Timing Closure Guide

September 3, 2017 | Author: jagadish_05ec752258 | Category: Program Optimization, Mathematical Optimization, Electronic Data Interchange, Capacitor, Library (Computing)
Share Embed Donate


Short Description

Download Timing Closure Guide...

Description

EDI Timing Closure Guide

Cadence Design Systems, Inc.

Application Note

Timing Closure Guide

Encounter Digital Implementation (EDI) System, 10.1 Rev – 1.0 August - 2011

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 1

EDI Timing Closure Guide

Table Of Content Introduction .................................................................................................................. 4 Recommended Timing Closure Flow ........................................................................... 4 Software ................................................................................................................... 4 Foundation Flow....................................................................................................... 4 Data Preparation and Validation .................................................................................. 5 Data Preparation ...................................................................................................... 5 Data Validation ......................................................................................................... 8 Determining and Setting the RC Scaling Factors ................................................... 12 Data Preparation and Validation for Low Power Designs....................................... 12 Flow Preparation........................................................................................................ 13 Setting the Design Mode ........................................................................................ 13 Extraction ............................................................................................................... 14 Timing Analysis ...................................................................................................... 15 Pre-Placement Optimization ...................................................................................... 16 Floorplanning and Initial Placement ........................................................................... 17 Ensuring Routability ............................................................................................... 18 Validating the Floorplan ......................................................................................... 20 Timing-Driven Placement ....................................................................................... 20 Pre-CTS Optimization ................................................................................................ 23 Guidelines for Pre-CTS Optimization ..................................................................... 23 Pre-CTS optDesign Command Sequences............................................................ 24 Path Group Optimization ........................................................................................ 27 Checking & Debugging Timing ............................................................................... 28 Clock Tree Synthesis ................................................................................................. 30 Creating the Clock Specification File...................................................................... 31 Synthesizing the Clock Tree .................................................................................. 32 Analyzing and Debugging the Clock Tree Results ................................................. 36 Optimizing the Clock Tree ...................................................................................... 36 Clock Specification File Example ........................................................................... 37 Post-CTS Optimization .............................................................................................. 39 Post-CTS Optimization Command Sequences ...................................................... 40 Hold Optimization ................................................................................................... 41 COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 2

EDI Timing Closure Guide

Checking Timing .................................................................................................... 45 Detailed Routing ........................................................................................................ 45 Improving Timing during Routing ........................................................................... 45 Routing Command Sequence ................................................................................ 45 PostRoute Extraction ............................................................................................. 47 Checking Timing .................................................................................................... 48 Post-Route Optimization ............................................................................................ 48 Data Preparation .................................................................................................... 49 Post-Route Optimization Command Sequences .................................................... 50 Checking Timing .................................................................................................... 52 Optimizing With 3rd Party SPEF or SDF ................................................................ 52 Timing Sign Off .......................................................................................................... 53 Additional Resources ................................................................................................. 54

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 3

EDI Timing Closure Guide

Introduction Achieving timing closure on a design is the process of creating a design implementation that is free from logical, physical, and design rule violations and meets or exceeds the timing specifications for the design. For a production chip, all physical effects, such as metal fill and coupling, must be taken into account before you can confirm that timing closure has been achieved. Timing Closure is not just about timing optimization. It is a complete flow that has to converge, including placement, timing optimization, clock tree synthesis (CTS), routing and SI fixing. Each step has to reach the expected targets or else timing closure will likely not be achieved. This document discusses each step in the implementation flow as it relates to timing closure in EDI 10.1, and provides recommended settings specific to high performance, congested or high utilization designs.

Recommended Timing Closure Flow Below is a high level diagram showing the steps in the timing closure flow:

Software The EDI System software is constantly being improved to provide better quality of results, reliability and ease of use. To ensure you're running with the latest improvement we recommend running with the latest software version available from http://downloads.cadence.com. Foundation Flow If you've developed your own flow scripts, you know maintaining and updating them can be time consuming and error prone. Also, ensuring you're running with the latest recommended commands and options can be challenging. Therefore, we recommend using the Foundation Flow scripts. The Foundation Flow provides Cadence-recommended procedures for implementing flat, hierarchical, and low-power/CPF designs using the EDI COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 4

EDI Timing Closure Guide

System software. The Foundation Flow is a starting point for building an implementation environment, but you can augment them with designspecific content. Utilizing the Foundation Flow helps achieve timing closure because it provides the latest recommended flow which you can further customize based on your design requirements. If you are new to the Foundation Flow, we recommend you start with the Foundation Flow video demonstrations. These provide examples of setting up and using the flows. They are accessible from within the EDI System GUI by selecting Flows - Foundation Flow Demo.

Data Preparation and Validation This section outlines the data (libraries, constraints, netlist, etc.) required for implementing the timing closure flow and how to validate that data. The goals of data preparation and validation include: • • • • •

Confirming that the EDI System has a complete and consistent set of design data (all library views and versions must be consistent). Ensuring that all tools in the flow interpret the timing constraints consistently. Making sure logically equivalent cells are defined properly. Creating a capacitance table file that matches the process technology. Correlating parasitics among the prototyping and sign-off extraction tools.

Data Preparation This section lists the data required and data setup recommended for the timing closure flow. Timing Libraries •

Every cell used in the design should be defined in the timing library. o If multiple delay corners are being analyzed then each cell needs to be characterized for each corner. o EDI System supports Non-linear Delay Models (NLDM), ECSM and CCS.  ECSM/CCS require the Signal Storm delay calculator (setDelayCalMode -engine signalStorm)  CCS/ECSM are less pessimistic than NLDM and therefore you can gain about 5% to 10% of the clock period on the slack using CCS/ECSM libraries.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 5

EDI Timing Closure Guide

Physical Libraries • • •

You need to have an abstract defined for every cell in either a LEF file or OpenAccess database. Define Non-Default Rules (NDRs) for routing as needed. These can be defined in the LEF or added within EDI System. The technology LEF should have an optimized set of vias to be used for routing. Confirm you have the latest technology LEF from your library vendor or foundry.

Verilog Netlist •

Check the Verilog netlist for assign statements. If assign statements exist, use the following procedure to enable optimization to work on those nets: o Run "setDoAssign on" before loading the design data OR use "set rda_Input(assign_buffer) {1}" in the configuration file (these commands do the same thing). o And set "setImportMode -bufferFeedThruAssign true". • The netlist should also be unique. o Run uniquifyNetlist at the Linux command line to uniquify the netlist.

Timing Constraints Timing constraints in the form of SDCs are required. You should have an SDC file for each operational mode required for analysis. Capacitance Table File A capacitance table is used by EDI System to accurately extract parasitics. A cap table is required for each RC corner to be analyzed. You may obtain the cap table(s) from your foundry. Alternatively, the generateCapTbl command can be used to generate the cap table. generateCapTbl reads a detailed process description file in Interconnect Technology (ICT) format and a LEF technology file, and outputs a capacitance table file. The Interconnect Technology (ICT) file describes the detailed process information for a given technology (layer thicknesses, materials, profiles, dielectric constants, and so forth). Any non-default rules for the design should be defined in the LEF file you are using.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 6

EDI Timing Closure Guide

The capacitance table file consists of two parts: •

Basic Capacitance Table Provides a table of spacing versus capacitance information for each layer.



Extended Capacitance Table Provides encoded extraction patterns that are derived using a 3D field solver which provides much greater accuracy during detailed extraction. A capacitance table file (basic or extended) is processspecific and not design dependent, so it only needs to be created once per technology. If a capacitance table already exists for your process, use it and do not recreate it.

The following example shows how to generate an extended capacitance table using the default field solver (Coyote). generateCapTbl -lef tech.lef -ict process.ict -output process.xCapTbl QRC Tech File A QRC technology file (ICT file) is required in order to run turbo-QRC (tQRC), integrated QRC (iQRC) or standalone QRC. A QRC tech file is required for each RC corner to be analyzed. Signal Integrity (SI) Libraries SI libraries in the form of cdB's are required for SI analysis and optimization. These can be generated from SPICE using the make_cdb command. Multi-Mode Multi-Corner (MMMC) Setup Multi-Mode Multi-Corner (MMMC) setup is required for optimizing designs over multiple operating conditions. MMMC setup is also required for accurate push-out/pull-in reporting during Signal Integrity (SI) analysis. Therefore, we highly recommend setting up your analysis views using the MMMC environment. This will be required in EDI System 11.1. See the chapter Performing Multi-Mode Multi-Corner Timing Analysis and Optimization in the EDI System User Guide for information on defining the MMMC environment. The flow described in this document assumes you are running in MMMC mode.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 7

EDI Timing Closure Guide

Data Validation This section explains how to identify problems when importing the data and checks you can run to catch data issues early in the flow. Loading the Design Once you have prepared the necessary data you are ready to import it. Run loadConfig to import the libraries, netlist and timing environment. loadConfig design.conf loadConfig executes a number of checks to validate the data and highlight problems. It is important you review the log file to understand and resolve warnings and error messages it reports. We recommend using the Log Viewer (Tools - Log Viewer) to review the loadConfig results. The Log Viewer expands/collapses the log output by command and color codes messages to make them easy to identify. Following are things to look for when reviewing the loadConfig output: •

loadConfig reports cells in the LEF which are not defined in the timing libraries. Look for the following and confirm if these cells need to be analyzed for timing: **WARN: (ENCSYC-2): Timing is not defined for cell INVXL.



A blackbox is an instance declaration in the netlist for which no module or macro definition is found. Unless your design is being done using a blackbox style of floorplanning, there should be no blackboxes in the design. If there are blackboxes reported, be sure to load the Verilog file that defines the logic module, and make sure you include the LEF file that defines the macro being referenced in the netlist. The following is reported for blackbox (empty) modules: *** Found empty module (bbox).



Verify the netlist is unqiue. The following is reported if it is not unique: *** Netlist is NOT unique.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 8

EDI Timing Closure Guide



By default, EDI System utilizes a "footprintless" flow. This means instead of relying on the "footprint" definitions inside the timing libraries, it uses the "function" statement to determine cells which are functionally equivalent and can be swapped during optimization. Additionally, it identifies buffers, inverters and delay cells. Inconsistencies in how the cell functions are defined can lead to suboptimal or erroneous optimization results. Review the log file to confirm the buffers, inverters and delay cells are properly identified. Below is an example of what you'll see: List of usable buffers: BUFX2 BUFX1 BUFX12 BUFX16 BUFX20 BUFX3 BUFX4 BUFX8 BUFXL CLKBUFX2 CLKBUFX1 CLKBUFX12 CLKBUFX16 CLKBUFX20 CLKBUFX3 CLKBUFX4 CLKBUFX8 CLKBUFXL Total number of usable buffers: 18 List of unusable buffers: Total number of unusable buffers: 0 List of usable inverters: CLKINVX2 CLKINVX1 CLKINVX12 CLKINVX16 CLKINVX20 CLKINVX3 CLKINVX4 CLKINVX8 CLKINVXL INVX1 INVX2 INVX12 INVX16 INVX20 INVX3 INVXL INVX4 INVX8 Total number of usable inverters: 18 List of unusable inverters: Total number of unusable inverters: 0 List of identified usable delay cells: DLY2X1 DLY1X1 DLY4X1 DLY3X1 Total number of identified usable delay cells: 4 List of identified unusable delay cells: Total number of identified unusable delay cells: 0 Also, look for cells which do not have a function defined for them: No function defined for cell 'HOLDX1'. The cell will only be used for analysis.

Checking Timing Constraint Syntax In addition to checking the libraries, the loadConfig command also checks the syntax of timing constraints. After running loadConfig, check for the following problems: •

Unsupported constraints o The EDI System software may not support the SDC constraints being used with the design, or the constraints may not match the netlist. If the constraints are not supported, they may need

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 9

EDI Timing Closure Guide

to be re-expressed (if possible) in constraints that the EDI System software does support. •

Ignored timing constraints o Syntax errors can cause the tools to ignore certain constraints resulting in the misinterpretation of important timing considerations. Check for warnings or errors about unaccepted SDC constraints. o The following are possible causes for ignored constraints.  A design object is not found. If the constraints refer to pins, cells, or nets that are not found in the netlist, then consider the following possible causes:  There could be a naming convention problem in the constraint file.  The netlist and constraints are out of sync, and a new set of constraints and/or a new netlist needs to be obtained.  An incorrect type of object is being passed to a constraint.  An option is being used incorrectly or an unknown option is used.  Illegal endpoints are used in assertions. Use the primary IOs (top-level ports, CK or D register pin) to define starting and endpoints of set_false_path and set_multicycle_path. A combinatorial pin or a Q register pin is not valid.



Other things to consider when defining constraints: o set_ideal_net, set_ideal_network -> will prevent optimization on these nets o set_propagated_clock -> will limit preCTS optimization by not allowing resize on sequential elements. Set this only after clock trees are inserted. o set_dont_use, set_dont_touch -> confirm that the proper settings are used o Have a constraint file for every mode required for signoff timing analysis o Understand and adjust clock uncertainty depending on the stage of the design flow (preCTS / postCTS / postRoute / signoff).

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 10

EDI Timing Closure Guide

Cap Table Checks •

• •

Make sure the cap table is current and generated with a recent version of generateCapTbl. This ensures the cap table information is used most effectively by extraction. Make sure the cap table, ICT, and LEF files match. The layer names, widths, spacings and pitches should be consistent between the files. Use correct temperature for resistance extraction.

Validating Timing Constraints As described in the previous section, the loadConfig command checks the syntax of specified timing constraints. However, it is also important to ensure that the timing constraints are valid for the design. A good first-pass method is to check the zero wire-load model timing. To validate timing constraints, use the command: timeDesign -prePlace -outDir preplaceTimingReports This command generates a quick timing report using zero wire load and provides a first indication, before placement and routing, of how much effort will be required to close timing and whether the timing constraints are valid for the design. During pre-placement timing analysis, high fanout nets are temporarily set as ideal so that more immediate timing issues can be addressed first. Additionally, you can run the command "check_timing -verbose" to report timing problems the Common Timing Engine (CTE) sees. Checking Logically Equivalent Cells Available for Optimization •



Run "checkFootPrint" to report any problems with footprint functions. o If there are problems reported run "reportFootPrint -outfile file_name" command to create a footprints file. You can review this file to see which cells are identified as logically equivalent. o You can edit the footprints file if needed and run "loadFootPrint infile file_name" to load it. If you made updates, run checkFootPrint again to verify the file you loaded does not have problems.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 11

EDI Timing Closure Guide

checkDesign command You can run "checkDesign -all" to check the design after loadConfig. This will run a number of checks and output the results to a text file. Review the file to understand any problems found. Determining and Setting the RC Scaling Factors RC Scaling factors are recommended to correlate EDI System's native parasitic extractor with your signoff extractor. This provides more accurate timing and predictability throughout the flow. Use Ostrich to obtain the parasitic measurements to determine the appropriate scaling factors. This command calculates the capacitance factors by comparing EDI System extraction with the results of either QRC Extraction or a SPEF file. Please see the solution How to Generate Scaling Factors for RC Correlation for the steps to generate the correlation factors with Ostrich. Once the scaling factors are determined specify them in your MMMC setup using the create_rc_corner or update_rc_corner commands. Data Preparation and Validation for Low Power Designs If your design is utilizing a low power flow using a Common Power Format (CPF) file you should also check the following: •

• • •

Make sure the CPF points to the proper libraries and constraints. o When using CPF most of the MMMC setup is defined within the CPF. Delay corner names are 'CPF-generated' so keep that in mind when attaching RC corners and derating timing. For low power designs utilizing power shutdown, make sure an alwayson buffer is available and usable. Run runCLP to verify the CPF and make sure the results are well understood before proceeding.

Optimization of leakage and/or dynamic power is typically on top of the presented flows: •

For designs where leakage is a high priority, the recommended flow is to enable leakage optimization at the beginning of the flow and allow the tool to manage the optimization. This is done by setting: setOptMode -leakagePowerEffort high

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 12

EDI Timing Closure Guide

o

o



High effort leakage does come with a run time and area penalty and is particularly time consuming as the ratio of higher VT cells drops. Low leakage effort, enabled via "setOptMode -leakagePower low", only does postRoute reclaim.  This flow has the least impact to run time but typically results in a 5-10% loss of potential leakage savings.

For designs where dynamic power is a high priority, dynamic power optimization can be enabled with the following and also set at the beginning of the flow: setOptMode -dyanmicPowerEffort [ low | high ] o

High effort does reclaim throughout the flow. Low effort only performs postRoute downsizing.  This method typically works best with a VCD or TCF to properly apply the activity rates.

Flow Preparation Setting the design mode and understanding how extraction and timing analysis are utilized during the flow are important to achieving timing closure. Setting the Design Mode setDesignMode specifies the process technology value and flow effort level. •

"setDesignMode -process" specifies the process technology you are designing at. You can use this command to change the process technology dependant default settings globally for each application instead of setting several mode options. When you specify a process technology value using the setDesignMode command, the EDI System software assigns coupling capacitance threshold values to the RC extraction filters automatically. These values determine whether the coupling capacitances of the nets in a design will be lumped to the ground. Note: In detailed extraction mode, the grounding of coupling capacitances also depends on the capacitance filtering mode set by the -capFilterMode parameter of the setExtractRCMode command.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 13

EDI Timing Closure Guide



"setDesignMode -flowEffort" can be used to force every supercommand to use their high-effort settings and the additional nondefault options. In this mode, the target is to achieve best possible timing/yield at the expense of some more CPU runtime. none : Leaves every super-command to use their default settings. In this mode, the target is to achieve good QoR with minimum runtime. high : Forces every super-command to use their high-effort settings and the additional non-default options. In this mode, the target is to achieve best possible timing/yield at the expense of some more CPU runtime.



The following example sets the process to 45nm and effort level to high: setDesignMode -process 45nm -effortLevel high

Extraction Resistance and Capacitance (RC) extraction using the extractRC command is run frequently in the flow each time timing analysis is performed. The setExtractRCMode options -engine and -effortLevel control which extractor is used by extractRC. The lower the effort level, the faster extraction runs at the expense of being less accurate. Fast extraction is used early in the flow to provide fast turnaround times so you can experiment with different floorplans and solutions. As you progress through the flow, the effort level is increased which improves the accuracy of the extraction at the expense of run time. The -engine option indicates whether to use the preRoute or postRoute extraction engine. •



Use "-engine preRoute" when the design has not been detail routed by NanoRoute yet. When "-engine preRoute" is set RC extraction is done by the fast density measurements of the surrounding wires; coupling is not reported. Use "-engine postRoute" after the design has been detail routed by NanoRoute. RC extraction is done by the detailed measurement of the distance to the surrounding wires; coupling is reported. The effortLevel parameter further specifies which postroute engine is used for balancing performance versus accuracy needs.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 14

EDI Timing Closure Guide

The -effortLevel value controls which extractor is used when the postRoute engine is used. • •





low - Invokes the native detailed extraction engine. This is the same as specifying the "-engine postRoute" setting. medium - Invokes the Turbo QRC (TQRC) extraction mode. TQRC performance and accuracy falls between native detailed extraction and IQRC engine. This engine supports distributed processing. o TQRC engine is recommended for process nodes < 65nm. Note: This setting does not require a QRC license. high - Invokes the Integrated QRC (IQRC) extraction engine. o IQRC provides superior accuracy compared to TQRC. IQRC is recommended for extraction after ECO. In addition, IQRC supports distributed processing. Note: IQRC requires a QRC license. signoff - Invokes the Standalone QRC extraction engine. This engine choice provides the highest accuracy. The engine has several run modes, thereby, providing maximum flexibility.

The default value for -effortLevel depends on the value of setDesignMode. The default for nodes above 65nm is low. TQRC (-effortLevel medium) is the default extraction engine in the postroute flow for 65nm and below design. However, TQRC and IQRC do not support the obsoleted 3 corner flow (defineRCCorner flow) and require a QRC Tech file. Therefore, Native Detailed (effortLevel low) engine remains the default engine if no QRC Tech file has been defined or if the defineRCCorner command has been used. Timing Analysis Timing analysis is typically run after each step in the timing closure flow using the timeDesign command. If timing violations exist, we recommend using the Global Timing Debug (GTD) GUI to analyze and debug the results: •

GTD is an invaluable tool which provides forms and graphs to help you visually see timing problems. o To learn more about GTD we recommend completing the Workshop provided with EDI 10.1. Download instructions are available at http://support.cadence.com/wps/mypoc/cos?uri=deeplinkmin:Vi ewSolution;solutionNumber=11681606.

Initial timing analysis should not be performed after placement. Instead, we recommend always waiting until after pre-CTS optimization to report your initial timing: COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 15

EDI Timing Closure Guide



After placement, there will be some basic buffering and resizing to get cells into their characterized region of timing as defined by their lookup tables. Until the cells are in that region, any timing analysis results are highly suspect.



A placement should always be followed by an optimization run to bring the cells into line prior to making a timing and routability judgment. It is not uncommon to have the original "worse" placement come out better post-optimization than the "better" placement. This seems nonintuitive, but it is often the case.

Pre-Placement Optimization Data preparation is complete and you are now ready to implement your design. It is important to note that using options from a previous design might not necessarily apply to your current design. Therefore, we recommend starting with the default flow (or Foundation Flow) then apply additional options based on your design requirements. Additionally, as you proceed through the flow you should investigate each flow step and validate it before moving to the next one. The goals of pre-placement optimization are to optimize the netlist to: • • • •

Improve the logic structure Reduce congestion Reduce area Improve timing

In some situations, the input netlist (typically from a poor RTL synthesis) is not a good candidate for placement since it might contain buffer trees or logic that is poorly structured for timing closure. In most cases, high-fanout nets should be buffered after placement. It is more reliable to allow buffer insertion algorithms to build and place buffer trees rather than to rely on the placer to put previously inserted trees in optimal locations. Additionally, having buffer trees in the initial netlist can adversely affect the initial placement. Because of these effects, it can be advantageous to run pre-placement optimization or simple buffer and double-inverter removal (area reclamation) prior to initial placement. This can be accomplished by using the deleteBufferTree and removeClockTree commands. Note deleteBufferTree is run by placeDesign by default. COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 16

EDI Timing Closure Guide

For designs where the logical structure of the critical paths or high congestion are the sources of closure problems, restructuring or remapping the elements on the paths (or related portions of the design) can provide better timing results. This is done by running netlist-to-netlist optimization using the runN2NOpt command on the initial netlist prior to place and route. Additional restructuring can be performed later in the flow using the EDI System optimization commands with more physical information.

Floorplanning and Initial Placement The goals of floorplanning and initial placement include: • • •

Creating prototypes using multiple iterations with a focus on routability Moving toward timing-driven placement as routability stabilizes Adding power routing once timing and congestion converge

The initial floorplan and placement have a primary impact on the performance of a design. EDI System allows you to use prototypes to analyze various placements and floorplans before you begin the optimization process. Prototyping allows you to create a floorplan that can be implemented with high confidence before you spend time and effort on optimization and routing. Prototyping involves multiple placement iterations that converge on a solution which meets a design's requirements for routability, timing (including clocks), power, and signal integrity. The initial floorplan drives the constraints leveraged by placement and partitioning to meet these objectives. The following steps outline a basic procedure for obtaining an initial placement. 1. Run the initial placement without any regions and guides. It is best to get a baseline placement without constraining the placer.

o

o

Early on use "setPlaceMode -fp true" to run placement in prototyping mode for faster turnaround.  Prototype placement does not produce legal placement so make sure you run placement with "setPlaceMode -fp false" as you converge on a floorplan. Use Automatic Floorplan Synthesis if your design contains a large number of hard macros.  Automatic Floorplan Synthesis is a set of nativelyintegrated automatic floorplan capabilities that can create a quick, prototype floorplan. Given a gate-level netlist and

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 17

EDI Timing Closure Guide

design physical boundary, Automatic Floorplan Synthesis can analyze the signal flow and generate a floorplan that includes automatic module and macro placement for large chips. See Automatic Floorplan Synthesis in the EDI System User Guide for more information. 2. Analyze the placement for timing and routability issues, and make necessary adjustments.

3. Employ module guides, placement blockages, and other techniques to refine the floorplan. The placement engine automatically detects lowutilized designs and turns on the options required to achieve an optimal placement. Ensuring Routability Initial prototype iterations should focus on routability as the key to achieving predictable timing closure. You should attempt to resolve congestion before attempting timing closure. Designs which are congested are more likely to have timing jumps during timing and Signal Integrity (SI) closure. Tools such as module guides, block placement, block halos, obstructions, and partial placement blockages (density screens) are used to control the efficient routing of the design. Use the following guidelines during floorplanning and placement to avoid congestion. 1. Choose an appropriate floorplan style. •



If possible, review a data flow diagram or a high-level design description from the chip designer to determine an appropriate floorplan style. Assess different floorplan styles such as hard macro placement in periphery, island, or doughnut (periphery and island). Keep the macro depth at 1 to 2 for best CTS, optimization, and Design-for-test (DFT) results. If possible, consider different aspect ratios to accommodate a shallower macro depth. Consider using Automatic Floorplan Synthesis and relative floorplanning constraints to simplify floorplan iterations.

2. Preplace I/Os and macros.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 18

EDI Timing Closure Guide

• •

Review hard macro connectivity and placement based on the minimum distance from a hard macro to its target connectivity. Preplace high-speed and analog cores based on their special requirements for noise isolation and power domains.

3. Review I/O placement to identify I/O anchors and associated logic. •

Verify that logic blocks and hard macros which communicate with I/O buffers are properly placed and have optimal orientation for routability. Push down into module guides to further assess the quality of the floorplan and resulting placement.

4. Allow enough space between preplaced blocks. •





Allow space between I/Os and peripheral macros for critical logic such as JTAG, PCI, or Power management logic. Use the specifyJtag and placeJtag commands prior to placing blocks. Use block halos, placement obstructions or fences around blocks prior to optimization or Clock Tree Synthesis (CTS). Placement generally does not do a good job of placing cells between macros. Reduce or remove halos and obstructions after placement to make sufficient space available around macros for optimization, CTS, DRV, or SI fixing to add buffers. Place other cells such as endcaps, welltaps and decaps prior to placement as required.

5. Use module guides carefully. •





Place module guides, regions, or fences only when greater control is required. Be careful not to place too many module constraints early in the floorplanning process because it is time consuming and greatly constrains the placement. Module guides should be used for floorplan refinement or hierarchical partitioning. Review the placement of module guides related to datapath and control logic relative to the associated hard macros. Datapath logic can be a source of congestion problems due to poor aspect ratios, high fanout, and large amounts of shifting. Consider tuning the locations of these module guides and lowering the density to reduce congestion. Review the placement of module guides related to memories. These modules can typically have higher densities due to the inclusion of the memories.

6. Reorder scan chains

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 19

EDI Timing Closure Guide



When evaluating congestion, make sure that scan chains are reordered to eliminate "false" hot spots in the design. Failing to reorder the scan chain can cause a routable floorplan to appear unroutable. By default, placeDesign performs scan tracing and scan reordering based on global or user-specified scan reorder settings and specified or imported scan chain information. If scan chain information is missing, no reordering is performed.

Validating the Floorplan A congested or unroutable design at this stage will not get better during optimization. With a good floorplan you should be able to: • •

Place the design in the floorplan without issues Create a routeable placement

Make sure to consider the following when finalizing the floorplan: •



• •

The power grid should be defined: o Global net connections properly defined using globalNetConnect command. o Nets requiring optimization should be defined as signal (regular) nets. Optimization treats nets in the SPECIALNETS as dont_touch. o PINS marked with + SPECIAL cannot be optimized. o Followpin routing should align to rows/cells with correct orientation (VDD pin to VDD followpin). Gaps between standard cells and blocks should be covered with soft or hard blockages. Placement cannot place cells in the area of soft blockages but optimization and CTS can. All blocks should be marked fixed (setBlockPlacementStatus allHardMacros -status fixed). Tracks should match IO pins and placement grid (rows). Use generateTracks to update the tracks.

Timing-Driven Placement As the routability of the floorplan stabilizes, you should shift your focus to timing-driven placement. Timing-driven placement (setPlaceMode -timingDriven true) and placement optimization (buffer tree deletion) are enabled by default: placeDesign COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 20

pre-

EDI Timing Closure Guide

Following are additional tips for performing placement on high performance, high congestion and high utilization designs. Tips for Placing High Performance Designs •

By default, optDesign optimizes only the most critical paths. For high performance designs it can help to optimize all paths. "setOptMode allEndPoints true" enables optimization of all paths and replaces "setOptMode -criticalRange 1.0" from previous releases. setOptMode -allEndPoints true Note in EDI 10.1 we no longer recommend running "placeDesign inPlaceOpt". Furthermore, the -inPlaceOpt option is ignored if "setOptMode -allEndPoints true" or "setDesignMode flowEffort high" is enabled. Pre-CTS optimization has been improved in EDI 10.1 and therefore -inPlaceOpt is no longer needed. Overall, running "placeDesign + optDesign -preCTS" with "setOptMode -allEndPoints true" enabled reduces the run time of placement and pre-CTS optimization while achieving similar results as "placeDesign inPlaceOpt + optDesign -preCTS".



The following will disable buffer tree deletion prior to placement. Usually, removing the existing buffer tree before placement produces better timing result and wire length for the majority of designs. However, in some designs, keeping the existing buffer tree generates better timing results. placeDesign –noPreplaceOpt The following option is useful when clock gating checks are in the critical paths and to help limit jumps seen by CTS moving gating elements. You must read in the clock tree specification file prior to placeDesign for this option to work. Use "specifyClockTree -file specFile". setPlaceMode -clkGateAware true



Net weights can be used in specific situations to instruct the placer to minimize the distance between the driver and sinks. Net weights should be used on a limited basis when the designer determines the placer needs to put more effort placing specified logic closer together. For example, if there are half cycle paths to or from memories or latches, these nets may benefit from higher net weights. Or if you

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 21

EDI Timing Closure Guide

need greater control over how the fanout of clock gating cells are placed. Use specifyNetWeight to assign a greater weight to a net (default is 2). For example: specifyNetWeight netName 10 Tips for Placing Congested or High Utilization Designs When optimization contributes to the congestion or the design has high utilization, running in high effort mode helps (High utilization is typically associated with utilizations > 80%, but depending on the technology, 70% might already start to be difficult to handle): setDesignMode -flowEffort high true The following option increases the numerical iterations and makes the instance bloating more aggressive. It also automatically enables the congRepair command. setPlaceMode -congEffort high The following optimizes wire length by swapping cells. Optimization can reduce the total wire length up to two percent without degrading placement. In high-effort mode, it might add up to ten percent to the run time. setPlaceMode -wireLenOptEffort high •

If clock routing is a source of congestion due to NDRs or other rules, setting the clock attributes before placement and allowing the placer and trialRoute to see the effect can help. Here is some example code that can be called before placement. The reason timeDesign is called is to make sure CTE does the clock net marking to identify all the clocks: specifyClockTree -file changeClockStatus -all -useClock timeDesign -prePlace -outDir RPT -prefix preplace dbForEachCellNet [dbgTopCell] net { if {[dbIsNetDefInClock $net]==1||[dbIsNetClock $net]==1} { setAttribute -net $net -non_default_rule \ -avoid_detour true -bottom_preferred_routing_layer 4 \ -top_preferred_routing_layer 6 \ -preferred_extra_space 1 } }

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 22

EDI Timing Closure Guide



There is a standalone command called congRepair which can be called in any part of the pre-route flow to attempt to relieve congestion. The command uses trialRoute + incremental placement. This can have a significantly detrimental effect on timing and often will require additional optDesign calls (and may not converge). So congRepair should be used with caution.

Pre-CTS Optimization The goal of pre-CTS optimization is to repair • • •

Setup slack (WNS - Worst Negative Slack) Design rule violations (DRVs) Setup times (TNS - Total Negative Slack)

Use the "optDesign -preCTS" command only if optimization has not been done already. During subsequent optimization runs, use the -incr option. The use of the optDesign command for post-CTS optimization and post-route optimization is described in the corresponding sections of this document. For more information on the optDesign command, see optDesign in the EDI System Text Command Reference, and refer to the Optimizing Timing chapter in the EDI System User Guide for more procedures on optimizing the design. Guidelines for Pre-CTS Optimization Consider the following points during pre-CTS optimization: • •



Handle assign statements correctly. See "Data Preparation" for more information. Set input transitions for the high fanout nets for delay calculation to use on high fanout networks. o For nets with a fanout higher than 1000, the default transition is 120ps, the default delay is 1ns, and the default capacitance is 0.5pf. o The defaults can be changed with the commands: setDefaultNetDelay, setDefaultNetLoad, and setInputTransitionDelay. o During optimization, all nets with a fanout greater than 100 will be buffered with the exception of clock nets. Make sure all required views are active using set_analysis_view.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 23

EDI Timing Closure Guide

Pre-CTS optDesign Command Sequences You can use optDesign for pre-CTS optimization in the following ways. You can use any of these features separately or in combination. Additionally, you may choose to optimize specific timing paths using path groups. See the section "Path Group Optimization" later in this document for details. •

To optimize timing placed designs for the first time (with ideal clocks), use the following command: optDesign -preCTS



To perform rapid timing optimization for design prototyping, use the following commands: setOptMode -effort low optDesign -preCTS In this mode optDesign performs gate resizing and global buffer insertion, but does not perform netlist restructuring.



To further optimize a design after you have already run optDesign preCTS, use the following command: optDesign -preCTS -incr

The following are additional tips for performing preCTS optimization on high performance, congested and high utilization designs. Tips for Optimizing High Performance Designs •

Enable optimization for all end points by setting the following. This should only be set if the WNS/TNS are manageable. If there is very large TNS or high number of violating paths this could significantly increase runtime and area. Note "setDesignMode -flowEffort high" automatically sets "setOptMode -allEndPoints true". setOptMode -allEndPoints true

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 24

EDI Timing Closure Guide



Useful skew is often required on the toughest designs. In preCTS, useful skew optimization advances and delays sequential elements, and often has little to no penalty in terms of the size of the tree. In postCTS, useful skew optimization delays sequential elements and physically adds additional delay through the use of buffers. Use setUsefulSkewMode to control the buffer types and other options used by useful skew optimization. To enable useful skew optimization run: setOptMode -usefulSkew true



The following option temporarily over constrains clock gating checks during preCTS optimization so the tool is able to work harder before the actual tree is inserted. Often clock gating checks don't become visibly critical until the tree is synthesized due to the larger clock skew between them and the enabling register. This option requires the clock tree specification file to be specified prior to optDesign (specifyClockTree -file specFile). setOptMode -clkGateAware true If clock gating cells are still on the critical path after performing placement and optimization using -clkGateAware true consider using clock gate decloning and/or cloning, or over-constraining the clock gating paths. Clock Gate Cloning & Decloning Clock gate decloning merges clock gating cells with the same inputs. Clock gate cloning duplicates clock gating cells and redistribute their gated loads. An example flow using clock gate decloning and cloning is: # Specify the clock tree constraints and run clock gate aware placement: specifyClockTree -file ctsConstraintsFile setPlaceMode -clkGateAware true placeDesign # Declone the clock gates (optional step) ckDecloneGate -ignoreDontTouch -ignorePreplaced # Clone the clock gates based on placement (optional step)

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 25

EDI Timing Closure Guide

ckCloneGate # Perform preCTS optimization setOptMode -clkGateAware true optDesign -preCTS # Clone the clock gates in time driven mode. This considers the setup timing # to the clock gating cells. ckCloneGate -timingDriven timeDesign -preCTS # Run preCTS optimization again (optional) optDesign -preCTS Over-constraining Clock Gating Paths You can create separate path groups for the clock gating cells and over constrain them. Following is a sample script to do this. Set the clkgate_target_slack to your desired value. Use care when over constraining designs as it increases run time: # Set target overshoot slack for clock gating elements preCTS (in nanoseconds): set clkgate_target_slack 0.15 # Turn on pathgroup flow setAnalysisMode -honorClockDomains false # Create separate reg2reg and clkgate groups group_path -name reg2reg -from [all_registers] -to [filter_collection [all_registers] "is_integrated_clock_gating_cell != true"] group_path -name clkgate -from [all_registers] -to [filter_collection [all_registers] "is_integrated_clock_gating_cell == true"] setPathGroupOptions reg2reg -effortLevel high -criticalRange 1 setPathGroupOptions clkgate -effortLevel high -targetSlack $clkgate_target_slack -criticalRange 1 Tips for Optimizing Congested and High Utilization Designs •

The risk for timing optimization on designs with high routing congestion mainly comes from nets detouring that are unpredictable. In addition to floorplan and placement guidances made previously, it is usually beneficial to identify the weak cells and to set them as dont_use.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 26

EDI Timing Closure Guide

setDontUse cellName(s) Note it's recommended to use setDontUse rather than the SDC set_dont_use. This is because setDontUse applies to all operating modes while set_dont_use only affects the operating mode(s) it is applied to. The following reduces local congestion during timing optimization: setOptMode -congOpt true •

The following option will run additional iterations for resolving routing congestion in trialRoute: setTrialRouteMode -highEffort true Allowing M1 routing can help to reduce congestion and thus removes some pessimism during RC extraction: setTrialRouteMode -useM1 true



Monitor the density increase. If needed, reduce/remove extra margins on setup target and DRV. Usually the density starts blowing up when trying to fix the last tens of picoseconds and the better timing seen pre-CTS will lead to flow divergence later on. Properly setting the setup target slack will reduce area and improve runtime. For example, if the target is -0.5ns, it might be helpful to set "setOptMode setupTargetSlack -0.2". The same applies to DRV fixing by using drcMargin option: setOptMode -setupTargetSlack -drcMargin



Depending on the quality of the initial netlist, it could be that some redundant or useless logic could be safely removed to reduce area. setOptMode -simplifyNetlist true

Path Group Optimization You can focus timing optimization on specific paths using path groups. The flow to create and optimize path groups is as follows. Use setPathGroupOptions to control how optimization works on each path group:

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 27

EDI Timing Closure Guide

clearClockDomains setAnalysisMode -honorClockDomains false group_path [-name path_group_name] [-from from_list] [-to to_list] [through through_list] setPathGroupOptions ... optDesign -preCTS [-incr] For example, you could create a path group to all the D inputs of SRAM instances using wildcards: clearClockDomains setAnalysisMode -honorClockDomains false group_path -name to_srams -to top/inst_sram*/D* setPathGroupOptions to_srams -slackAdjustment 1000 slackAdjustmentPriority 10 optDesign -preCTS [-incr] Here is another example where separate path groups are created for reg2reg and clock gating paths. The goal here is to over constrain the clock gating paths so optDesign works harder on them: setAnalysisMode -honorClockDomains false group_path -name reg2reg -from [all_registers] -to [filter_collection [all_registers] "is_integrated_clock_gating_cell != true"] group_path -name clkgate -from [all_registers] -to [filter_collection [all_registers] "is_integrated_clock_gating_cell == true"] setPathGroupOptions reg2reg -effortLevel high -criticalRange 1 setPathGroupOptions clkgate -effortLevel high -targetSlack 0.15 criticalRange 1 optDesign -preCTS [-incr] Checking & Debugging Timing You can use the command timeDesign -preCTS to check the current timing: timeDesign -preCTS -outDir preCTSOptTiming If timing violations exist using Global Timing Debug to analyze the violations. Following are suggestions for resolving common types of timing problems seen during pre-CTS optimization: •

Some nets are not being optimized. Run "reportIgnoredNets -outfile filename" to output a report of ignored nets during optimization. Use

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 28

EDI Timing Closure Guide

the abbreviation in the last column with the key at the bottom of the file to determine why certain nets are not being optimized. •

Monitor "reportFootprint -reportDontTouchNUse -outfile fileName" to identify cells with a set_dont_touch or set_dont_use attribute set. The file will identify these cells in the far right column. Make sure cells intended for optimization are available to use.



Similar paths are not meeting timing. Create a custom path group for these paths and optimize them separately. See "Path Group Optimization" earlier in this section.



Critical path(s) go through congested area. Try using cell padding (specifyCellPad or specifyInstPad commands) or partial placement blockages (also known as density screens) to reduce the congestion.



Evaluate the timing constraints and identify false paths that may be affecting the critical path. Conformal Constraint Designer offers a complete, functional constraint validation solution that enables rapid timing closure when working with implementation tools. Conformal Constraint Designer includes the following:

o o o o o



Formal validation of SDC exceptions, such as false paths (FPs). SDC Quality checks Extensive debug and analysis, which leverages both RTL and gate-level information Ability to identify false paths from critical paths Ability to validate the consistency between top-level and lowerlevel constraints

Evaluate Throughout the flow it is important to achieve desirable timing results at each step before proceeding. The goal as you proceed through each step is to achieve similar timing results as the previous step. Timing jumps are unexpected increases in negative slack which occur during or command, or when going from one step to the next. When running optDesign, timing jumps may occur when it runs refinePlace or trialRoute. If you see in the log file the worst negative slack (WNS) increases unexpectedly, it's likely caused by one of three things:

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 29

EDI Timing Closure Guide

1. Large cell movements during refinePlace 2. Discrepancies between optDesign's internal steiner router and trialRoute (the steiner router is used after a net has been modified by optimization). 3. Differences in layer selection between the steiner router and trialRoute, or between different calls to trialRoute. You can debug this by running the following which will separate out the refinePlace and trialRoute steps from optDesign: setOptMode -allEndPoints false -rPlace false optDesign -preCTS -incr -outDir RPT -prefix prects.mixed_steiner_tr saveDesign prects.mixed_steiner_tr.enc timeDesign -preCTS -outDir RPT -prefix prects.tr_nonlegal refineplace timeDesign -preCTS -outDir RPT -prefix prects.tr_legal The initial optDesign will finish the design without calling refinePlace or trialRoute. This will be a mix of steiner and trialRoute nets so you can observe the optimized timing. Next, trialRoute is run by timeDesign. If the timing degrades a lot at this point, the timing jump is due to discrepancies between steiner route and trialRoute. You can compare the path after trialRoute to the same exact path in the mixed case to determine the net(s) causing the problem and attempt to identify what's happening. For example, is the net going over a hard macro? Lastly, refinePlace is called to legalize the placement and re-trialRoute the design with timeDesign. Compare the timing after legalization. If it has increased significantly, compare the worst path before and after refinePlace and observe how the cells are moved. Is the moved in a congested area or is a cell having to be moved a large distance to avoid a blockage?

Clock Tree Synthesis The goal of clock tree synthesis is to build a buffer distribution network to meet the timing requirements among the leaf pins. It includes the following jobs. • • •

Creating clock tree spec file Building a buffer distribution network Routing clock nets using NanoRoute

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 30

EDI Timing Closure Guide

Creating the Clock Specification File Clock tree synthesis (CTS) is a series of procedures to build a buffer distribution network to meet the design's timing targets. The clock tree specification file is used to direct clock tree synthesis and includes: • • •



Design constraints including latency, skew, and design rules Buffer and routing type definitions Trace and synthesis controls like: MacroModel, ClkGroup, NoGating, LeafPin, ExcludedPin, PreservePin, ThroughPin, and GatingGroupInstances Flow controls like: o Whether or not to generate a detail report o Whether or not to route the clock net o Whether or not to perform post-CTS optimization

You can generate the default clock tree spec file with the command: createClockTreeSpec -file filename Automatically generating a clock tree specification translates the following information from the timing constraint file into suitable records for the clock tree spec file. • • • • • •

create_clock - Becomes AutoCTSRootPin in CTS constraints file set_clock_transition - Becomes SinkLeafTran and BufMaxTran (Default: 400 ps) set_clock_latency value - Becomes MaxDelay (Default: clock period) MinDelay (Default: 0) set_clock_latency -source value - Becomes SrcLatency value in ns set_clock_uncertainty - Becomes MaxSkew (Default: 300 ps) create_generated_clock - Will add necessary ThroughPin statement to the CTS constraints file

You can control the buffer types to build clock trees by listing the buffer types in the config file or use "createClockTreeSpec -bufferList bufferList". You can also control the routing types for the clock nets. Route types for nets connected to leaf cells and nets connected to non-leaf cells can be specified separately with LeafRouteType and Routetype. Also, you can use setCTSMode before running specifyClockTree to change the default routing type and global clock tree synthesis controls. Note: Set the preferredExtraSpace to 0 in the LeafRouteType definition in the spec file. If you do not set it to zero, the software uses the value COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 31

EDI Timing Closure Guide

specified in RouteType definition, which might be greater than zero. This can cause congestion. The clock tree specification file is very important and directly affects the result of clock tree synthesis. A good clock tree plan including suitable constraints and placement space can improve the results of clock tree synthesis and avoid problems for post-CTS timing closure. The Pre-CTS Clock Tree Tracer (Clock - Trace Pre-CTS Clock Tree) user interface can be used to traverse the clock tree structure logically and physically based on the applied clock specification file before committing clock tree synthesis. You can use it as a basis for changing the clock tree specification file to consolidate the clock tree structure and improve the results of clock tree synthesis. Synthesizing the Clock Tree To generate the clock tree, use the clockDesign command. This command performs the following operations during clock tree synthesis: • • • •

Deletes any existing buffers on the clock nets Builds a buffer distribution network to distribute the clock signal(s) to the registers Routes the clock nets using NanoRoute Optimizes the clock tree

clockDesign is a super-command which runs the commands in the CTS flow (i.e. createClockTreeSpec, specifyClockTree, deleteClockTree, ckSynthesis, etc.) It's important to note clockDesign automatically sets some CTS options which are disabled by default. So if you are comparing a clockDesign run to a run where each command is run separately, ensure the settings are consistent. The table below compares the clockDesign settings to the default settings: Option clockDesign Setting Default Setting RouteClkNet Yes No PostOpt Yes Yes OptAddBuffer Yes No The clockDesign command generates the default clock tree specification file (if not specified), deletes existing clock trees, builds the clock tree, calls NanoRoute to route the clock nets, and then optimizes the clock tree to

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 32

EDI Timing Closure Guide

improve the skew including resizing buffers or inverters, adding buffers, refining placement, and correcting routing. If you performed useful skew optimization (setOptMode -usefulSkew true), clockDesign automatically checks for any scheduling file in the working directory, or checks for "rda_Input ui_scheduling_file", and honors the scheduling file while building the clock tree. If the clockDesign command calls NanoRoute to route the clock nets, direct NanoRoute to follow the route guide by using the command "setCTSMode routeGuide true". This is enabled by default. This operation can improve the correlation between pre-route and post-route clock nets. Tips for Performing CTS on High Performance Designs Skew, slew and latency are the primary methods to get the highest performance design but there is a balance between tightening these constraints and getting the best clock tree (tighter values often increase the area/power of the tree). Reducing latency often lessens the impact of derating, but sometimes can come at the expense of common paths which get filtered via CPPR. Mode settings to reduce latency: •

The following affects actual tree construction. It can increase runtime considerably and ignores MinDelay constructs: setCTSMode -synthLatencyEffort high



The following performs optimization after the tree construction mostly by optimizing the location of the tree elements and can also add significant runtime: setCTSMode -optLatency true



The following reduces the size of the tree by performing optimization after the tree construction to delete and downsize elements to recover area: setCTSMode -optArea true

Routing layer selection •

It is advised to have CTS route the clock nets and fix the wires.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 33

EDI Timing Closure Guide







Using the upper thicker layers often improves performance due to the lower lateral capacitance. o Try to pair and limit the layers (i.e. horizontal and vertical) with the same pitch/width/spacing so CTS estimation is more accurate. Using a preferred extra spacing of 1 helps to reduce capacitance and future SI impact. o Using values larger than 1 often do not improve the situation and often have no effect on SI pushouts. That is because once the wires are >1 track away from the clock, the primary coupling occurs between the cross over/under wires. Routing with wider widths (2 to 3 times the default width) is an effective way to reduce resistance which is a big factor for designs at 40nm and below.

Buffer/Inverter selection •

Limiting the list of available cells to 3 or 4 often improves both runtime and quality of results. The low drive clock cells typically shouldn't be used as they are susceptible to routing changes. If the tree can be built with LVT cells, use those as the tree will be smaller and more SI immune. Additionally, you can run tests to see if trees built with all buffers or all inverters are faster.

Setting constraints •

Using MaxCap constraints on elements in the tree can help reduce the potential for large jumps through these cells. For example: MaxCap + SGCLATNX4 0.030pF + CLKINVX16 0.20pF



Apply a realistic minDelay, maxDelay and maxSkew value because it will help to balance the different clocks from each other.



Make sure all clock gating cells are not marked dont_use or dont_touch otherwise they will not be resized. o It is very common for the default library definition to have them marked dont_use and/or dont_touch.

Manual skewing COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 34

EDI Timing Closure Guide



Sometimes certain elements must be manually skewed. This can arise when preCTS useful skew is not enabled or preCTS cannot predict the magnitude of the problem due to skew/derating. In preCTS you can model this using the set_clock_latency SDC construct. The following example show the clock delay to A/B/RAM1/CLKA is pulled in 500ps: set_clock_latency -0.5 A/B/RAM1/CLKA



To model this in the CTS spec file it would appear as follows. The +0.5ns means that 500ps of latency is "inside" the CLKA pin of A/B/RAM1: MacroModel pin A/B/RAM1/CLKA 0.5ns 0.5ns 0.5ns 0.5ns 0pF

Tips for Performing CTS on Congested and High Utilization Designs CTS congestion normally results from either: • • •

Too many cells inserted due to overly tight constraints Poor choice of top/bottom preferred routing layers o For example, restricting the top layer so it cannot go over RAMS NDRs or other routing rules

For CTS in high utilization designs, typically the goal is to make the tree as small as possible. The following setting is used to reduce the size of the tree. CTS performs optimization after the tree construction to delete and downsize elements to recover area. Also, when too many cells are inserted, try relaxing the constraints (typically the Buf/Sink MaxTran): setCTSMode -optArea true When routing rules are causing the problems, consider only using the rules for the non-sink levels (or using a less restrictive rule for the sinks) •

In the CTS spec file: o RouteType controls the routing rules for non-sink levels o LeafRouteType controls the routing rules for the sink level

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 35

EDI Timing Closure Guide



MaxTran constraints typically have the largest effect on size of tree so relaxing these helps reduce impact.

Buffer/Inverter selection •

If the tree can be built with LVT cells, use those as the tree will be smaller and more SI immune.

Analyzing and Debugging the Clock Tree Results You can use the Clock Tree Browser (Clock - Browse Clock Tree) user interface to fine tune the clock tree to improve the results. From the user interface you can perform the following operations: • • • •

Add buffers Delete buffers Size cells Change net connections

Use Global Clock Tree Debug (Clock - Debug Clock Tree) to debug the timing result. Refer to the chapter, Synthesizing Clock Trees, in the EDI System User Guide for more information. Also, see the chapter, Clock Menu, in the Encounter Menu Reference for descriptions of the forms and fields of the user interface. Sometimes a degradation in clock delay or skew occurs during CTS when comparing the results before and after the clocks are routed. If this occurs try the following: • •



Confirm the RC scaling factors for the clocks are set properly. See How to Generate Scaling Factors for RC Correlation. Constrain the routing to two upper routing layers using a RouteType in the CTS specification file. Constraining the routing to two layers reduces differences in layer assignment between CTS and NanoRoute. Use displayClockMinMaxPaths with the -preRoute and -clkRouteOnly options to compare pre-route and clock route paths.

Optimizing the Clock Tree After clockDesign, ckECO can be used to improve the tree based on the parasitics and timing seen by the optimizer. •

ckECO by default can use all the allowed buffers/inverters. To limit it to only those in the CTS spec file use the -useSpecFileCellsOnly option.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 36

EDI Timing Closure Guide

ckECO -postCTS –useSpecFileCellsOnly



A similar flow can be used after detailed routing. Be aware that if useful skew was applied during post-CTS optimization, "ckECO postRoute" may undo this because its goal is to minimize skew.: ckECO -postRoute [-useSpecFileCellsOnly]



If you are looking for local skew reduction (skew between talking flipflops) use the -localSkew option: ckECO -postCTS -useSpecFileCellsOnly –localSkew



Check the CTS log file for clock gating element movement during optDesign -postCTS: o Use setPlaceMode "-clockGateAware true" option in placement (see placeDesign section) o Or don't allow gated elements to move during CTS: setCTSMode -optLatencyMoveGate false

Clock Specification File Example The following example shows a clock specification file. # # FirstEncounter(TM) Clock Synthesis Technology File Format # MacroModel pin freg/mod004048/CLK 20ps 18ps 20ps 18ps 30ff ClkGroup + CGEN_1 + CGEN_2 RouteTypeName CK1 PreferredExtraSpace 1 TopPreferredLayer 4

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 37

EDI Timing Closure Guide

BottomPreferredLayer 3 End RouteTypeName LF1 PreferredExtraSpace 0 TopPreferredLayer 4 BottomPreferredLayer 3 End AutoCTSRootPin cgen/i_5/Y MaxDelay 5.0ns MinDelay 0ns MaxFanout 2 MaxSkew 250ps SinkMaxTran 550ps BufMaxTran 550ps NoGating NO DetailReport YES Obstruction YES useCTSRouteGuide YES RouteType CK1 LeafRouteType LF1 RouteClkNet YES PostOpt YES OptAddBuffer YES OptAddBufferLimit 100 Buffer BUFX4 BUFX8 BUFX12 INVX1 MaxCap + BUFX4 1pf + BUFX8 1pf + BUFX12 1pf ThroughPin + df/mod000446/CK ThroughPort + df/mod002300/ax2 LeafPin + PCLK66_gate_i/A rising LeafPort + ssfd2s/D rising PreservePin + cgen/mod000043/A ExcludedPin + freg/mod004048/CLK ExcludedPort COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 38

EDI Timing Closure Guide

+ DFF_B/CLK End

Post-CTS Optimization The goals of post-CTS optimization include: • • •

Fixing remaining design rule violations Optimizing remaining setup violations Correcting timing with propagated clocks

At this point in the design flow, the clocks are inserted and preferably routed. Since timing analysis uses the actual clock delays, you should adjust the timing constraints as follows. Typically, users have separate SDC files for pre-CTS and post-CTS timing analysis. Use the update_constraint_mode command to update the SDC files for each operating mode. •

Set the clocks to propagated by adding the following to your SDC file(s): set_propagated_clock [all_clocks]



Adjust the clock uncertainty (set_clock_uncertainty SDC) to model only jitter. You need to update the constraints after clock tree synthesis to adjust clock jitter according to the design. Modeling only the jitter avoids making the timing appear worse than it is. Remember that, since the actual clock skew data is now available, it is possible that critical path timing will be worse.



Remove or change the SDC constraints that are not valid postCTS like clock_uncertainty or clock_latency. You may need to adjust the clock latencies on the IOs so IO timing does not become the critical path by either adjusting the virtual clock source latencies: set_clock_latency -source xxx Or adjust IO constraints directly (set_input_delay/set_output_delay) or use the update_io_latency command.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 39

EDI Timing Closure Guide



Adjust

derating

applied

per

each

delay

corner.



Make sure RC scaling factors are tuned properly. o Clock routing will use a different scale factor than the signal nets.  Signal nets use -preRoute_cap  Clock nets use -postRoute_clkcap (if defined) or postRoute_cap. Run timing analysis to check the timing: timeDesign -postCTS -outDir ctsTimingReports If the timing is not similar to pre-CTS timing results check the following:



Was CTS able to achieve the skew constraints? Are these skew constraints within the set_clock_uncertainty set during preCTS timing?



Was clock uncertainty adjusted after CTS to only model jitter?

Post-CTS Optimization Command Sequences Typically the same options applied during preCTS optimization are used for postCTS optimization. •

Use

the

timeDesign

command

to

check

the

post-CTS

timing:

timeDesign -postCTS -outDir postctsTimingReports The timing at this point should be similar to the pre-CTS timing results. If there is a large jump in negative slack, investigate whether the end points are latches. A jump in negative slack may occur for latches now that their real skew is used for timing analysis. If the large slack is occurring at other points used GTD to analyze the paths. Double-check the clocks are propagated and that the skew is within your specifications.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 40

EDI Timing Closure Guide



To optimize timing after the clock tree has been built, use the following commands: optDesign -postCTS -outDir postctsOptTimingReports



To take advantage of useful skew when optimizing timing in post-CTS mode, use the following commands: setOptMode -usefulSkew true optDesign -postCTS [-incr] If you have already performed detail routing on the clock tree, EDI System performs global and detailed ECO routing automatically using NanoRoute in post-CTS useful skew mode. If you do not want EDI System to do this, specify the -noECORoute parameter as follows: setOptMode -usefulSkew true optDesign -postCTS -noECORoute [-incr] If you specify -noECORoute before running optimization, EDI System performs trial routing to estimate clock delays.



If clock gating cells are on the critical path you can enable clock gate cloning during post-CTS optimization: setOptMode -postCtsClkGateCloning true

Hold Optimization At this point run timing analysis to report hold violations: timeDesign -postCTS -hold -outDir postctsHoldTimingReports Timing optimization to fix hold violations can be performed at this point. This is recommended if your design has a significant number of hold violations because they are easier to fix prior to routing. If the number of hold violation is low you can wait until after routing the signal nets. COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 41

EDI Timing Closure Guide

To perform hold optimization: optDesign -postCTS -hold -outDir postctsOptHoldTimingReports Tips for Performing Hold Optimization on High Peformance Designs



Try to avoid providing very weak buffers for Hold fixing because they are more sensitive to routing detour and SI.



For Multi-Vth design, make sure to run leakage optimization (through optDesign or optLeakagePower) before running Hold fixing since leakage reduction improves Hold timing.



For a design with many hold violated paths, it is highly recommended to run hold fixing at postCts stage already (although there is no need to achieve 0ns slack at this stage). Then, use postRoute hold fixing to fix the remaining violations. You can use a negative hold target slack to focus hold fixing on the paths with large violations and fix the remaining hold violations after routing. For example, the following sets a hold target slack of -200ps: setOptMode -holdTargetSlack -0.2



By default Hold fixing can degrade setup TNS (but not Setup WNS). This can be changed through: setOptMode -fixHoldAllowSetupTnsDegrade true|false



To exclude some path_group/clockDomains from hold fixing you can apply: setOptMode -ignorePathGroupsForHold {groupA groupB ...}

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 42

EDI Timing Closure Guide



Things to watch out for: o Make sure that the clock trees are well balanced (inter and intra clock trees). o Avoid over constraining the clock uncertainty. o Reduce local placement congestions. Adding cell padding may sometime help. o If you want to allow Setup WNS degradation, you should set a negative setup target slack (setOptMode -setupTargetSlack )



In Verbose mode (setOptMode -verbose true), Hold fixing will print a detailed report where each net not buffered will be sorted through several categories. o This will allow the user to identify what are the most critical issues to resolve. o It will help the user in understanding why a given net was not buffered. o The different categories listed in the log file are : *info: XXX net(s): Could not be location' *info: XXX net(s): Could not be "degrade setup reg2reg WNS". *info: XXX net(s): Could not be "degrade setup TNS" *info: XXX net(s): Could not be "degrade setup WNS". *info: XXX net(s): Could not be "Degrade Hold". *info: XXX net(s): Could not be "Degrade max_cap violations". *info: XXX net(s): Could not be "Degrade max_tran violations". *info: XXX net(s): Could not be "Violating DRV". *info: XXX net(s): Could not be dont_touch. *info: XXX net(s): Could not be as badly routed. *info: XXX net(s): Could not be Driver nets. *info: XXX net(s): Could not be nets. *info: XXX net(s): Could not be

fixed because of 'no legal fixed because they would fixed because they would fixed because they would fixed because they would further fixed because they would fixed because they would fixed because they already have fixed because they are set as fixed because they are flagged fixed because they are Multifixed because they are clock fixed because no Always-On-

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 43

EDI Timing Closure Guide

Buffers are usable. *info: XXX net(s): Could not be fixed because "no valid node" were found on net. *info: XXX net(s): Could not be fixed because of "internal failure". •

"setOptMode -fixHoldAllowOverlap" controls if hold fixing is limited to purely legal moves (no overlaps). When set to "true" hold optimization allows initial cell insertion to overlap cells and then refinePlace legalizes the cells placement. This provides optimization more opportunity to fix violations. When set to the default value of "auto" hold optimization is allowed to create overlaps during post-CTS optimization but not during post-route optimization. To allow overlaps during post-route optimization as well set the following: setOptMode -fixHoldAllowOverlap true This can be also used post-route but should not be used with SI optimization because the refinePlace call may not be made.

Tips for Performing Hold Optimization on Congested and High Utilization Designs •

In a multi-Vth library scenario, make sure that Leakage Optimization was run before Hold fixing.



Hold fixing is allowed to overlap cells during post-CTS hold optimization but not during post-route optimization. To enable overlaps during post-route optimization as well set the following: setOptMode -fixHoldAllowOverlap true



Control hold time margins o Beyond certain hold margin, delay buffer addition rate increases exponentially o Try useful skew optimization for RAM and Register files

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 44

EDI Timing Closure Guide

Checking Timing You can use the following commands to report setup and hold time violations after post-CTS optimization. timeDesign -postCTS -outDir postctsOptTimingReports timeDesign -postCTS -hold -outDir postctsOptTimingReports

Detailed Routing After post-CTS optimization, there should be few, if any, timing violations left in the design. The goals of detailed routing include: • •

Routing the design without DRC or LVS violations. NanoRoute performs a DRC and cleans up violations. Routing the design without degrading timing or creating signal integrity violations.

NanoRoute can perform timing-driven and SI-driven routing concurrently. NanoRoute routes the signals which are critical for signal integrity appropriately to minimize cross-coupling between these nets which would lead to post-route signal integrity issues. Improving Timing during Routing The following tips can help achieve better timing results during the routing phase of the design. •

• •

• • •

Make sure the LEF file contains a sufficient number and variety of vias (hammer-head, stacked, and so forth). o Check with your library provided or foundry for the latest technology LEF to use. Check the definition of tracks in the DEF file. If the tracks are poorly defined, regenerate tracks with the generateTracks command. If timing is way off and/or there is local or global congestion, return to post-CTS optimization and optimize further or run non-timing-driven routing. Unfix the clock nets before using NanoRoute Make sure the top max routing later is set appropriately. Specify required NonDefaultRules (NDRs) and/or shield routing.

Routing Command Sequence The following example shows the use of NanoRoute to do detailed routing. COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 45

EDI Timing Closure Guide

changeUseClockNetStatus -noFixedNetWires routeDesign Note: routeDesign automatically sets "setNanoRouteMode –routeWithTimingDriven true -routeWithSiDriven true". If you are using globalDetailRoute in place of routeDesign make sure you manually set these options to true. Tips for Routing High Performance Designs •

For many RC corners at 40nm and below, vias are extremely expensive for resistance. Achieving the highest possible double cut coverage helps reduce resistance for vias that cannot be eliminated: setNanoRouteMode –routeConcurrentMinimizeViaCountEffort medium setNanoRouteMode -drouteUseMultiCutViaEffort medium | high



When using routeDesign, by default timing driven and SI driven are enabled. High effort SI routing can be enforced using: setNanoRouteMode -routeSiEffort high



If filler cells containing metal obstruction other than the followpins are to be used, make sure they are inserted prior to the initial route.



Post route wire spreading significantly reduces SI impact. o After routeDesign, use non-timing driven wire spread to spread all potential wires (for best SI results). setNanoRouteMode -routeWithTimingDriven false setNanoRouteMode -droutePostRouteSpreadWire true routeDesign -wireOpt setNanoRouteMode -droutePostRouteSpreadWire false

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 46

EDI Timing Closure Guide

Tips for Routing Congested Designs



The following will help the routing with some runtime impact and setup timing degradation as NanoRoute will detour some of the nets to ease the congestion: setNanoRouteMode -grouteMinimizeCongestion true



For 65nm and below, we recommend cover blockage over MACRO's (RAM/ROM) rather than pin cut-outs from OBS (blockage).

PostRoute Extraction All nets are now routed. It is important to now set the extraction mode to post-route and specify the extractor to use. •

Specify the engine to be postRoute: setExtractRCMode -engine postRoute



Set -effortLevel so extractRC uses your desired extractor: setExtractRCMode -effortLevel low|medium|high o

o

o

low - Invokes the native detailed extraction engine. This is the same as specifying the "-engine postRoute" setting. medium - Invokes the Turbo QRC (TQRC) extraction mode. TQRC performance and accuracy falls between native detailed extraction and IQRC engine. This engine supports distributed processing.  TQRC engine is recommended for process nodes < 65nm. Note: This setting does not require a QRC license. high - Invokes the Integrated QRC (IQRC) extraction engine.  IQRC provides superior accuracy compared to TQRC. IQRC is recommended for extraction after ECO. In addition,

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 47

EDI Timing Closure Guide

IQRC supports distributed processing. Note: IQRC requires a QRC license. Checking Timing Use the following command to do a post-route timing check: timeDesign -postRoute -outDir postrouteTimingReports timeDesign -postRoute -hold -outDir postrouteTimingReports If timing jumps at this point compared to timing before routing check the following:

• •

Are the post-route RC scaling factors set properly? Is the routing topology similar between trialRoute and NanoRoute? Compare the same paths between post-CTS and post-Route databases and for large differences in loads.

Post-Route Optimization During post-route optimization, there should be very few violations that need correction. The primary sources of these timing violations include: • •

Inaccurate prediction of the routing topology during optimization due to congestion-based detour routing Incremental delays due to parasitics coupling

pre-route

Since the violations at this stage are due to inaccurate modeling of the final route topology and the attendant parasitics, it is critical at this point not to introduce any additional topology changes beyond those needed to fix the existing violations. Making unnecessary changes to the routing at this point can lead to a scenario where fixing one violation leads to the creation of others. This cascading effect creates a situation where it becomes impossible to close on a final timing solution with no design rule violations. One of the strengths of post-route optimization is the ability to simultaneously cut a wire and insert buffers, create the new RC graph at the corresponding point, and modify the graph to estimate the new parasitics for the cut wire without re-doing extraction. In addition to the timing violations caused by inaccurate route topology modeling, the parasitics cross-coupling of neighboring nets can cause the following problems that need to be addressed in high speed designs: COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 48

EDI Timing Closure Guide

• •

An increase or decrease in incremental delay on a net due to the coupling of its neighbors and their switching activity. Glitches (voltage spikes) that can be caused in one signal route by the switching of a neighbor resulting in a logic malfunction.

These effects need to be analyzed and corrected before a design is completed. They are magnified in designs with small geometries and in designs with high clock speeds. Data Preparation SI optimization requires certain preparation: • •

• • •

Make sure cdB libraries are specified for each cell for each delay corner. You must be in On Chip Variation (OCV) mode to see simultaneous clock pushout / pullin. Enable this using "setAnalysisMode analysisType onChipVariation" Confirm your CeltIC settings match your signoff tool Confirm coupling filter settings match your signoff tool Enable SI CPPR through "set_global timing_enable_si_cppr true" (this is the default)

Additionally, consider the following techniques if you have difficulty achieving signal integrity closure on your design. •







Watch for routing congestion during floorplanning and especially after detailed routing. o Consider running the congOpt command on the design to eliminate local hot spots or adjust your floorplan Use NanoRoute advanced timing with SI-driven routing options during detailed routing. For example: o setNanoRouteMode -routeWithTimingDriven true o setNanoRouteMode -routeWithSiDriven true Fix transition time violations. o Slow transitions introduce a larger delay penalty or incremental delay. Prepare all of the required noise models. o Highly-accurate CeltIC analysis requires the use of accurate noise models like cdB, ECHO, or xILM. The use of blackbox (missing) models could lead to a significant number of false violations.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 49

EDI Timing Closure Guide

You must recalibrate the noise library with each release of CeltIC. New features may not function properly when used with an old .cdB file. Characterize your memory and analog block using the make_cdb utility. Create XILM models for sub blocks to model their noise sensitivity at the chip level. 

o o

Post-Route Optimization Command Sequences The Advanced Analysis Engine (AAE) is a new unified delay calculation engine that simultaneously computes base and SI delays. AAE can perform incremental SI delay computation during timing optimization; hence provide accurate feedback to optimization engine. AAE also capable of multithreading thus gives substantial runtime gains in timing closure. Unlike the default flow, AAE based optimization does not require four steps optimization: Default: optDesign optDesign optDesign optDesign

-postRoute -postRoute -hold -postRoute -si -hold -postRoute -si

AAE (recommended): optDesign -postRoute optDesign -postRoute -hold To run AAE based SI fixing: setAnalysisMode -analysisType onChipVariation -cppr both setDelayCalMode -engine default -SIAware true optDesign -postRoute optDesign -postRoute -hold For more information on the AAE flow see the application note: PostRoute Optimization Using the Advanced Analysis Engine (AAE). Tips for Performing Post-Route Optimization High Performance Designs

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 50

EDI Timing Closure Guide



For multi-VT designs, after routing there can be large WNS/TNS jumps due to the sensitivity of HVT cells o runtime of postRoute optimization can be significantly impacted by large TNS o Using optLeakagePower -fixTimingOnly prior to optDesign postRoute can massively speed up closure  Only works for multi-VT  Does not have much impact when designs has already mainly LVT cells.



The registers by default are marked + FIXED by clockDesign and cannot be resized eventhough postRoute optimization can resize flops. To temporarily set them as + PLACED during optDesign, you can apply: setOptMode -unfixClkInstForOpt true



The local density by default is limited to 98%. If optimization is prematurely exiting due to local placement hotspots this can be increased to 100% using setOptMode -maxLocalDensity 1.0



Make sure you are in OCV analysis (MMMC setup required) to calculate pushin/pullout timing properly: setAnalysisMode -analysisType onChipVariation Note if the design is not in OCV mode, clock SI pushout/pullin will not be correct.



Make sure your extraction filters correlate to your signoff extraction o when using iQRC the filters typically can be set to the exact signoff values  make sure to use setExtractRCMode -capFilterMode relAndCoup  StarRC has no total_c_th equivalent so set this to 0 for these flows For postRoute optimization, you may want to force the tool

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 51

EDI Timing Closure Guide

to only allow legal resizing to avoid placement legalization and thus limit routing changes: setOptMode -postRouteAllowOverlap false Checking Timing AAE is a fast SI engine so it does not perform circuit simulations to compute SI delays. Being a fast SI engine, it does not completely correlate with Celtic (signoff SI engine). AAE is tuned well to work with optimization engine in order to provide better QOR. So there is a tendency to be pessimistic compared to Celtic. AAE being pessimistic compared to Celtic means it does not need 'setSIMode -analysisType pessimistic' setting for Primetime correlation. Please use 'setSIMode -analysisType default' in AAE and Celtic flows for better correlation. Optimization final summary is AAE estimation based report only. So if you want to check your QOR, you need to run signalStorm+Celtic based timeDesign between setup and hold optimization or at the end of hold optimization. For example: timeDesign -postRoute -si -outDir final_setup timeDesign -postRoute -si -hold -outDir final_hold If you are using Encounter Timing System (ETS) for signoff timing make sure the CeltIC settings are consistent between EDI System and ETS. If you are using PrimeTime-SI for signoff set "setSIMode -analysisType pessimistic" prior to timeDesign for better correlation. Optimizing With 3rd Party SPEF or SDF If you are using a 3rd Party tool for extraction or timing analysis the most important step is to make sure they correlate with EDI System's corresponding function. For example, if you are using a 3rd Party extractor make sure the RC scaling factors are set properly within EDI System. If you are using a 3rd Party timing analysis tool, make sure it correlates with EDI System by verifying SDC constraints are applied consistently between the tools. The delay calculation and timing analysis results between EDI System and the 3rd Party tool should also correlate.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 52

EDI Timing Closure Guide

If timing violations occur when using SPEF from a 3rd Party extractor you can import the SPEF into EDI System and perform optimization. You must import a SPEF for each RC corner. The flow is: spefIn rc_corner1.spef -rc_corner rc_corner1 spefIn rc_corner2.spef -rc_corner rc_corner2 ... optDesign -postRoute [-si] [-hold] -outDir spefFlowTimingReports





The -si and -hold options are optional in the above command. o Use -hold if you need to perform hold fixing based on the the SPEF. o Use -si if you need to perform SI fixing based on the SPEF. optDesign will use the SPEF for initial timing to determine the best location to optimize the paths.

optDesign can also optimize based on SDF from a 3rd Party timing analyzer. This is useful when the timing analyzer reports timing violations EDI System did not. The flow to optimize based on SDF is: read_sdf -view view1 sdf1.sdf read_sdf -view view2 sdf2.sdf optDesign -postRoute -useSDF [-hold] -outDir sdfFlowTimingReports





Note in order to use SDF information in MMMC analysis mode, you should provide an SDF file for each active view in the design. If a view was not given an SDF file, EDI System will run detail extraction and delay calculation for that specific view. optDesign will run extraction to calculate slew values because SDF doesn't contain slews. The delays in the SDF will be used, but the slews must be generated. You can also spefIn to avoid the extraction.

Timing Sign Off The goal of timing sign off is to verify that the design meets the specified timing constraints. This is accomplished by first using QRC to generate detailed extraction data and then using the specified timing analysis engine for a final analysis of setup and hold data.

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 53

EDI Timing Closure Guide

At this point in the design process, final routing and post-route optimization is complete. The following command sequence generates the reports needed to verify timing: timeDesign -signoff -outDir signOffTimingReports timeDesign -signoff -hold -reportOnly -outDir signOffTimingReports These commands perform the following operations: 1. Runs QRC to generate detailed parasitics. 2. Uses the detailed parasitics and generates the setup timing reports. 3. Generates the hold timing reports. For QRC to run, you need to ensure that the EDI System configuration file includes a technology file and a LibGen cell library database created with the runLibGen command. (?? To Do?) For more information about setting up the libraries for QRC, refer to the chapter RC Extraction in the EDI System User Guide. If the timing degrades compared to post-route timing check the following:

• •

Were the post-route RC scaling factors properly set which correlate to the signoff extractor? See the Application Note "EDI to ETS Correlation: Guidelines on How to Debug and Correlate Timing and SI results" at the end of this document.

If you are using a 3rd party extractor use spefIn to read the SPEF for each RC corner then run timeDesign using the -reportOnly option: spefIn rc_corner1.spef -rc_corner rc_corner1 spefIn rc_corner2.spef -rc_corner rc_corner2 ... timeDesign -signoff -reportOnly -outDir signOffTimingReports timeDesign -signoff -hold -reportOnly -outDir signOffTimingReports

Additional Resources Following are additional application notes, training and documentation related to timing closure. These can be found at http://support.cadence.com: COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 54

EDI Timing Closure Guide



NanoRoute Recommended Options with emphasis on 32nm and below advance node/technology (EDI)



Encounter Digital Implementation System Foundation Flows Guide



How to Generate Scaling Factors for RC Correlation



PostRoute Optimization Using the Advanced Analysis Engine (AAE)



Guide to Clock Tree Synthesis (CTS) in a block or flat chip using EDI system



EDI to ETS Correlation - Guidelines on How to Debug and Correlate Timing and SI results

COPYRIGHT © 2011, CADENCE DESIGN SYSTEMS, INC. ALL RIGHTS RESERVED.

PAGE 55

View more...

Comments

Copyright © 2017 KUPDF Inc.
SUPPORT KUPDF