External Memory Interfaces (EMIF) IP User Guide: Agilex™ 3 FPGAs and SoCs

Download PDF
ID 847458
Date 4/24/2025
Public
Document Table of Contents
Document Table of Contents x
1. About the External Memory Interfaces Agilex™ 3 FPGA IP 2. Agilex™ 3 FPGA EMIF IP – Introduction 3. Agilex™ 3 FPGA EMIF IP - Configuring and Generating the IP 4. Agilex™ 3 FPGA EMIF IP – Simulating Memory IP 5. Agilex™ 3 FPGA EMIF IP - Validating the IP 6. Agilex 3 FPGA EMIF IP Debugging 7. Document Revision History for External Memory Interfaces (EMIF) IP User Guide A. Agilex™ 3 FPGA EMIF IP – Product Architecture B. Agilex™ 3 FPGA EMIF IP – End-User Signals
1. About the External Memory Interfaces Agilex™ 3 FPGA IP x
1.1. Release Information
2. Agilex™ 3 FPGA EMIF IP – Introduction x
2.1. Agilex™ 3 EMIF IP Protocol and Feature Support 2.2. Protocol and Maximum Interface Width Support 2.3. Agilex 3 EMIF IP Design Flow 2.4. Agilex™ 3 EMIF IP Design Checklist
3. Agilex™ 3 FPGA EMIF IP - Configuring and Generating the IP x
3.1. Creating an EMIF Project 3.2. Generating and Configuring the EMIF IP 3.3. EMIF IP LPDDR4 Parameter Descriptions 3.4. Generating HDL for Synthesis and Simulation 3.5. Generating the Synthesizable EMIF Design Example 3.6. Agilex™ 3 FPGA EMIF IP Pin and Resource Planning 3.7. Compiling the Agilex™ 3 EMIF Design Example 3.8. Agilex™ 3 FPGA EMIF IP – Timing Closure 3.9. Agilex™ 3 FPGA EMIF IP – Controller Optimization
3.3. EMIF IP LPDDR4 Parameter Descriptions x
3.3.1. Configuring DQ Pin Swizzling
3.3.1. Configuring DQ Pin Swizzling x
3.3.1.1. Example: DQ Pin Swizzling Within DQS Group for a x32 LPDDR4 Interface 3.3.1.2. Example: Byte Swizzling for x32 LPDDR4 Interface 3.3.1.3. Example: Combining Pin and Byte Swizzling 3.3.1.4. Example: DQ Pin Swizzling Within DQS Group for 2 Channel x16 LPDDR4 Interface 3.3.1.5. Example: Byte Swizzling for 2 Channel x16 LPDDR4 Interface
3.5. Generating the Synthesizable EMIF Design Example x
3.5.1. Synthesis Design Example
3.6. Agilex™ 3 FPGA EMIF IP Pin and Resource Planning x
3.6.1. Agilex™ 3 FPGA EMIF IP Interface Pins 3.6.2. Pin Guidelines for Agilex™ 3 FPGA EMIF IP 3.6.3. Agilex™ 3 FPGA EMIF IP Resources
3.6.1. Agilex™ 3 FPGA EMIF IP Interface Pins x
3.6.1.1. Estimating Pin Requirements 3.6.1.2. LPDDR4 Component Options 3.6.1.3. Maximum Number of Interfaces 3.6.1.4. Address and Command Pin Placement for LPDDR4 3.6.1.5. LPDDR4 Data Width Mapping 3.6.1.6. Pin Placement for Agilex™ 3 EMIF IP
3.6.2. Pin Guidelines for Agilex™ 3 FPGA EMIF IP x
3.6.2.1. General Guidelines 3.6.2.2. Specific Pin Connection Requirements 3.6.2.3. Command and Address Signals 3.6.2.4. Clock Signals 3.6.2.5. Pin Swapping Guidelines
3.6.3. Agilex™ 3 FPGA EMIF IP Resources x
3.6.3.1. OCT 3.6.3.2. PLL
3.8. Agilex™ 3 FPGA EMIF IP – Timing Closure x
3.8.1. Timing Closure 3.8.2. Timing Analysis 3.8.3. PHY or Core 3.8.4. Optimizing Timing
3.9. Agilex™ 3 FPGA EMIF IP – Controller Optimization x
3.9.1. Optimizing Efficiency for Secondary Controller Asychronous Clocking Mode (Fabric Direct - User Clock Asychronous to PHY) Synchronous Clocking Mode (Fabric Direct – User Clock Synchronous to PHY) 3.9.2. Interface Standard 3.9.3. Bank Management Efficiency 3.9.4. Data Transfer 3.9.5. Improving Controller Efficiency
3.9.5. Improving Controller Efficiency x
3.9.5.1. Series of Reads or Writes 3.9.5.2. AXI to Memory Mapping
4. Agilex™ 3 FPGA EMIF IP – Simulating Memory IP x
4.1. Simulation Walkthrough 4.2. Generating the EMIF Design Example for Simulation 4.3. Simulation Design Example 4.4. Simulating the Design Example
4.1. Simulation Walkthrough x
4.1.1. Calibration 4.1.2. Simulation Scripts 4.1.3. Functional Simulation with Verilog HDL
5. Agilex™ 3 FPGA EMIF IP - Validating the IP x
5.1. Validating the Design Example with the Test Engine IP 5.2. Checking the EMIF Design Example with the Performance Monitor
6. Agilex 3 FPGA EMIF IP Debugging x
6.1. Interface Configuration Performance Issues 6.2. Functional Issue Evaluation 6.3. Timing Issue Characteristics 6.4. Evaluating FPGA Timing Issues 6.5. Verifying Memory IP Using the Signal Tap Logic Analyzer 6.6. Guidelines for Developing HDL for Traffic Generator 6.7. Debugging with the External Memory Interface Debug Toolkit 6.8. Guidelines for Traffic Generator Status Check 6.9. Hardware Debugging Guidelines 6.10. Categorizing Hardware Issues 6.11. Agilex™ 3 FPGA EMIF IP - Mailbox Support
6.1. Interface Configuration Performance Issues x
6.1.1. Interface Configuration Bottleneck and Efficiency Issues
6.2. Functional Issue Evaluation x
6.2.1. Altera IP Memory Model 6.2.2. Vendor Memory Model 6.2.3. Transcript Window Messages
6.4. Evaluating FPGA Timing Issues x
6.4.1. Evaluating External Memory Interface Timing Issues
6.7. Debugging with the External Memory Interface Debug Toolkit x
6.7.1. Prerequisites for Using the EMIF Debug Toolkit 6.7.2. Configuring a Design to Use the Debug Toolkit 6.7.3. Launching the EMIF Debug Toolkit 6.7.4. Using the EMIF Debug Toolkit
6.7.2. Configuring a Design to Use the Debug Toolkit x
6.7.2.1. Generating a Design Example with the Debug Toolkit
6.7.4. Using the EMIF Debug Toolkit x
6.7.4.1. Running Re-Calibration 6.7.4.2. Running the Test Engine 6.7.4.3. Generating Traffic with the Test Engine IP
6.8. Guidelines for Traffic Generator Status Check x
6.8.1. Status Check Using the Signal Tap Logic Analyzer 6.8.2. Exporting the Status Interface to the Top-Level Design
6.9. Hardware Debugging Guidelines x
6.9.1. Create a Simplified Design that Demonstrates the Same Issue 6.9.2. Measure Power Distribution Network 6.9.3. Measure Signal Integrity and Setup and Hold Margin 6.9.4. Vary Voltage 6.9.5. Operate at a Lower Speed 6.9.6. Determine Whether the Issue Exists in Previous Versions of Software 6.9.7. Determine Whether the Issue Exists in the Current Version of Software 6.9.8. Try A Different PCB 6.9.9. Try Other Configurations 6.9.10. Debugging Checklist
6.10. Categorizing Hardware Issues x
6.10.1. Signal Integrity Issues 6.10.2. Hardware and Calibration Issues
6.10.1. Signal Integrity Issues x
6.10.1.1. Characteristics of Signal Integrity Issues 6.10.1.2. Evaluating Signal Integrity Issues 6.10.1.3. Skew 6.10.1.4. Crosstalk 6.10.1.5. Power System 6.10.1.6. Clock Signals 6.10.1.7. Address and Command Signals 6.10.1.8. Read Data Valid Window and Eye Diagram 6.10.1.9. Write Data Valid Window and Eye Diagram
6.10.2. Hardware and Calibration Issues x
6.10.2.1. Verifying High-Level Configuration 6.10.2.2. Verifying Memory Timing Parameters 6.10.2.3. Verifying the Correct Memory Component Installed 6.10.2.4. Debugging Intermittent Issues
6.11. Agilex™ 3 FPGA EMIF IP - Mailbox Support x
6.11.1. Agilex™ 3 Mailbox Structure and Register Definitions 6.11.2. Accessing Read-Only Registers 6.11.3. Sending a Mailbox Command 6.11.4. IOSSM Mailbox Access Script
6.11.1. Agilex™ 3 Mailbox Structure and Register Definitions x
6.11.1.1. Mailbox Supported Commands 6.11.1.2. Mailbox Command Definitions 6.11.1.3. ECC Syndrome Codes 6.11.1.4. ECC Error Handling
6.11.2. Accessing Read-Only Registers x
6.11.2.1. Example 1: Reading IP_TYPE and IP_INSTANCE_ID of All the Interfaces in the IO96B Through Read-Only Registers 6.11.2.2. Example 2: Reading the Memory Clock Frequency for an Interface
6.11.3. Sending a Mailbox Command x
6.11.3.1. Example 3: Sending Recalibration Request and Reading Calibration Status Using the Mailbox 6.11.3.2. Example 4: Sending ECC_INJECT_ERROR Using the Mailbox and Reading ECC Error Status, ECC Error Buffer Registers
6.11.4. IOSSM Mailbox Access Script x
6.11.4.1. Mailbox Script Overview 6.11.4.2. Running the Mailbox Script
6.11.4.1. Mailbox Script Overview x
6.11.4.1.1. mb_iossm_command.tcl 6.11.4.1.2. mb_user_input.tcl
A. Agilex™ 3 FPGA EMIF IP – Product Architecture x
A.1. Agilex™ 3 EMIF Architecture: Introduction
A.1. Agilex™ 3 EMIF Architecture: Introduction x
A.1.1. Agilex™ 3 EMIF Architecture: I/O Subsystem A.1.2. Agilex™ 3 EMIF Architecture: I/O SSM A.1.3. Agilex™ 3 EMIF Architecture: HSIO Bank A.1.4. Agilex™ 3 EMIF Architecture: I/O Lane A.1.5. Agilex™ 3 EMIF Architecture: Input DQS Clock Tree A.1.6. Agilex™ 3 EMIF Architecture: PHY Clock Tree A.1.7. Agilex™ 3 EMIF Architecture: PLL Reference Clock Networks A.1.8. Agilex™ 3 EMIF Architecture: Clock Phase Alignment A.1.9. User Clock in Different Core Access Modes A.1.10. Agilex™ 3 EMIF Sequencer A.1.11. Agilex™ 3 EMIF Controller A.1.12. Hard Memory Controller A.1.13. Agilex™ 3 EMIF IP for Hard Processor Subsystem (HPS)
A.1.3. Agilex™ 3 EMIF Architecture: HSIO Bank x
A.1.3.1. Narrow Read Transfer Support A.1.3.2. HSIO Sub-Bank Usage
A.1.12. Hard Memory Controller x
A.1.12.1. Hard Memory Controller Features
A.1.13. Agilex™ 3 EMIF IP for Hard Processor Subsystem (HPS) x
A.1.13.1. Restrictions on I/O Bank Usage for Agilex™ 3 EMIF IP with HPS A.1.13.2. Using the Legacy EMIF Debug Toolkit with Agilex™ 3 HPS Interfaces
B. Agilex™ 3 FPGA EMIF IP – End-User Signals x
B.1. IP Interfaces for External Memory Interfaces (EMIF) IP - LPDDR4 B.2. s0_axi4_clock_in for External Memory Interfaces (EMIF) IP - LPDDR4 B.3. core_init_n for External Memory Interfaces (EMIF) IP - LPDDR4 B.4. s0_axi4_clock_in for External Memory Interfaces (EMIF) IP - LPDDR4 B.5. core_init_n for External Memory Interfaces (EMIF) IP - LPDDR4 B.6. s0_axi4_ctrl_ready for External Memory Interfaces (EMIF) IP - LPDDR4 B.7. s0_axi4_clock_out for External Memory Interfaces (EMIF) IP - LPDDR4 B.8. s1_axi4_ctrl_ready for External Memory Interfaces (EMIF) IP - LPDDR4 B.9. s0_axi4 for External Memory Interfaces (EMIF) IP - LPDDR4 B.10. s1_axi4 for External Memory Interfaces (EMIF) IP - LPDDR4 B.11. io96b0_to_hps for External Memory Interfaces (EMIF) IP - LPDDR4 B.12. io96b1_to_hps for External Memory Interfaces (EMIF) IP - LPDDR4 B.13. s0_axi4lite_clock for External Memory Interfaces (EMIF) IP - LPDDR4 B.14. s0_axi4lite_reset_n for External Memory Interfaces (EMIF) IP - LPDDR4 B.15. s0_axi4lite for External Memory Interfaces (EMIF) IP - LPDDR4 B.16. mem_0 for External Memory Interfaces (EMIF) IP - LPDDR4 B.17. mem_ck_0 for External Memory Interfaces (EMIF) IP - LPDDR4 B.18. mem_1 for External Memory Interfaces (EMIF) IP - LPDDR4 B.19. mem_ck_1 for External Memory Interfaces (EMIF) IP - LPDDR4 B.20. mem_reset_n for External Memory Interfaces (EMIF) IP - LPDDR4 B.21. oct_0 for External Memory Interfaces (EMIF) IP - LPDDR4 B.22. oct_1 for External Memory Interfaces (EMIF) IP - LPDDR4 B.23. ref_clk for External Memory Interfaces (EMIF) IP - LPDDR4
1. About the External Memory Interfaces Agilex™ 3 FPGA IP
2. Agilex™ 3 FPGA EMIF IP – Introduction
3. Agilex™ 3 FPGA EMIF IP - Configuring and Generating the IP
4. Agilex™ 3 FPGA EMIF IP – Simulating Memory IP
5. Agilex™ 3 FPGA EMIF IP - Validating the IP
6. Agilex 3 FPGA EMIF IP Debugging
7. Document Revision History for External Memory Interfaces (EMIF) IP User Guide
A. Agilex™ 3 FPGA EMIF IP – Product Architecture
B. Agilex™ 3 FPGA EMIF IP – End-User Signals

3.9.1. Optimizing Efficiency for Secondary Controller

You can encounter lower than expected efficiency for sequential access patterns on the secondary memory controller in the following EMIF configurations:
  • 2ch x16 LPDDR4
  • 1ch x16 of LPDDR4 on the top sub-bank

Asychronous Clocking Mode (Fabric Direct - User Clock Asychronous to PHY)

You can improve the efficiency for sequential access pattern on the secondary controller by increasing the AXI clock frequency. Recommendation is to set the AXI clock frequency to 1/4 of memory clock frequency. You can further improve the efficiency of sequential write access pattern on the secondary controller by increasing the burst length as shown in the Table 1. The data shown in the table is obtained by using a 2ch x16 LPDDR4 design, running at 1066.667MHz.

Table 32.  Controller Efficiency with Different AXI Clock Frequency and Burst Length
AXI Clock Frequency (MHz) Pattern Burst Length Controller Efficiency (%)
Primary Secondary

133.333

(Equivalent to Sync Mode) 		Sequential write 2 ~90 ~40
Sequential read ~90 ~50

266.667

 Sequential write 2 ~90 ~75
Sequential read ~90 ~90
Sequential write 4 ~90 ~80
Sequential read ~90 ~90
Sequential write 8 ~90 ~90
Sequential read ~90 ~90
Sequential write 16 ~90 ~90
Sequential read ~90 ~90

Random access patterns or mixed traffic patterns can mitigate the differences in the efficiency for the primary controller and secondary controller.

Table 33.  Controller Efficiency with Different AXI Clock Frequency for Random Traffic Pattern
AXI Clock Frequency (MHz Pattern Burst Length Controller Efficiency (%)
Primary Secondary
133.333 Random write 2 ~40 ~39
Random read ~38 ~38
266.667 Random write 2 ~41 ~41
Random read ~38 ~38

With increased AXI clock frequency, the efficiency number reported by the PMON IP represents the utilization of the AXI bus instead of controller efficiency. The following equation represents the efficiency of the memory controller:

Controller efficiency = AXI transactions accepted by AXI bus / Memory Controller Bandwidth * 100 %

= ( PMON efficiency * AXI Clock Freq * 256-bit ) / (16-bit * Mem Clock Freq ) * 100%

Synchronous Clocking Mode (Fabric Direct – User Clock Synchronous to PHY)

When using x16 2-channel configuration for long burst sequential access in your LPDDR4 interface during write or read operation, performance is limited to 50% per read/write sub-channel for the secondary controller.

If your application requires high controller efficiency for sequential access pattern, it is recommended to use asynchronous clocking mode. Random access patterns or mixed traffic patterns can mitigate the differences in the efficiency for the primary controller and secondary controller.

Note that in synchronous clocking mode, the efficiency for the primary controller used for implementing a x32 LPDDR4 memory interface has reduced performance. This is because the bandwidth for AXI bus is only half of the bandwidth for the memory controller.

AXI Bus Bandwidth = AXI Clock Freq x 256-bit = ( Memory Clock Freq / 8 ) * 256 = 32 * Memory Clock Frequency

Memory Controller Bandwidth = Memory Clock Freq * 2 * 32-bit = 64 * Memory Clock Frequency = 2 * AXI Bus Bandwidth

Level Two Title