Quartus® Prime Pro Edition User Guide: Design Recommendations

Download PDF
ID 683082
Date 9/30/2024
Public
Document Table of Contents
Document Table of Contents x
Answers to Top FAQs 1. Recommended HDL Coding Styles 2. Recommended Design Practices 3. Managing Metastability with the Quartus® Prime Software 4. Quartus® Prime Pro Edition User Guide: Design Recommendations Archive A. Quartus® Prime Pro Edition User Guides
1. Recommended HDL Coding Styles x
1.1. Using Provided HDL Templates 1.2. Instantiating IP Cores in HDL 1.3. Inferring Multipliers and DSP Functions 1.4. Inferring Memory Functions from HDL Code 1.5. Register and Latch Coding Guidelines 1.6. General Coding Guidelines 1.7. Designing with Low-Level Primitives 1.8. Cross-Module Referencing (XMR) in HDL Code 1.9. Using force Statements in HDL Code 1.10. Recommended HDL Coding Styles Revision History
1.1. Using Provided HDL Templates x
1.1.1. Inserting HDL Code from a Provided Template
1.3. Inferring Multipliers and DSP Functions x
1.3.1. Inferring Multipliers 1.3.2. Inferring Multiply-Accumulator and Multiply-Adder Functions
1.4. Inferring Memory Functions from HDL Code x
1.4.1. Inferring RAM functions from HDL Code 1.4.2. Inferring ROM Functions from HDL Code 1.4.3. Inferring Shift Registers in HDL Code 1.4.4. Inferring FIFOs in HDL Code
1.4.1. Inferring RAM functions from HDL Code x
1.4.1.1. Use Synchronous Memory Blocks 1.4.1.2. Avoid Unsupported Reset and Control Conditions 1.4.1.3. Check Read-During-Write Behavior 1.4.1.4. Controlling RAM Inference and Implementation 1.4.1.5. Single-Clock Synchronous RAM with Old Data Read-During-Write Behavior 1.4.1.6. Single-Clock Synchronous RAM with New Data Read-During-Write Behavior 1.4.1.7. Simple Dual-Port, Dual-Clock Synchronous RAM 1.4.1.8. True Dual-Port Synchronous RAM 1.4.1.9. Mixed-Width Dual-Port RAM 1.4.1.10. RAM with Byte-Enable Signals 1.4.1.11. Specifying Initial Memory Contents at Power-Up Verilog HDL RAM with Initialized Contents Verilog HDL RAM Initialized with the readmemb Command VHDL RAM with Initialized Contents Verilog HDL Single-Clock, Simple Dual-Port RAM with Synchronous Reset Verilog HDL Single-Clock, Simple Dual-Port RAM with Asynchronous Reset
1.4.3. Inferring Shift Registers in HDL Code x
1.4.3.1. Simple Shift Register 1.4.3.2. Shift Register with Evenly Spaced Taps
1.4.4. Inferring FIFOs in HDL Code x
1.4.4.1. Dual Clock FIFO Example in Verilog HDL 1.4.4.2. Dual Clock FIFO Timing Constraints
1.5. Register and Latch Coding Guidelines x
1.5.1. Register Power-Up Values 1.5.2. Secondary Register Control Signals Such as Clear and Clock Enable 1.5.3. Latches
1.5.1. Register Power-Up Values x
1.5.1.1. Specifying a Power-Up Value
1.5.3. Latches x
1.5.3.1. Avoid Unintentional Latch Generation 1.5.3.2. Inferring Latches Correctly
1.6. General Coding Guidelines x
1.6.1. Tri-State Signals 1.6.2. Clock Multiplexing 1.6.3. Adder Trees 1.6.4. State Machine HDL Guidelines 1.6.5. Multiplexer HDL Guidelines 1.6.6. Cyclic Redundancy Check Functions 1.6.7. Comparator HDL Guidelines 1.6.8. Counter HDL Guidelines
1.6.3. Adder Trees x
1.6.3.1. Architectures with 6-Input LUTs in Adaptive Logic Modules 1.6.3.2. Change Adder Tree Styles
1.6.4. State Machine HDL Guidelines x
1.6.4.1. State Machine Processing 1.6.4.2. State Machine Power-Up 1.6.4.3. Verilog HDL State Machines 1.6.4.4. VHDL State Machines
1.6.4.3. Verilog HDL State Machines x
1.6.4.3.1. Verilog-2001 State Machine Coding Example 1.6.4.3.2. SystemVerilog State Machine Coding Example
1.6.4.4. VHDL State Machines x
1.6.4.4.1. VHDL State Machine Coding Example
1.6.5. Multiplexer HDL Guidelines x
1.6.5.1. Quartus® Prime Software Option for Multiplexer Restructuring 1.6.5.2. Multiplexer Types 1.6.5.3. Implicit Defaults in IF Statements 1.6.5.4. default or OTHERS CASE Assignment
1.6.5.2. Multiplexer Types x
1.6.5.2.1. Binary Multiplexers 1.6.5.2.2. Selector Multiplexers 1.6.5.2.3. Priority Multiplexers
1.6.6. Cyclic Redundancy Check Functions x
1.6.6.1. If Performance is Important, Optimize for Speed 1.6.6.2. Use Separate CRC Blocks Instead of Cascaded Stages 1.6.6.3. Use Separate CRC Blocks Instead of Allowing Blocks to Merge 1.6.6.4. Take Advantage of Latency if Available 1.6.6.5. Save Power by Disabling CRC Blocks When Not in Use 1.6.6.6. Initialize the Device with the Synchronous Load (sload) Signal
2. Recommended Design Practices x
2.1. Following Synchronous FPGA Design Practices 2.2. HDL Design Guidelines 2.3. Use Clock and Register-Control Architectural Features 2.4. Implementing Embedded RAM 2.5. Design Assistant Design Rule Checking 2.6. Recommended Design Practices Revision History
2.1. Following Synchronous FPGA Design Practices x
2.1.1. Implementing Synchronous Designs 2.1.2. Asynchronous Design Hazards
2.2. HDL Design Guidelines x
2.2.1. Considerations for the Hyperflex® FPGA Architecture 2.2.2. Optimizing Combinational Logic 2.2.3. Optimizing Clocking Schemes 2.2.4. Optimizing Physical Implementation and Timing Closure 2.2.5. Optimizing Power Consumption 2.2.6. Managing Design Metastability
2.2.2. Optimizing Combinational Logic x
2.2.2.1. Avoid Combinational Loops 2.2.2.2. Avoid Unintended Latch Inference 2.2.2.3. Avoid Delay Chains in Clock Paths 2.2.2.4. Use Synchronous Pulse Generators
2.2.3. Optimizing Clocking Schemes x
2.2.3.1. Register Combinational Logic Outputs 2.2.3.2. Avoid Asynchronous Clock Division 2.2.3.3. Avoid Ripple Counters 2.2.3.4. Use Multiplexed Clocks 2.2.3.5. Use Gated Clocks 2.2.3.6. Use Synchronous Clock Enables
2.2.3.5. Use Gated Clocks x
2.2.3.5.1. Recommended Clock-Gating Methods
2.2.4. Optimizing Physical Implementation and Timing Closure x
2.2.4.1. Planning Physical Implementation 2.2.4.2. Planning FPGA Resources 2.2.4.3. Optimizing for Timing Closure 2.2.4.4. Optimizing Critical Timing Paths
2.3. Use Clock and Register-Control Architectural Features x
2.3.1. Use Global Reset Resources 2.3.2. Use Global Clock Network Resources 2.3.3. Use Clock Region Assignments to Optimize Clock Constraints 2.3.4. Avoid Asynchronous Register Control Signals
2.3.1. Use Global Reset Resources x
2.3.1.1. Use Synchronous Resets 2.3.1.2. Using Asynchronous Resets 2.3.1.3. Use Synchronized Asynchronous Reset
2.3.3. Use Clock Region Assignments to Optimize Clock Constraints x
2.3.3.1. Clock Region Assignments in Stratix® 10 Devices 2.3.3.2. Clock Region Assignments in Arria® 10 and Older Device Families
2.5. Design Assistant Design Rule Checking x
2.5.1. Setting Up Design Assistant 2.5.2. Running Design Assistant During Compilation 2.5.3. Running Design Assistant in Analysis Mode 2.5.4. Cross-Probing from Design Assistant 2.5.5. Managing Design Assistant Rules 2.5.6. Design Assistant Rule Categories
2.5.1. Setting Up Design Assistant x
2.5.1.1. Design Assistant Rule Severity Levels
2.5.2. Running Design Assistant During Compilation x
2.5.2.1. Opening Design Assistant Rule Help
2.5.3. Running Design Assistant in Analysis Mode x
2.5.3.1. Launching Design Assistant from Chip Planner 2.5.3.2. Launching Design Assistant from Timing Analyzer
2.5.4. Cross-Probing from Design Assistant x
2.5.4.1. Cross-Probing from Design Assistant to Timing Analyzer 2.5.4.2. Cross-Probing from Design Assistant to Visualization Tools
2.5.5. Managing Design Assistant Rules x
2.5.5.1. Changing the Default Number of Violations per Rule 2.5.5.2. Enabling Rules for Specific Compiler Stages 2.5.5.3. Specifying Rule Parameters for a Specific Compiler Stage 2.5.5.4. Modifying Rule Severity Levels 2.5.5.5. Waiving Design Assistant Rules 2.5.5.6. Design Assistant Tags
2.5.5.5. Waiving Design Assistant Rules x
2.5.5.5.1. Creating Design Assistant Waivers 2.5.5.5.2. Design Assistant Waiver Dialog Box 2.5.5.5.3. Deleting Design Assistant Waivers 2.5.5.5.4. Design Assistant Waiver Tcl Commands 2.5.5.5.5. drc::add_waiver Command 2.5.5.5.6. drc::get_waivers Command 2.5.5.5.7. drc::report_waivers Command
3. Managing Metastability with the Quartus® Prime Software x
3.1. Metastability Analysis in the Quartus® Prime Software 3.2. Metastability and MTBF Reporting 3.3. MTBF Optimization 3.4. Reducing Metastability Effects 3.5. Scripting Support 3.6. Managing Metastability 3.7. Managing Metastability with the Quartus® Prime Software Revision History
3.1. Metastability Analysis in the Quartus® Prime Software x
3.1.1. Data Synchronization Register Chains 3.1.2. Identify Synchronizers for Metastability Analysis 3.1.3. How Timing Constraints Affect Synchronizer Identification and Metastability Analysis
3.2. Metastability and MTBF Reporting x
3.2.1. Metastability Reports 3.2.2. Synchronizer Data Toggle Rate in MTBF Calculation
3.2.1. Metastability Reports x
3.2.1.1. MTBF Summary Report 3.2.1.2. Synchronizer Summary Report 3.2.1.3. Synchronizer Chain Statistics Report in the Timing Analyzer
3.2.1.1. MTBF Summary Report x
3.2.1.1.1. Typical and Worst-Case MTBF of Design 3.2.1.1.2. Synchronizer Chains 3.2.1.1.3. Increasing Available Settling Time
3.3. MTBF Optimization x
3.3.1. Synchronization Register Chain Length
3.4. Reducing Metastability Effects x
3.4.1. Apply Complete System-Centric Timing Constraints for the Timing Analyzer 3.4.2. Force the Identification of Synchronization Registers 3.4.3. Set the Synchronizer Data Toggle Rate 3.4.4. Optimize Metastability During Fitting 3.4.5. Increase the Length of Synchronizers to Protect and Optimize 3.4.6. Increase the Number of Stages Used in Synchronizers 3.4.7. Select a Faster Speed Grade Device
3.5. Scripting Support x
3.5.1. Identifying Synchronizers for Metastability Analysis 3.5.2. Synchronizer Data Toggle Rate in MTBF Calculation 3.5.3. report_metastability and Tcl Command 3.5.4. MTBF Optimization 3.5.5. Synchronization Register Chain Length
Answers to Top FAQs
1. Recommended HDL Coding Styles
2. Recommended Design Practices
3. Managing Metastability with the Quartus® Prime Software
4. Quartus® Prime Pro Edition User Guide: Design Recommendations Archive
A. Quartus® Prime Pro Edition User Guides

1.4.1.11. Specifying Initial Memory Contents at Power-Up

Your synthesis tool may offer various ways to specify the initial contents of an inferred memory. There are slight power-up and initialization differences between dedicated RAM blocks and the MLAB memory, due to the continuous read of the MLAB.

Intel FPGA dedicated RAM block outputs always power-up to zero, and are set to the initial value on the first read. For example, if address 0 is pre-initialized to FF, the RAM block powers up with the output at 0. A subsequent read after power-up from address 0 outputs the pre-initialized value of FF. Therefore, if a RAM powers up and an enable (read enable or clock enable) is held low, the power-up output of 0 maintains until the first valid read cycle. The synthesis tool implements MLAB using registers that power-up to 0, but initialize to their initial value immediately at power-up or reset. Therefore, the initial value is seen, regardless of the enable status. The Quartus® Prime software maps inferred memory to MLABs when the HDL code specifies an appropriate ramstyle attribute.

In Verilog HDL, you can use an initial block to initialize the contents of an inferred memory. Quartus® Prime Pro Edition synthesis automatically converts the initial block into a Memory Initialization File (.mif) for the inferred RAM.

Verilog HDL RAM with Initialized Contents

module ram_with_init(
   output reg [7:0] q,
   input [7:0] d,
   input [4:0] write_address, read_address,
   input we, clk
);
   reg [7:0] mem [0:31];
   integer i;

   initial begin
      for (i = 0; i < 32; i = i + 1)
         mem[i] = i[7:0];
   end

   always @ (posedge clk) begin
      if (we)
         mem[write_address] <= d;
      q <= mem[read_address];
   end
endmodule

Quartus® Prime Pro Edition synthesis and other synthesis tools also support the $readmemb and $readmemh attributes. These attributes allow RAM initialization and ROM initialization work identically in synthesis and simulation.

Verilog HDL RAM Initialized with the readmemb Command

reg [7:0] ram[0:15];
initial 
begin
	$readmemb("ram.txt", ram);
end

In VHDL, you can initialize the contents of an inferred memory by specifying a default value for the corresponding signal. Quartus® Prime Pro Edition synthesis automatically converts the default value into a .mif file for the inferred RAM.

VHDL RAM with Initialized Contents

LIBRARY ieee;
USE ieee.std_logic_1164.all;
use ieee.numeric_std.all;

ENTITY ram_with_init IS
    PORT(
            clock: IN STD_LOGIC;
            data: IN UNSIGNED (7 DOWNTO 0);
            write_address: IN integer RANGE 0 to 31;
            read_address: IN integer RANGE 0 to 31;
            we: IN std_logic;
            q: OUT UNSIGNED (7 DOWNTO 0));
END;

ARCHITECTURE rtl OF ram_with_init IS

    TYPE MEM IS ARRAY(31 DOWNTO 0) OF unsigned(7 DOWNTO 0);
    FUNCTION initialize_ram
        return MEM is
        variable result : MEM;
    BEGIN 
        FOR i IN 31 DOWNTO 0 LOOP
            result(i) := to_unsigned(natural(i), natural'(8));
        END LOOP; 
        RETURN result;
    END initialize_ram;

    SIGNAL ram_block : MEM := initialize_ram;
BEGIN
    PROCESS (clock)
    BEGIN
        IF (rising_edge(clock)) THEN
            IF (we = '1') THEN
            ram_block(write_address) <= data;
            END IF;
            q <= ram_block(read_address);
        END IF;
    END PROCESS;
END rtl;

Verilog HDL Single-Clock, Simple Dual-Port RAM with Synchronous Reset

// Simple Dual-Port Block with Single Clock

module simple_dual_port_ram_single_clock
#(
   parameter DATA_WIDTH = 8,
   parameter ADDR_WIDTH = 6
)
(
	input clk, clr, we,
	input [ADDR_WIDTH-1:0] waddr,
	input [ADDR_WIDTH-1:0] raddr,
	input [DATA_WIDTH-1:0] data_in,
	output reg [DATA_WIDTH-1:0] data_out
);

(*ramstyle = "M20K"*) reg [DATA_WIDTH-1:0] memory[2**ADDR_WIDTH-1:0];

always @(posedge clk) begin
	if (clr) begin
	   data_out <= 0;
	end
	else begin
          data_out <= memory[raddr];
	end
end

always @(posedge clk) begin
	if (we) begin
	   memory[waddr] <= data_in;
	end
end	
	

endmodule

RAM inference with synchronous reset is supported for single-port, simple-dual port, and true dual port RAM.

Verilog HDL Single-Clock, Simple Dual-Port RAM with Asynchronous Reset

// Simple Dual-Port Block with Single Clock

module simple_dual_port_ram_single_clock_aclr
#(
   parameter DATA_WIDTH = 8,
   parameter ADDR_WIDTH = 6
)
(
	input clk, clr, we,
	input [ADDR_WIDTH-1:0] waddr,
	input [ADDR_WIDTH-1:0] raddr,
	input [DATA_WIDTH-1:0] data_in,
	output reg [DATA_WIDTH-1:0] data_out
);

(*ramstyle = "M20K"*) reg [DATA_WIDTH-1:0] memory[2**ADDR_WIDTH-1:0];


always @(posedge clk or posedge clr) begin
	if (clr) begin
	   data_out <= 0;
	end
	else begin
           data_out <= memory[raddr];
	end
end

always @(posedge clk) begin
	if (we) begin
		memory[waddr] <= data_in;
	end
end	
	

endmodule

RAM inference with asynchronous reset is supported for single-port and simple-dual port RAM.

Level Two Title