Agilex™ 5 FPGAs and SoCs Device Overview

Download PDF
ID 762191
Date 4/01/2024
Public
Document Table of Contents
Document Table of Contents x
1. Overview of the Agilex™ 5 FPGAs and SoCs 2. Agilex™ 5 FPGAs and SoCs Family Plan 3. Second Generation Hyperflex® Core Architecture 4. Adaptive Logic Module in Agilex™ 5 FPGAs and SoCs 5. Internal Embedded Memory in Agilex™ 5 FPGAs and SoCs 6. Variable-Precision DSP in Agilex™ 5 FPGAs and SoCs 7. Core Clock Network in Agilex™ 5 FPGAs and SoCs 8. General Purpose I/Os in Agilex™ 5 FPGAs and SoCs 9. I/O PLLs in Agilex™ 5 FPGAs and SoCs 10. External Memory Interface in Agilex™ 5 FPGAs and SoCs 11. Hard Processor System in Agilex™ 5 SoCs 12. Transceivers in Agilex™ 5 FPGAs and SoCs 13. MIPI* Protocols Support in Agilex™ 5 FPGAs and SoCs 14. Variable Pitch BGA (VPBGA) Package Design of Agilex™ 5 FPGAs and SoCs 15. Configuration via Protocol Using PCIe* for Agilex™ 5 FPGAs and SoCs 16. Device Configuration and the SDM in Agilex™ 5 FPGAs and SoCs 17. Partial and Dynamic Configuration of Agilex™ 5 FPGAs and SoCs 18. Device Security for Agilex™ 5 FPGAs and SoCs 19. SEU Error Detection and Correction in Agilex™ 5 FPGAs and SoCs 20. Power Management for Agilex™ 5 FPGAs and SoCs 21. Software and Tools for Agilex™ 5 FPGAs and SoCs 22. Revision History for the Agilex™ 5 FPGAs and SoCs Device Overview
1. Overview of the Agilex™ 5 FPGAs and SoCs x
1.1. Key Features and Innovations in Agilex™ 5 FPGAs and SoCs 1.2. Agilex™ 5 FPGAs and SoCs Block Diagram 1.3. Agilex™ 5 FPGAs and SoCs Summary of Features 1.4. Additional Features for Agilex™ 5 SoCs
2. Agilex™ 5 FPGAs and SoCs Family Plan x
2.1. Agilex™ 5 FPGAs and SoCs D-Series 2.2. Agilex™ 5 FPGAs and SoCs E-Series 2.3. Agilex™ 5 FPGAs and SoCs Package Options 2.4. Part Number Decoder
10. External Memory Interface in Agilex™ 5 FPGAs and SoCs x
10.1. External Memory Interface Performance 10.2. Features of the Hard Memory Controller
12. Transceivers in Agilex™ 5 FPGAs and SoCs x
12.1. PMA Features in Agilex™ 5 FPGA GTS Transceivers 12.2. PCS Features in Agilex™ 5 FPGA GTS Transceivers 12.3. GTS Transceiver PLL in Agilex™ 5 FPGAs and SoCs
1. Overview of the Agilex™ 5 FPGAs and SoCs
2. Agilex™ 5 FPGAs and SoCs Family Plan
3. Second Generation Hyperflex® Core Architecture
4. Adaptive Logic Module in Agilex™ 5 FPGAs and SoCs
5. Internal Embedded Memory in Agilex™ 5 FPGAs and SoCs
6. Variable-Precision DSP in Agilex™ 5 FPGAs and SoCs
7. Core Clock Network in Agilex™ 5 FPGAs and SoCs
8. General Purpose I/Os in Agilex™ 5 FPGAs and SoCs
9. I/O PLLs in Agilex™ 5 FPGAs and SoCs
10. External Memory Interface in Agilex™ 5 FPGAs and SoCs
11. Hard Processor System in Agilex™ 5 SoCs
12. Transceivers in Agilex™ 5 FPGAs and SoCs
13. MIPI* Protocols Support in Agilex™ 5 FPGAs and SoCs
14. Variable Pitch BGA (VPBGA) Package Design of Agilex™ 5 FPGAs and SoCs
15. Configuration via Protocol Using PCIe* for Agilex™ 5 FPGAs and SoCs
16. Device Configuration and the SDM in Agilex™ 5 FPGAs and SoCs
17. Partial and Dynamic Configuration of Agilex™ 5 FPGAs and SoCs
18. Device Security for Agilex™ 5 FPGAs and SoCs
19. SEU Error Detection and Correction in Agilex™ 5 FPGAs and SoCs
20. Power Management for Agilex™ 5 FPGAs and SoCs
21. Software and Tools for Agilex™ 5 FPGAs and SoCs
22. Revision History for the Agilex™ 5 FPGAs and SoCs Device Overview

6. Variable-Precision DSP in Intel Agilex® 5 FPGAs and SoCs

The Intel Agilex® 5 FPGAs and SoCs are the first midrange or edge-centric FPGAs with an AI tensor block, making it the ideal choice for edge AI applications.

For INT8 operations in a single DSP block, the Intel Agilex® 5 FPGAs and SoCs improve peak theoretical TOPS:

  • D-Series—up to 2.5 times than Intel® Stratix® 10 FPGAs
  • E-Series—up to 37 times than Cyclone® V FPGAs

Through a large increase in arithmetic density13 , the Intel Agilex® 5 FPGAs and SoCs fit more multipliers and accumulators in the same footprint of a standard DSP block.

The Intel® FPGA AI Suite (Intel® FPGA AI) supports the new AI features. The Intel® FPGA AI Suite enables push-button flow from industry standard frameworks—such as Caffe, PyTorch* , and TensorFlow* —to FPGA bitstream.

Additionally, the Intel Agilex® 5 FPGAs and SoCs also carry over the variable-precision DSP architecture from previous Intel® FPGAs with hard fixed point and IEEE 754-compliant floating point capabilities.

In fixed point mode, you can configure the DSP blocks to support signal processing with precisions from 9×9 up to 54×54:

  • Increased 9×9 multipliers count, with three 9×9 multipliers for every 18×19 multiplier
  • A pipeline register increases the maximum DSP block operating frequency and reduces the power consumption
  • Dynamically switch multiplier inputs through scanin and chainout signals
  • Compile each DSP block independently as six 9×9, dual 18×19, or single 27×27 multiply-accumulate.

The variable-precision DSP supports floating point addition, multiplication, multiply-add, and multiply-accumulate:

  • Single-precision 32-bit arithmetic FP32 floating point mode
  • Half-precision 16-bit arithmetic FP16 and FP19 floating point modes, and BFLOAT16 floating point format

With a dedicated 64-bit cascade bus, you can cascade multiple variable-precision DSP blocks to efficiently implement even higher-precision DSP functions.

Table 13.  Variable-Precision DSP Block Configurations in Intel Agilex® 5 FPGAs and SoCsThis table lists the way Intel Agilex® 5 FPGAs and SoCs accommodate the different precisions within a DSP block or by utilizing multiple DSP blocks.
Multiplier DSP Block Resource Usage Expected Application
9×9 bits

One-sixth of of a variable-precision DSP block

(One DSP block can support six 9×9)

 Low-precision fixed point
18×19 bits Half of a variable-precision DSP block Medium-precision fixed point
27×27 bits One variable-precision DSP block High-precision fixed point
19×36 bits One variable-precision DSP block with external adder Fixed point fast Fourier transform (FFT)
36×36 bits Two variable-precision DSP blocks with external adder Very high-precision fixed point
54×54 bits Four variable-precision DSP blocks with external adder Double-precision fixed point
Half-precision floating point

One variable-precision DSP block

(Contains adder for two FP16, FP19, or BFLOAT16 multipliers with one accumulator)

 Half-precision floating point
Single-precision floating point

One variable-precision DSP block

(Contains one FP32 multipliers with one accumulator)

 Single-precision floating point
AI tensor block Two sums of ten INT8×INT8 multipliers tensor fixed-point and floating-point computation Tensor dot products of 10-element vectors computation
Complex multiplication mode

One variable-precision DSP block

(16×16 ± 16×16)

 INT16 complex multiplication
13 Arithmetic density is a measure of how many dot products can fit into a 1 mm2 of silicon on any given process node.
Level Two Title