AN 763: Intel® Arria® 10 SoC Device Design Guidelines

Download PDF
ID 683192
Date 5/17/2022
Public
Document Table of Contents
Document Table of Contents x
1. Overview of Design Guidelines for Intel® Arria® 10 SoC FPGAs 2. Guidelines for Interconnecting the Intel® Arria® 10 HPS and FPGA 3. Design Guidelines for HPS Portion of Arria 10 SoC FPGAs 4. Board Design Guidelines for Arria 10 SoC FPGAs 5. Embedded Software Design Guidelines for Arria 10 SoC FPGAs
1. Overview of Design Guidelines for Intel® Arria® 10 SoC FPGAs x
1.1. SoC FPGA Designer's Checklist 1.2. Overview of HPS Design Guidelines for SoC FPGA design 1.3. Overview of Board Design Guidelines for SoC FPGA Design 1.4. Overview of Embedded Software Design Guidelines for SoC FPGA Design 1.5. Overview of Design Guidelines for Intel® Arria® 10 SoC FPGAs Revision History
2. Guidelines for Interconnecting the Intel® Arria® 10 HPS and FPGA x
2.1. Overview of HPS Memory-Mapped Interfaces 2.2. Recommended System Topologies 2.3. Guidelines for Interconnecting the Intel® Arria® 10 HPS and FPGA Revision History
2.1. Overview of HPS Memory-Mapped Interfaces x
2.1.1. HPS-to-FPGA Bridge 2.1.2. Lightweight HPS-to-FPGA Bridge 2.1.3. FPGA-to-HPS Bridge 2.1.4. FPGA-to-SDRAM Ports 2.1.5. Interface Bandwidths
2.2. Recommended System Topologies x
2.2.1. HPS Accesses to FPGA Fabric 2.2.2. Maintaining Cache Coherency GUIDELINE: Ensure that data accessed through the ACP slave fits in the 512 KB L2 cache to avoid thrashing overhead. GUIDELINE: For small buffers of data shared between the MPU and system masters, consider having the system master perform cacheable accesses to avoid overhead caused by cache flushing operations. Guideline: Avoid needing to coherently access back the HPS through ACP from the fabric in order to complete an access coming from HPS, as this may result in a deadlock situation. 2.2.3. MPU Sharing Data with FPGA 2.2.4. Examples of Cacheable and Non-Cacheable Data Accesses From the FPGA
2.2.4. Examples of Cacheable and Non-Cacheable Data Accesses From the FPGA x
2.2.4.1. Example 1: FPGA Reading Data from HPS SDRAM Directly 2.2.4.2. Example 2: FPGA Writing Data into HPS SDRAM Directly 2.2.4.3. Example 3: FPGA Reading Cache Coherent Data from HPS 2.2.4.4. Example 4: FPGA Writing Cache Coherent Data to HPS
3. Design Guidelines for HPS Portion of Arria 10 SoC FPGAs x
3.1. Start your SoC FPGA design here 3.2. Design Considerations for Connecting Device I/O to HPS Peripherals and Memory 3.3. HPS Clocking and Reset Design Considerations 3.4. HPS EMIF Design Considerations 3.5. DMA Considerations 3.6. Design Guidelines for HPS Portion of Intel® Arria® 10 SoC FPGAs Revision History
3.1. Start your SoC FPGA design here x
3.1.1. Recommended Starting Point for HPS-to-FPGA Interface Designs 3.1.2. Determining your SoC FPGA Topology
3.2. Design Considerations for Connecting Device I/O to HPS Peripherals and Memory x
3.2.1. HPS Pin Multiplexing Design Considerations 3.2.2. HPS I/O Settings: Constraints and Drive Strengths
3.3. HPS Clocking and Reset Design Considerations x
3.3.1. HPS Clock Planning 3.3.2. Early Pin Planning and I/O Assignment Analysis 3.3.3. Pin Features and Connections for HPS Clocks, Reset and PoR 3.3.4. Internal Clocks 3.3.5. HPS Reset During FPGA Reconfiguration and FPGA Configuration Failures 3.3.6. HPS Peripheral Reset Management
3.4. HPS EMIF Design Considerations x
3.4.1. Considerations for Connecting HPS to SDRAM 3.4.2. HPS SDRAM I/O Locations 3.4.3. Integrating the Arria 10 HPS EMIF with the SoC FPGA Device 3.4.4. HPS Memory Debug
3.4.2. HPS SDRAM I/O Locations x
3.4.2.1. I/O Bank 2K, Lanes 0,1,2 (Addr/Cmd) 3.4.2.2. I/O Bank 2K, Lane 3 (ECC) 3.4.2.3. I/O Bank, 2J (Data) 3.4.2.4. I/O Bank, 2I (Data, 64-, 72-bit interfaces)
3.5. DMA Considerations x
3.5.1. Choosing a DMA Controller 3.5.2. Optimizing DMA Master Bandwidth through HPS Interconnect
4. Board Design Guidelines for Arria 10 SoC FPGAs x
4.1. Power On Board Bring Up and Boot ROM/Boot Loader Debugging 4.2. FPGA Reconfiguration 4.3. HPS Power Design Considerations 4.4. Boundary Scan for HPS 4.5. Design Guidelines for HPS Interfaces 4.6. Connection Guidelines for Unused HPS Block 4.7. Board Design Guidelines for Intel® Arria® 10 SoC FPGAs Revision History
4.2. FPGA Reconfiguration x
4.2.1. Flash Update with HPS Reboot 4.2.2. Partial Reconfiguration of the SoC FPGA
4.3. HPS Power Design Considerations x
4.3.1. Early System and Board Planning 4.3.2. Design Considerations for HPS and FPGA Power Supplies for SoC FPGA devices 4.3.3. Pin Connection Considerations for Board Designs 4.3.4. Power Analysis 4.3.5. Power Optimization
4.3.1. Early System and Board Planning x
4.3.1.1. Early Power Estimation
4.3.1.1. Early Power Estimation x
4.3.1.1.1. Main Worksheet 4.3.1.1.2. IO Worksheet 4.3.1.1.3. IO-IP Worksheet 4.3.1.1.4. HPS Worksheet
4.3.2. Design Considerations for HPS and FPGA Power Supplies for SoC FPGA devices x
4.3.2.1. Consider Device Power Consumption and HPS Performance 4.3.2.2. Consider Desired HPS Boot Clock Frequency
4.3.3. Pin Connection Considerations for Board Designs x
4.3.3.1. Device Power-Up 4.3.3.2. Power Pin Connections and Power Supplies
4.3.5. Power Optimization x
4.3.5.1. Processor and Memory Clock Speeds 4.3.5.2. MPU Standby Modes and Dynamic Clock Gating 4.3.5.3. Managing Peripheral Power 4.3.5.4. Managing Power by Shutting Down Supplies
4.5. Design Guidelines for HPS Interfaces x
4.5.1. HPS EMAC PHY Interfaces 4.5.2. USB Interface Design Guidelines 4.5.3. QSPI Flash Interface Design Guidelines 4.5.4. SD/MMC and eMMC Card Interface Design Guidelines 4.5.5. Provide Flash Memory Reset for QSPI and SD/MMC/eMMC 4.5.6. NAND Flash Interface Design Guidelines 4.5.7. UART Interface Design Guidelines 4.5.8. I2C Interface Design Guidelines
4.5.1. HPS EMAC PHY Interfaces x
4.5.1.1. PHY Interfaces Connected Through Shared I/O 4.5.1.2. PHY Interfaces Connected Through FPGA I/O 4.5.1.3. MDIO 4.5.1.4. Common PHY Interface Design Considerations
4.5.1.1. PHY Interfaces Connected Through Shared I/O x
4.5.1.1.1. RMII 4.5.1.1.2. RGMII
4.5.1.2. PHY Interfaces Connected Through FPGA I/O x
4.5.1.2.1. GMII/MII 4.5.1.2.2. Adapting to RMII 4.5.1.2.3. Adapting to SGMII
4.5.1.4. Common PHY Interface Design Considerations x
4.5.1.4.1. Signal Integrity
5. Embedded Software Design Guidelines for Arria 10 SoC FPGAs x
5.1. Embedded Software for HPS Design Guidelines 5.2. Support and Documentation 5.3. Embedded Software Design Guidelines for Intel® Arria® 10 SoC FPGAs Revision History
5.1. Embedded Software for HPS Design Guidelines x
5.1.1. Purpose 5.1.2. Assembling the components of your Software Development Platform 5.1.3. Selecting an Operating System for your application 5.1.4. Assembling your Software Development Platform for Linux 5.1.5. Assembling your Software Development Platform for a Bare-Metal Application 5.1.6. Assembling your Software Development Platform for Partner OS or RTOS 5.1.7. Choosing Boot Loader Software 5.1.8. Selecting Software Tools for Development, Debug and Trace 5.1.9. Board Bring Up Considerations 5.1.10. Boot and Configuration Design Considerations 5.1.11. Flash Device Driver Design Considerations 5.1.12. HPS ECC Design Considerations 5.1.13. Security Design Considerations 5.1.14. Embedded Software Debugging and Trace
5.1.2. Assembling the components of your Software Development Platform x
5.1.2.1. Golden Hardware Reference Design (GHRD)
5.1.3. Selecting an Operating System for your application x
5.1.3.1. Using Linux or RTOS 5.1.3.2. Developing a Bare-Metal Application 5.1.3.3. Using the Bootloader as a Bare-Metal Framework 5.1.3.4. Using Symmetrical vs. Asymmetrical Multiprocessing (SMP vs. AMP) Modes
5.1.4. Assembling your Software Development Platform for Linux x
5.1.4.1. Golden System Reference Design (GSRD) for Linux 5.1.4.2. GSRD for Linux Build Flow 5.1.4.3. Source Code Management Considerations 5.1.4.4. Linux Device Tree Design Considerations
5.1.8. Selecting Software Tools for Development, Debug and Trace x
5.1.8.1. Selecting Software Build Tools 5.1.8.2. Selecting Software Debug Tools 5.1.8.3. Selecting Software Trace Tools
5.1.9. Board Bring Up Considerations x
5.1.9.1. Plan the SDRAM Initialization
5.1.10. Boot and Configuration Design Considerations x
5.1.10.1. Boot Design Considerations 5.1.10.2. Configuration
5.1.10.1. Boot Design Considerations x
5.1.10.1.1. Boot Source 5.1.10.1.2. Select Desired Flash Device 5.1.10.1.3. BSEL Options 5.1.10.1.4. Boot Clock 5.1.10.1.5. Determine Boot Fuses Usage 5.1.10.1.6. CSEL Options 5.1.10.1.7. Determine Flash Programming Method 5.1.10.1.8. Selecting NAND Flash Devices 5.1.10.1.9. Selecting QSPI Flash Devices 5.1.10.1.10. Reference Materials
5.1.10.2. Configuration x
5.1.10.2.1. Traditional Configuration 5.1.10.2.2. HPS-Initiated Configuration
5.1.12. HPS ECC Design Considerations x
5.1.12.1. General ECC Design Considerations 5.1.12.2. ECC for External SDRAM Interface 5.1.12.3. ECC for L2 Cache Data Memory 5.1.12.4. ECC for Flash Memory
5.2. Support and Documentation x
5.2.1. Support 5.2.2. Hardware Documentation 5.2.3. Software Documentation
5.2.3. Software Documentation x
5.2.3.1. Example of Prebuilt Bootloaders for the Intel® Arria® 10 SoC Development Kit
1. Overview of Design Guidelines for Intel® Arria® 10 SoC FPGAs
2. Guidelines for Interconnecting the Intel® Arria® 10 HPS and FPGA
3. Design Guidelines for HPS Portion of Arria 10 SoC FPGAs
4. Board Design Guidelines for Arria 10 SoC FPGAs
5. Embedded Software Design Guidelines for Arria 10 SoC FPGAs

2.2.2. Maintaining Cache Coherency

Cache coherency is a fundamental topic to understand any time data must be shared amongst multiple masters in a system. In the context of a SoC device these masters can be the MPU, DMA, peripherals with master interfaces, and masters in the FPGA connected to the HPS. Since the MPU contains level 1 and level 2 cache controllers it can hold more up-to-date contents than main memory in the system. The HPS supports two mechanisms to make sure masters in the system observe a coherent view of memory: ensuring main memory contains the latest value, or have masters access the ACP slave of the HPS.

The MPU can allocate buffers to be non-cacheable which ensures data is never cached by the L1 and L2 caches. The MPU can also access cacheable data and either flush it to main memory or copy it to a non-cacheable buffer before other masters attempt to access the data. Operating systems typically provide mechanisms for maintaining cache coherency both ways described above.

Masters in the system access coherent data by either relying on the MPU to place data into main memory instead of having it cached, or by having the master in the system perform a cacheable access via the ACP slave. Which mechanism you use typically depends on the size of the buffer of memory the master is accessing.

GUIDELINE: Ensure that data accessed through the ACP slave fits in the 512 KB L2 cache to avoid thrashing overhead.

Since the L2 cache size is 512 KB, if a master in the system frequently accesses buffers whose total size exceeds 512 KB, thrashing results.

Cache thrashing is a situation where the size of the data exceeds the size of the cache, causing the cache to perform frequent evictions and prefetches to main memory. Thrashing negates the performance benefits of caching the data.

In potential thrashing situation, it makes more sense to have the masters access non-cache coherent data and allow software executing on the MPU maintain the data coherency throughout the system.

GUIDELINE: For small buffers of data shared between the MPU and system masters, consider having the system master perform cacheable accesses to avoid overhead caused by cache flushing operations.

If a master in the system requires access to smaller coherent blocks of data then you should consider having the MPU access the buffer as cacheable memory and the master in the system perform cacheable accesses to the data. Cacheable accesses are automatically routed to the MPU ACP slave ensuring that the master and MPU access the same copy of the data. By having the MPU use cacheable buffers and the system master performing cacheable accesses, software does not have to maintain system wide coherency ensuring both the MPU and system master observe the same copy of data.

Guideline: Avoid needing to coherently access back the HPS through ACP from the fabric in order to complete an access coming from HPS, as this may result in a deadlock situation.

Certain coherent access scenarios can create deadlock through the ACP and the CPU. However, you can avoid this type of deadlock with a simple access strategy.

In the following example, the CPU creates deadlock by initiating an access to the HPS through the FPGA fabric:
  1. The CPU initiates a device memory access to the FPGA fabric. The CPU pipeline must stall until this type of access is complete.
  2. Before the FPGA fabric state machine can respond to the device memory access, it must access the HPS coherently. It initiates a coherent access, which requires the ACP.
  3. The ACP must perform a cache maintenance operation before it can complete the access. However, the CPU’s pipeline stall prevents it from performing the cache maintenance operation. The system deadlocks.

You can implement the desired access without deadlock, by breaking it into smaller pieces. For example, you can initiate the operation with one access, then determine the operation status with a second access.

Level Two Title