Intel® Quartus® Prime Pro Edition User Guide: Timing Analyzer

Download PDF
ID 683243
Date 10/04/2021
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents
Document Table of Contents x
1. Timing Analysis Introduction 2. Using the Intel® Quartus® Prime Timing Analyzer A. Intel® Quartus® Prime Pro Edition User Guides
1. Timing Analysis Introduction x
1.1. Timing Analysis Basic Concepts 1.2. Timing Analysis Overview Document Revision History
1.1. Timing Analysis Basic Concepts x
1.1.1. Timing Path and Clock Analysis 1.1.2. Clock Setup Analysis 1.1.3. Clock Hold Analysis 1.1.4. Recovery and Removal Analysis 1.1.5. Multicycle Path Analysis 1.1.6. Metastability Analysis 1.1.7. Timing Pessimism 1.1.8. Clock-As-Data Analysis 1.1.9. Multicorner Timing Analysis 1.1.10. Time Borrowing
1.1.1. Timing Path and Clock Analysis x
1.1.1.1. The Timing Netlist 1.1.1.2. Timing Paths 1.1.1.3. Data and Clock Arrival Times 1.1.1.4. Launch and Latch Edges
1.1.5. Multicycle Path Analysis x
1.1.5.1. Multicycle Clock Hold 1.1.5.2. Multicycle Clock Setup
1.1.10. Time Borrowing x
1.1.10.1. Time Borrowing Limitations 1.1.10.2. Time Borrowing with Latches
2. Using the Intel® Quartus® Prime Timing Analyzer x
2.1. Timing Analysis Flow 2.2. Step 1: Specify Timing Analyzer Settings 2.3. Step 2: Specify Timing Constraints 2.4. Step 3: Run the Timing Analyzer 2.5. Step 4: Analyze Timing Reports 2.6. Applying Timing Constraints 2.7. Timing Analyzer Tcl Commands 2.8. Timing Analysis of Imported Compilation Results 2.9. Using the Intel® Quartus® Prime Timing Analyzer Document Revision History 2.10. Intel® Quartus® Prime Pro Edition User Guide: Timing Analyzer Archive
2.4. Step 3: Run the Timing Analyzer x
2.4.1. Setting the Operating Conditions 2.4.2. Enabling Time Borrowing Optimization
2.5. Step 4: Analyze Timing Reports x
2.5.1. Generating Timing Reports 2.5.2. Cross-Probing with Design Assistant 2.5.3. Launching Design Assistant from Timing Analyzer 2.5.4. Locating Timing Paths in Other Tools 2.5.5. Correlating Constraints to the Timing Report
2.5.1. Generating Timing Reports x
2.5.1.1. Report Fmax Summary 2.5.1.2. Report Timing 2.5.1.3. Report Data Delay 2.5.1.4. Report Clocks and Clock Networks 2.5.1.5. Report Clock Transfers 2.5.1.6. Report Logic Depth 2.5.1.7. Report Neighbor Paths 2.5.1.8. Report Register Spread 2.5.1.9. Report Route Net of Interest 2.5.1.10. Report Retiming Restrictions 2.5.1.11. Report Reset Statistics 2.5.1.12. Report Pipelining Information 2.5.1.13. Report Asynchronous CDC 2.5.1.14. Report CDC Viewer 2.5.1.15. Report Time Borrowing Data 2.5.1.16. Report Exceptions and Exceptions Reachability
2.5.1.8. Report Register Spread x
2.5.1.8.1. Registers with High Timing Path Endpoint Tension 2.5.1.8.2. Registers with High Immediate Fan-Out Tension
2.5.2. Cross-Probing with Design Assistant x
2.5.2.1. Cross-Probing from Design Assistant to Timing Analyzer
2.6. Applying Timing Constraints x
2.6.1. Recommended Initial SDC Constraints 2.6.2. SDC File Precedence 2.6.3. Modifying Iterative Constraints 2.6.4. Using Entity-bound SDC Files 2.6.5. Creating Clocks and Clock Constraints 2.6.6. Creating I/O Constraints 2.6.7. Creating Delay and Skew Constraints 2.6.8. Creating Timing Exceptions 2.6.9. Using Fitter Overconstraints 2.6.10. Example Circuit and SDC File
2.6.1. Recommended Initial SDC Constraints x
2.6.1.1. Create Clock (create_clock) 2.6.1.2. Derive PLL Clocks (derive_pll_clocks) 2.6.1.3. Derive Clock Uncertainty (derive_clock_uncertainty) 2.6.1.4. Set Clock Groups (set_clock_groups)
2.6.4. Using Entity-bound SDC Files x
2.6.4.1. Entity-bound Constraint Scope 2.6.4.2. Entity-bound Constraint Examples
2.6.5. Creating Clocks and Clock Constraints x
2.6.5.1. Creating Base Clocks 2.6.5.2. Creating Virtual Clocks 2.6.5.3. Creating Generated Clocks (create_generated_clock) 2.6.5.4. Deriving PLL Clocks 2.6.5.5. Creating Clock Groups (set_clock_groups) 2.6.5.6. Accounting for Clock Effect Characteristics 2.6.5.7. Constraining CDC Paths
2.6.5.1. Creating Base Clocks x
2.6.5.1.1. Automatic Clock Detection and Constraint Creation
2.6.5.2. Creating Virtual Clocks x
2.6.5.2.1. Specifying I/O Interface Uncertainty 2.6.5.2.2. I/O Interface Clock Uncertainty Example
2.6.5.3. Creating Generated Clocks (create_generated_clock) x
2.6.5.3.1. Clock Divider Example (-divide_by) 2.6.5.3.2. Clock Multiplexer Example
2.6.5.5. Creating Clock Groups (set_clock_groups) x
2.6.5.5.1. Exclusive Clock Groups (-logically_exclusive or -physically_exclusive) 2.6.5.5.2. Asynchronous Clock Groups (-asynchronous) 2.6.5.5.3. set_clock_groups Constraint Tips
2.6.5.6. Accounting for Clock Effect Characteristics x
2.6.5.6.1. Set Clock Latency (set_clock_latency) 2.6.5.6.2. Clock Uncertainty
2.6.6. Creating I/O Constraints x
2.6.6.1. Input Constraints (set_input_delay) 2.6.6.2. Output Constraints (set_output_delay)
2.6.7. Creating Delay and Skew Constraints x
2.6.7.1. Advanced I/O Timing and Board Trace Model Delay 2.6.7.2. Maximum Skew (set_max_skew) 2.6.7.3. Net Delay (set_net_delay)
2.6.8. Creating Timing Exceptions x
2.6.8.1. Timing Exception Precedence 2.6.8.2. False Paths (set_false_path) 2.6.8.3. Minimum and Maximum Delays 2.6.8.4. Multicycle Paths 2.6.8.5. Multicycle Exception Examples 2.6.8.6. Delay Annotation 2.6.8.7. Constraining Design Partition Ports
2.6.8.4. Multicycle Paths x
2.6.8.4.1. Common Multicycle Applications 2.6.8.4.2. Relaxing Setup with Multicycle (set_multicyle_path) 2.6.8.4.3. Accounting for a Phase Shift (-phase)
2.6.8.5. Multicycle Exception Examples x
2.6.8.5.1. Default Multicycle Analysis 2.6.8.5.2. End Multicycle Setup = 2 and End Multicycle Hold = 0 2.6.8.5.3. End Multicycle Setup = 2 and End Multicycle Hold = 1 2.6.8.5.4. Same Frequency Clocks with Destination Clock Offset 2.6.8.5.5. Destination Clock Frequency is a Multiple of the Source Clock Frequency 2.6.8.5.6. Destination Clock Frequency is a Multiple of the Source Clock Frequency with an Offset 2.6.8.5.7. Source Clock Frequency is a Multiple of the Destination Clock Frequency 2.6.8.5.8. Source Clock Frequency is a Multiple of the Destination Clock Frequency with an Offset
2.7. Timing Analyzer Tcl Commands x
2.7.1. The quartus_sta Executable 2.7.2. Collection Commands
2.7.2. Collection Commands x
2.7.2.1. Wildcard Characters 2.7.2.2. Adding and Removing Collection Items 2.7.2.3. Query of Collections 2.7.2.4. Using the get_pins Command
1. Timing Analysis Introduction
2. Using the Intel® Quartus® Prime Timing Analyzer
A. Intel® Quartus® Prime Pro Edition User Guides

1.1.10. Time Borrowing

Time borrowing can improve performance by enabling the path ending at a time-borrowing flip-flop or latch to "borrow" time from the next path in the register pipeline. The borrowed time subtracts from the next path, resulting in the same cumulative timing. In this way, time borrowing can shift slack to more critical parts of the design. Without time borrowing, the Timing Analyzer analyzes each path independently and normally assumes exactly one clock cycle for each transfer.
Figure 34. Time Borrowing Example

Some of the flip-flops in Intel® Stratix® 10 and Intel® Arria® 10 devices allow time borrowing. The exact size of the available time borrow window depends on hardware settings. The Fitter (Finalize) stage automatically configures the appropriate borrow window for each time-borrowing flip-flop, based on hardware restrictions and the available hold slack.1

Figure 35. Time Borrowing with Various Slack Conditions and Borrow Values

Intel FPGA devices generally support only a few borrow window sizes. For example, Intel® Stratix® 10 devices support narrow, medium, and wide. Typically, groups of several flip-flops must share the same setting. The actual borrowed amount is completely flexible within a given borrow window. The Timing Analyzer calculates the borrowed amount separately for each operating condition, clock, and signal rise and fall edge. Selecting a wider borrow window reduces hold slack. The Compiler only selects wider settings if hold slack allows. Furthermore, if the Compiler determines that a narrower window is sufficient for a given group of registers (based on the optimal time borrowing solution), the Compiler uses the narrower window, even if there is sufficient hold slack for a wider window.

For a given borrow window size, the exact size of the borrow window may depend on the register input (for example, d or sclr), the edge of the incoming signal (rising or falling), the device speed grade, and operating conditions.

You can enable automatic implementation of time borrowing without making any RTL changes. Once enabled, the Fitter automatically configures the window size. The Fitter also determines the optimal time borrow amount within the available borrow window for any design registers that the Fitter places in time-borrowing flip-flops.

Proper timing analysis of designs that contain level-sensitive latches typically requires time borrowing. However, the automatic Fitter time borrowing optimizations do not apply for level-sensitive latches, as Time Borrowing with Latches describes in detail.

Related Information
1 In Partial Reconfiguration designs, additional restrictions may apply to time-borrowing window size.
Level Two Title