 andBlock 3, which is different than the functional behavior defined by Design Module B and Design Module C. Design Module B and Design Module C respectively defineBlock 2 andBlock 4. - Each
block 101 is defined by one or more ports 103 (P11-P19, P21-P29, P210, P211, P31-39, and P41-P46), as shown inFIGS. 2 and 3 . For example, theports 103 are located on a first side as input ports to receive incoming data signals to theblock 101, and on a second side as output ports where the outgoing data signals are transmitted from eachblock 101. Theports 103 have connectedelements 107 that are initially located at specified positions relative to theports 3 of eachblock 101. During optimization of the time budgeting, which is discussed in more detail below, the position of one or more of theconnected elements 107 may be moved relative to a respective one of theports 103 depending on a change in the time budgeting. The location of eachport 103 andconnected element 107 in a final design of theIC 100 identifies the physical location of eachport 103 andconnected element 107 on theIC 100 when manufactured using an integrated circuit manufacturing and/or fabrication tool(s). - For clock time budgeting, a total clock delay of the
IC 100 at a top level of hierarchy is divided between theblocks 101 in parallel with the data time budgeting of the data signal paths. The total clock delay of theIC 100 is specified as either anexternal clock delay 104 or aninternal clock delay 105 of eachblock 101. - As shown in
FIGS. 2-4 , external delays 104 (e11, e1824, e2833, e3843, and e45) are set for delays outside of eachblock 101, and internal delays 105 (d11-d19, d21-d25, d210, d211, d31-d35, d38, d39, d41-d43, and d46) are set for delays for components within each of theblocks 101. In other words,external delays 104 are the segments outside of allblocks 103, which are values that are given from the “top” design.Internal delays 105 are the segments “inside” eachblock 101, and are values that are optimized by the optimization of the time budgeting. - As shown in
FIG. 4 , atiming path 106 connects betweendifferent blocks 101 of theIC 100. The timingpaths 106 indicate the data signals that propagate between the ports of thedifferent blocks 101. The timingpaths 106 may be configured to connect between varieties ofdifferent blocks 101 depending on the configuration of electronic components on theIC 100. - For example, as shown in
FIG. 4 , afirst timing path 106 extending fromBlock 1 to Block 2 to Block 3 indicates data signals propagating from withinBlock 1, accounting for internal delay d18, port P18, external delay e1824, port P24 internal delay d24, port P28, external delay e2833, port P33, and internal delay d33. Further, asecond timing path 106 extends fromBlock 3 toBlock 4, which indicates data signals propagating from withinBlock 3, accounting for internal delay d38, port P38, e3843, port P43, internal delay d43, port P45, and e45. - A clock cycle of the
IC 100 must be set to be equal to or larger than all of the timingpaths 106 in order for theIC 100 to function properly, where the data signals have sufficient time to propagate through theIC 100. Similarly, a clock cycle for eachblock 101 must be set equal to or larger than the time necessary for a data signal to propagate along atiming path 106 extending through therespective block 101. Thus, it is important to optimize the time budgeting of theblocks 101 of theIC 100 in order to ensure proper functioning of theIC 100. - Method of Time Budgeting an Integrated Circuit (IC)
-
FIG. 5 illustrates a flowchart explaining exemplary steps of a method oftime budgeting IC 100. The method may include fewer or additional steps than the exemplary embodiments set forth herein. - In step S200, a graph data structure of an IC design for
IC 100 is acquired based on an initial physical database of the electronic components of theIC 100.FIG. 3 illustrates an example of the graph data structure. The graph data structure divides electronic components of the IC design into the plurality ofblocks 101. Each of the plurality ofblocks 101 corresponding to one of thedesign modules 102. At least two blocks 101 (e.g.,Block 1 and Block 3) correspond to anidentical design module 102. - The graph data structure indicates an initial layout of ports 103 (P11-P19, P21-P29, P210, P211, P31-39, and P41-P46) of each block 101 (
Block 1,Block 2,Block 3, and Block 4). The graph data structure also indicatesexternal delays 104 andinternal delays 105 for eachport 103. - In step S210, values of the
internal delay 105 and values of theexternal delay 104 of the plurality ofports 103 of eachdesign module 103 are acquired and initially set corresponding to the plurality ofblocks 101. The values of theinternal delay 105 and values of theexternal delay 104 are time delays that are caused by a data signal propagating betweenports 103. - In step S210 through step S230, optimization of the values of the
internal delay 105 and the values of theexternal delay 104 is performed to satisfy a predetermined clock cycle of theIC 100. Optimization refers to a quantitative computer algorithm or method that is applied to find a feasible solution to time budgeting. For example, the algorithms may be linear optimization which may also be referred to as linear programming. Since the relationships between theinternal delays 105 and the externals delays 104, and the clock period of the clock cycle is a linear relationship, it is practical to use linear optimization or linear programming to calculate a solution for the variables. - Time budgeting algorithms distribute timing requirements for the
external delays 104 andinternal delays 105 relative to eachport 103 of theIC 100. The timing requirement could be in various formats, such as one clock period, half of a clock period, or multiple clock periods. As integrated circuit technology progresses, there could beseveral blocks 101 in the full chip which have the same composition of electronic components to each other, i.e., blocks 101 of thesame design module 102. - In step 220, these
blocks 101 that have thesame design module 102 are set as equivalent to each other such that when optimization of theexternal delays 104 and theinternal delays 105 is performed in step S230, as discussed below, the calculated time budgeting of theinternal delays 105 are identical for eachblock 101 having thesame design module 102 in theIC 100. By setting theblocks 101 having thesame design module 102 as equivalent, the method implements the multi-instantiation of theseblocks 101 having thesame design 102 during the optimization to reduce the computational requirements of the optimization of thefull IC 100 and maintains the time budgeting as identical for theseblocks 101 having thesame design module 102, as an improvement over the conventional techniques. - Under the conventional time budgeting, Blocks 1-4 of
FIG. 1 would have been assigned separate time budgeting distribution according to different algorithms, and engineers were required to optimize Blocks 1-4 one-by-one. However, the exemplary embodiments distribute an equal amount of time budgeting toblocks 101 of the same design modules 102 (i.e.,Block 1 and Block 3), while still achieving the clock cycle requirements of thefull IC 100. Therefore, there is a need to only optimize identical blocks (Block 1 and Block 3) once, instead of separately as inefficiently performed in the conventional art. - In step S230, optimization of the values of the
internal delay 105 and the values of theexternal delay 104 is performed to determine a solution that satisfies the predetermined clock cycle of theIC 100. The optimization is performed on each of theblocks 101 having adifferent design module 102. Meaning, in the example shown inFIGS. 3 and 4 , the optimization is performed onBlock 1 andBlock 3 together because bothBlock 1 andBlock 3 have the same design module 102 (Design Module A), as shown inFIG. 1 . The optimization is also performed on each ofBlock 2 andBlock 4 in addition to the optimization performed on the combination ofBlock 1 andBlock 4. - The optimization modifies a physical location of at least one of the
connected elements 107 ofports 103 in order to satisfy the acquired clock cycle requirements. - In the optimization, all delay variables/
values 104/105 along eachtiming path 106 are summed together, and the sum must be equal to or less than the clock period of eachtiming path 106. Using the timingpaths 106 inFIG. 4 as examples, assume for this example that the timingpaths 106 are all under one clock cycle. The following relationships for the internal and external delays are set as requirements: -
e 11 +d 11≤Clock period -
d 19 +e 19≤Clock period -
d 18 +e 1824 +d 24 +e 2833 +d 33≤Clock period -
d 38 +e 3843 +d 43 +e 45≤Clock period - In addition, step S220 requires setting the equivalence between the parameters of the
blocks 101 of thesame design module 102. BecauseBlock 1 andBlock 3 each have the same design module 102 (Design Module A), the following equivalence requirements are set to represent the relationship between blocks of the same design module 102: -
d 11 =d 31 -
d 12 =d 32 -
d 13 =d 33 -
d 14 =d 34 -
d 15 =d 35 -
d 18 =d 38 -
d 19 =d 39 - In step S240, a solution for the time budget of all of the
internal delays 105 and theexternal delays 104 is obtained as a result of the optimization performed on theIC 100. The solution at least indicates values of each dij value and eij value that result in a time budget equal to or less than the clock period requirement along eachtiming path 106. - In the solution, the values of the
internal delays 105 forBlock 1 andBlock 3 are identical due to the multi-instantiation ofBlock 1 andBlock 3, which have thesame design module 102, by setting the parameters ofBlock 1 andBlock 3 as identical to each other and performing the optimization ofBlock 1 andBlock 3 together. - In step S250, a Quality of Results (QoR) determination is performed. After calculating a set of input and output timing delay variables, i.e. dij, eij etc., the variables to guide a physical implementation process. Based on information derived from the variables, with adjusted configurations, an initial version of a physical database is generated. Different EDA tools may be used to verify different aspects of the physical database, and quality of results (QoR) are evaluated to examine whether the physical database is satisfies all design and manufacturing criteria.
- When the QoR do not satisfy the predetermined clock cycle requirements of the
IC 100, then the method returns to step S200 and iteratively performs steps S200 to S250 until the QoR in step S250 satisfies all design and manufacturing requirements of theIC 100. When returning to step S200 and prior to acquiring a new graph data structure in step S200, the parameters of all Blocks 1-4 are set based on the solution obtained in step S240. Then the graph data structure acquired in step S200 will be acquired based on the newly set parameters of theIC 100. - When the QoR satisfy all design and manufacturing requirements of the
IC 100, then the method continues to Step S260. - In step S260, a final design of the
IC 100 is generated. The final design indicates all parameters of theblocks 101, including theinternal delays 105, theexternal delays 104, the timingpaths 106, and the physical location of theconnected elements 107 of theports 103 for theIC 100. By performing optimization ofBlock 1 andBlock 3 together and setting the parameters ofBlock 3 to be identical toBlock 1 during the optimization, it is ensured thatBlock 1 andBlock 3 will be physically identical when manufactured because the physical location of theconnected elements 107 of the ports is set as the same and bothBlock 1 andBlock 3 correspond to the same design module 102 (Design Module A). - In step S260, the final design is output for manufacturing of the
IC 100 based on the final design. The outputting may be performed through a variety of methods, such as transmitting to a manufacturing tool, displaying on a display screen, notifying to a user, etc. - In step S270, the
IC 100 is manufactured according to the outputted final design. In the manufacturedIC 100, the values of theinternal delays 105, the values of theexternal delays 104, the timingpaths 106, and the physical location of theconnected elements 107 of theports 103 of the at least two identical blocks (Block 1 and Block 3) are identical. In contrast, the blocks with identical design modules of the conventional art were physical different when manufactured due to performing time budgeting individually on all of the blocks of an IC. - Steps S200 through step S270 are merely exemplary steps of the exemplary embodiments and may include additional steps or sub-steps.
- For example, after determining a set of solutions to the timing delay variables, i.e. dij, eij, these variables guide a physical implementation process. Based on the variables, an initial version of a physical database is generated. Different EDA tools may be used to check different aspects of this physical database. Then, the QoR are evaluated to determine whether the physical database satisfies all design and manufacturing requirements, including the clock cycle requirements.
- With each iteration of step S200 through step S250, the values of the
external delays 104 and the values of theinternal delays 105 are gradually improved to consider various physical factors, until the design converges to a physical database that satisfies all requirements. - Time Budgeting Device 110
- As shown in
FIG. 6 , the time budgeting device 110 may be configured as one or more computers including one ormore processors 120, one ormore storage devices 130, and an input/output (I/O)interface 140. The one ormore processors 120 may be any type of programmed computational device, including central processing units (CPU), microprocessors, microcontrollers, networked computer systems, etc., or a specialized processor for performing integrated circuit design. The one ormore storage devices 130 may be a computer readable storage medium that includes memory devices, storage media readable by a removable media drive, and/or a hard disk drive, such as random access memory (RAM), read-only memory (ROM), magnetic hard disks, optical storage discs, etc., for storing one or more software modules of instructions that control theprocessor 120 to perform various operations. - The I/
O interface 140 allows a user to input to and receive data from theprocessor 120. The I/O interface 140 also allows control of the various operations performed by theprocessor 102. For example, the I/O interface 140 may comprise one or more input devices, such as a keyboard, a pointing device (e.g., a mouse, a track ball), a touch-sensitive display, microphone, etc. The I/O interface 140 may also comprise one or more output devices, such as a display (including a touch-sensitive display). - The one or
more processors 120 and the one ormore storage devices 130 of the time budgeting device 110 perform at least the same functions as the above-mentioned method. The above description of the method also applies to the following exemplary embodiments of the time budgeting device 110, which will also be discussed in detail below. - The one or
more processors 120 acquire a graph data structure of an IC design forIC 100.FIG. 3 illustrates an example of the graph data structure that divides electronic components of the IC design into the plurality ofblocks 101. At least two blocks 101 (e.g.,Block 1 and Block 3) correspond to anidentical design module 102. The graph data structure indicates an initial layout ofconnected elements 107 ofports 103 of eachblock 101, andexternal delays 104 andinternal delays 105 for eachport 103. - The one or
more processors 120 acquire and initially set values of theinternal delay 105 and values of theexternal delay 104 of the plurality ofports 103. The values of theinternal delay 105 and values of theexternal delay 104 are periods of time delay that are caused by a data signal propagating betweenports 103. - The one or
more processors 120 perform optimization of the values of theinternal delay 105 and the values of theexternal delay 104 to determine whether the parameters, such as a predetermined clock cycle, of theIC 100 is satisfied. - The
blocks 101 that have thesame design module 102 are set as equivalent to each other such that when optimization of theexternal delays 104 and theinternal delays 105 is performed in step S230, as discussed herein, the time budgeting of theexternal delays 104 and theinternal delays 105 are identical for eachblock 101 having thesame design module 102 in theIC 100. By setting theblocks 101 having thesame design module 102 as equivalent, the method implements the multi-instantiation of theseblocks 101 having thesame design 102 during the optimization to reduce the computational requirements of the optimization of thefull IC 100 and maintains the time budgeting as identical for theseblocks 101 having thesame design module 102, as an improvement over the conventional techniques. - The one or
more processors 120 perform optimization of the parameters, such as values of theinternal delay 105 and the values of theexternal delay 104, to determine a solution that satisfies the requirements of theIC 100. The optimization is performed onBlock 1 andBlock 3 together because bothBlock 1 andBlock 3 have the same design module 102 (Design Module A), as shown inFIG. 1 . The optimization is also performed on each ofBlock 2 andBlock 4 in addition to the optimization performed on the combination ofBlock 1 andBlock 4. - Using the timing
paths 106 inFIG. 4 as examples, assume for this example that the timingpaths 106 are all under one clock cycle. The following relationships for the internal and external delays are set as requirements: -
e 11 +d 11≤Clock period -
d 19 +e 19≤Clock period -
d 18 +e 1824 +d 24 +e 2833 +d 33≤Clock period -
d 38 +e 3843 +d 43 +e 45≤Clock period - In addition, the equivalence between the parameters of the
blocks 101 of thesame design module 102 is set becauseBlock 1 andBlock 3 each have the same design module 102 (Design Module A). The following equivalence requirements are set to represent the relationship between blocks of the same design module 102: -
d 11 =d 31 -
d 12 =d 32 -
d 13 =d 33 -
d 14 =d 34 -
d 15 =d 35 -
d 18 =d 38 -
d 19 =d 39 - The one or
more processors 120 obtain a solution for the time budget of all of theinternal delays 105 and theexternal delays 104 as a result of the optimization performed on theIC 100. The solution at least indicates values of each du value and e1 value that result in a time budget equal to or less than the requirements along eachtiming path 106. In the solution, the values of theexternal delays 104 and theinternal delays 105 forBlock 1 andBlock 3 are identical due to the multi-instantiation ofBlock 1 andBlock 3, which have thesame design module 102, by setting the parameters ofBlock 1 andBlock 3 as identical to each other and performing the optimization ofBlock 1 andBlock 3 together in combination. - The one or
more processors 120 determine QoR of the solution that indicates whether the parameters, such as the values of theinternal delay 105 and the values of theexternal delay 104, along eachtiming path 106 satisfy predetermined clock cycle requirements for theIC 100. - When the QoR do not satisfy the predetermined requirements of the
IC 100, then the one ormore processors 120 iteratively re-perform acquiring the graph data structure, setting the parameters ofBlock 1 equal toBlock 3, and performing optimization, obtaining a solution based on the optimization, as discussed in detail above, until the QoR satisfies the predetermined requirements of theIC 100. Prior to acquiring a new graph data structure at the start of each iteration, the parameters of all Blocks 1-4 are set based on the obtained solution of the prior iteration. Then the graph data structure acquired will be acquired based on the newly set parameters of theIC 100. - When the QoR satisfy the predetermined requirements of the
IC 100, the one ormore processors 120 generate a final design of theIC 100. The final design indicates all parameters of theblocks 101, including theinternal delays 105, theexternal delays 104, the timingpaths 106, and the physical location of theconnected elements 107 of theports 103 for theIC 100. By performing optimization ofBlock 1 andBlock 3 together and setting the parameters ofBlock 3 to be identical toBlock 1 during the optimization, it is ensured thatBlock 1 andBlock 3 will be physically identical when manufactured because the physical location of theconnected elements 107 of the ports is set as the same and bothBlock 1 andBlock 3 correspond to the same design module 102 (Design Module A). - The one or
more processors 120 output the final design for manufacturing of theIC 100 based on the final design. The outputting may be performed through a variety of manners, such as transmitting to a manufacturing tool, displaying on the I/O interface 140, etc. - The one or
more processors 120 may control the manufacturing of theIC 100 according to the outputted final design. In the manufacturedIC 100, the values of theinternal delays 105, the values of theexternal delays 104, the timingpaths 106, and the physical location of theconnected elements 107 of theports 103 of the at least two identical blocks (Block 1 and Block 3) are identical. In contrast, the blocks with identical design modules of the conventional art were physical different when manufactured due to performing time budgeting individually on all of the blocks of an IC. - The above-mentioned embodiments are examples of the methods and devices that achieve the improvements set forth herein. Although the above-mentioned embodiments are disclosed in relation to EDA software, these exemplary embodiments may be applied to other types of software, such as scripts.
- The above-mentioned embodiments provide blocks of an IC with the same design module and perform time budgeting for the identical blocks together, which reduces computational requirements of the EDA, increases the speed and efficiency of the design process, and results in a final design of the IC chip where blocks with identical design modules have the same final time budgeted design. The same time budgeting of identical blocks also results in the same physical design of the components of the blocks when manufactured. Thus, the design process of the exemplary embodiments also changes the physical design of the blocks when the IC is manufactured to be identical, which improves the efficiency of the manufacturing process.
- In conventional process, a time budget for d18 may be “Clock period−e1824−d24−e2833−d33 ”, i.e., with an output delay constraint of “e1824+d24+e2833+d33.” While a time budget for d38, the budget is “Clock period−e3843−d43−e45”, which is the output delay constraint of “e3843+d43+e45.” Meaning, using the conventional process the values of d18 and d38 are different. On the contrary, the method and device of the exemplary embodiments performs time budgeting of
Block 1 andBlock 3 together, which results in the values of d18 and d38 being identical. - Further, the using the above-mentioned example, when the exemplary embodiments calculate a time budget of “Clock period−e1824−d24−e2833−d33,” which is smaller than the budget of a conventional algorithm “Clock period−e3843−d43−e45,” the block implementation moves the physical location of the connected elements of port P38 of
Block 3, so that the time budget can be met with a smaller value. The exemplary method and device also changes the physical location of the connected elements of port P18 to be identical. This results in a more efficient design and optimization process that reduce the computational requirements for the optimization, and also results in manufacturing ICs where at twoblocks 101 withidentical design modules 102 are physically identical, contrary to the conventional art.
Claims (9)
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US16/370,008 US10810344B1 (en) | 2019-03-29 | 2019-03-29 | Multi-instantiation time budgeting for integrated circuit design and manufacturing |
CN202010235960.1A CN111539176B (en) | 2019-03-29 | 2020-03-30 | Multi-instance time budget for integrated circuit design and fabrication |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US16/370,008 US10810344B1 (en) | 2019-03-29 | 2019-03-29 | Multi-instantiation time budgeting for integrated circuit design and manufacturing |
Publications (2)
Publication Number | Publication Date |
---|---|
US20200311224A1 true US20200311224A1 (en) | 2020-10-01 |
US10810344B1 US10810344B1 (en) | 2020-10-20 |
Family
ID=71976869
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US16/370,008 Active US10810344B1 (en) | 2019-03-29 | 2019-03-29 | Multi-instantiation time budgeting for integrated circuit design and manufacturing |
Country Status (2)
Country | Link |
---|---|
US (1) | US10810344B1 (en) |
CN (1) | CN111539176B (en) |
Family Cites Families (29)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1539114A (en) * | 2001-06-08 | 2004-10-20 | �������Զ�������ƹ�˾ | Method for generating design constraints for modulates in hierarchical integrated circuit design system |
US7451412B2 (en) * | 2005-08-04 | 2008-11-11 | Synopsys, Inc. | Speeding up timing analysis by reusing delays computed for isomorphic subcircuits |
US7590964B1 (en) * | 2005-12-19 | 2009-09-15 | Tensilica, Inc. | Method and system for automatic generation of processor datapaths using instruction set architecture implementing means |
US8365113B1 (en) * | 2007-01-10 | 2013-01-29 | Cadence Design Systems, Inc. | Flow methodology for single pass parallel hierarchical timing closure of integrated circuit designs |
US8024686B2 (en) * | 2008-11-25 | 2011-09-20 | Synopsys, Inc. | Retiming of multirate system |
CN101846724B (en) * | 2009-03-26 | 2013-06-12 | 台湾积体电路制造股份有限公司 | Method and apparatus for diagnosing an integrated circuit |
CN102073007B (en) * | 2009-11-25 | 2013-06-26 | 台湾积体电路制造股份有限公司 | System and method for detecting soft-fails |
CN102169515B (en) * | 2010-02-26 | 2014-04-16 | 国际商业机器公司 | Estimation method and system of clock tree delay time in specified integrated circuit |
US8296704B1 (en) * | 2010-07-09 | 2012-10-23 | Altera Corporation | Method and apparatus for simultaneous switching noise optimization |
US8341573B2 (en) * | 2010-10-15 | 2012-12-25 | Lsi Corporation | Modeling approach for timing closure in hierarchical designs leveraging the separation of horizontal and vertical aspects of the design flow |
US8397197B1 (en) * | 2011-05-25 | 2013-03-12 | Applied Micro Circuits Corporation | Integrated circuit module time delay budgeting |
US20120319752A1 (en) * | 2011-06-17 | 2012-12-20 | Telefonaktiebolaget Lm Ericsson(Publ) | Look-up tables for delay circuitry in field programmable gate array (fpga) chipsets |
KR101857677B1 (en) * | 2011-07-21 | 2018-05-14 | 에스케이하이닉스 주식회사 | Semiconductor integrated circuit and method of transmitting signal thereof |
US9558309B2 (en) * | 2014-05-09 | 2017-01-31 | University Of Southern California | Timing violation resilient asynchronous template |
WO2016012780A1 (en) * | 2014-07-23 | 2016-01-28 | Iruiz Technologies Ltd | Improvements related to forecasting systems |
CN106326510B (en) * | 2015-06-30 | 2020-02-21 | 新思科技(上海)有限公司 | Verifying clock tree delays |
CN105095604B (en) * | 2015-09-10 | 2017-12-15 | 中国人民解放军国防科学技术大学 | A kind of consideration distance and the timing budget method of clock |
US9632726B1 (en) * | 2015-10-09 | 2017-04-25 | Arch Systems Inc. | Modular device and method of operation |
US10073939B2 (en) * | 2015-11-04 | 2018-09-11 | Chronos Tech Llc | System and method for application specific integrated circuit design |
US10599788B2 (en) * | 2015-12-30 | 2020-03-24 | International Business Machines Corporation | Predicting target characteristic data |
US10783311B2 (en) * | 2016-10-31 | 2020-09-22 | Synopsys, Inc. | DRC processing tool for early stage IC layout designs |
CN106874593B (en) * | 2017-02-13 | 2020-11-13 | 上海兆芯集成电路有限公司 | Digital electronic device design adjustment method and server |
CN108984806B (en) * | 2017-05-31 | 2022-02-18 | 深圳市中兴微电子技术有限公司 | Clock tree synthesis method and computer readable storage medium |
CN107944073B (en) * | 2017-10-12 | 2021-01-08 | 北京时代民芯科技有限公司 | Ring oscillation integrated circuit for multi-channel time measurement |
CN108038312B (en) * | 2017-12-13 | 2021-08-03 | 嘉兴倚韦电子科技有限公司 | Integrated circuit semi-custom back-end design timing sequence budget method |
CN108647422B (en) * | 2018-05-03 | 2022-05-10 | 天津芯海创科技有限公司 | Port delay constraint method and device |
CN109376467B (en) * | 2018-11-19 | 2022-12-06 | 武汉工程大学 | Clock tree layout flow method and clock tree deviation compensation device in integrated circuit |
CN109444716B (en) * | 2018-11-27 | 2021-08-10 | 中科曙光信息产业成都有限公司 | Scanning test structure with positioning function and method |
CN109406902B (en) * | 2018-11-28 | 2021-03-19 | 中科曙光信息产业成都有限公司 | Logic scanning aging test system |
-
2019
- 2019-03-29 US US16/370,008 patent/US10810344B1/en active Active
-
2020
- 2020-03-30 CN CN202010235960.1A patent/CN111539176B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN111539176A (en) | 2020-08-14 |
CN111539176B (en) | 2023-04-07 |
US10810344B1 (en) | 2020-10-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8160858B2 (en) | Systems and methods of efficient library characterization for integrated circuit cell libraries | |
US8839171B1 (en) | Method of global design closure at top level and driving of downstream implementation flow | |
US7865855B2 (en) | Method and system for generating a layout for an integrated electronic circuit | |
US6026228A (en) | Integrated circuit design method, database apparatus for designing integrated circuit and integrated circuit design support apparatus | |
US8689158B2 (en) | System and method for performing static timing analysis in the presence of correlations between asserted arrival times | |
US8307317B2 (en) | Statistical on-chip variation timing analysis | |
US7467367B1 (en) | Method and system for clock tree synthesis of an integrated circuit | |
US10915685B1 (en) | Circuit stage credit based approaches to static timing analysis of integrated circuits | |
Papa et al. | Physical synthesis with clock-network optimization for large systems on chips | |
US10990733B1 (en) | Shared timing graph propagation for multi-mode multi-corner static timing analysis | |
US11003821B1 (en) | Deterministic loop breaking in multi-mode multi-corner static timing analysis of integrated circuits | |
US6976235B2 (en) | Region-based voltage drop budgets for low-power design | |
US11574101B2 (en) | Techniques for providing optimizations based on categories of slack in timing paths | |
US20120221995A1 (en) | System and method for employing signoff-quality timing analysis information concurrently in multiple scenarios to reduce dynamic power in an electronic circuit and an apparatus incorporating the same | |
US11748548B2 (en) | Hierarchical clock tree implementation | |
US8549448B2 (en) | Delay optimization during circuit design at layout level | |
US11042678B2 (en) | Clock gate latency modeling based on analytical frameworks | |
US9110689B2 (en) | Automatic pipeline stage insertion | |
US9047434B2 (en) | Clustering for processing of circuit design data | |
US10963610B1 (en) | Analyzing clock jitter using delay calculation engine | |
US10956639B1 (en) | Method to reduce full-chip timing violation through time budgeting in integrated circuit design | |
US10810344B1 (en) | Multi-instantiation time budgeting for integrated circuit design and manufacturing | |
US9600613B1 (en) | Block-level code coverage in simulation of circuit designs | |
US9384309B2 (en) | Global timing modeling within a local context | |
US20190392089A1 (en) | Automated region based optimization of chip manufacture |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
FEPP | Fee payment procedure |
Free format text: ENTITY STATUS SET TO UNDISCOUNTED (ORIGINAL EVENT CODE: BIG.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
STCF | Information on status: patent grant |
Free format text: PATENTED CASE |
|
AS | Assignment |
Owner name: HYGON INFORMATION TECHNOLOGY CO., LTD., CHINA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:LIANG, HONGCHANG;REEL/FRAME:055139/0035 Effective date: 20201204 Owner name: ARCADIA INNOVATION INCORPORATED, CALIFORNIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:LIN, YIZHOU;REEL/FRAME:055139/0050 Effective date: 20201106 Owner name: HIGON AUSTIN R & D CENTER CORPORATION, TEXAS Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:TANG, JIAN;REEL/FRAME:056193/0197 Effective date: 20210110 |
|
MAFP | Maintenance fee payment |
Free format text: PAYMENT OF MAINTENANCE FEE, 4TH YEAR, LARGE ENTITY (ORIGINAL EVENT CODE: M1551); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY Year of fee payment: 4 |