Stratix® 10 Variable Precision DSP Blocks User Guide

Download PDF
ID 683832
Date 7/08/2024
Public
Document Table of Contents
Document Table of Contents x
1. Stratix® 10 Variable Precision DSP Blocks Overview 2. Block Architecture Overview 3. Operational Mode Descriptions 4. Design Considerations 5. Stratix® 10 Variable Precision DSP Blocks Implementation Guide 6. Native Fixed Point DSP Stratix® 10 FPGA IP Core References 7. Multiply Adder IP Core References 8. ALTMULT_COMPLEX Intel® FPGA IP Core References 9. LPM_MULT Intel® FPGA IP Core References 10. Native Floating Point DSP Stratix® 10 FPGA IP References 11. LPM_DIVIDE (Divider) Intel FPGA IP Core 12. Stratix® 10 Variable Precision DSP Blocks User Guide Document Archives 13. Document Revision History for the Stratix® 10 Variable Precision DSP Blocks User Guide
1. Stratix® 10 Variable Precision DSP Blocks Overview x
1.1. Features 1.2. Supported Operational Modes in Stratix® 10 Devices 1.3. Resources
2. Block Architecture Overview x
2.1. Input Register Bank for Fixed-Point and Floating-Point Arithmetic 2.2. Pipeline Registers for Fixed-Point and Floating-Point Arithmetic 2.3. Pre-adder for Fixed-Point Arithmetic 2.4. Internal Coefficient for Fixed-Point Arithmetic 2.5. Multipliers for Fixed-Point and Floating-Point Arithmetic 2.6. Adder or Subtractor for Fixed-Point and Floating-Point Arithmetic 2.7. Accumulator, Chainout Adder, and Preload Constant for Fixed-Point Arithmetic 2.8. Systolic Register for Fixed-Point Arithmetic 2.9. Double Accumulation Register for Fixed-Point Arithmetic 2.10. Output Register Bank for Fixed-Point and Floating-Point Arithmetic 2.11. Exception Handling for Floating-Point Arithmetic
3. Operational Mode Descriptions x
3.1. Operational Modes for Fixed-Point Arithmetic 3.2. Operational Modes for Floating-Point Arithmetic
3.1. Operational Modes for Fixed-Point Arithmetic x
3.1.1. Independent Multiplier Mode 3.1.2. Multiplier Adder Sum Mode 3.1.3. Independent Complex Multiplier 3.1.4. 18 × 19 Multiplication Summed with 36-Bit Input Mode 3.1.5. Systolic FIR Mode
3.1.1. Independent Multiplier Mode x
3.1.1.1. 18 × 18 or 18 × 19 Independent Multiplier 3.1.1.2. 27 × 27 Independent Multiplier
3.1.5. Systolic FIR Mode x
3.1.5.1. Mapping Systolic Mode User View to Variable Precision Block Architecture View 3.1.5.2. 18-bit Systolic FIR Mode 3.1.5.3. 27-Bit Systolic FIR Mode
3.2. Operational Modes for Floating-Point Arithmetic x
3.2.1. Single Floating-Point Arithmetic Functions 3.2.2. Multiple Floating-Point Arithmetic Functions
3.2.1. Single Floating-Point Arithmetic Functions x
3.2.1.1. Multiplication Mode 3.2.1.2. Adder or Subtract Mode 3.2.1.3. Multiply Accumulate Mode
3.2.2. Multiple Floating-Point Arithmetic Functions x
3.2.2.1. Multiply-Add or Multiply-Subtract Mode 3.2.2.2. Vector One Mode 3.2.2.3. Vector Two Mode 3.2.2.4. Direct Vector Dot Product 3.2.2.5. Complex Multiplication
4. Design Considerations x
4.1. Internal Coefficient and Pre-Adder for Fixed-Point Arithmetic 4.2. Accumulator for Fixed-Point Arithmetic 4.3. Chainout Adder 4.4. Input Cascade for Fixed-Point Arithmetic 4.5. DSP Block Cascade Limit in Agilex™ 7 Devices
5. Stratix® 10 Variable Precision DSP Blocks Implementation Guide x
5.1. 'X' Propagation Support in Simulation
6. Native Fixed Point DSP Stratix® 10 FPGA IP Core References x
6.1. Native Fixed Point DSP Stratix® 10 FPGA IP Release Information 6.2. Supported Operational Modes 6.3. Maximum Input Data Width for Fixed-Point Arithmetic 6.4. Parameterizing Native Fixed Point DSP IP Core 6.5. Signals
6.3. Maximum Input Data Width for Fixed-Point Arithmetic x
6.3.1. Using Less Than 36-Bit Operand In 18 x 18 Plus 36 Mode Example
6.4. Parameterizing Native Fixed Point DSP IP Core x
6.4.1. Native Fixed Point DSP Stratix® 10 FPGA IP Parameters
7. Multiply Adder IP Core References x
7.1. Multiply Adder Intel FPGA IP Release Information 7.2. Features 7.3. Parameters 7.4. Signals
7.2. Features x
7.2.1. Pre-adder 7.2.2. Systolic Delay Register 7.2.3. Pre-load Constant 7.2.4. Double Accumulator
7.2.1. Pre-adder x
7.2.1.1. Pre-adder Simple Mode 7.2.1.2. Pre-adder Coefficient Mode 7.2.1.3. Pre-adder Input Mode 7.2.1.4. Pre-adder Square Mode 7.2.1.5. Pre-adder Constant Mode
7.3. Parameters x
7.3.1. General Tab 7.3.2. Extra Modes 7.3.3. Multipliers Tab 7.3.4. Preadder Tab 7.3.5. Accumulator Tab 7.3.6. Systolic/Chainout Tab 7.3.7. Pipelining Tab
8. ALTMULT_COMPLEX Intel® FPGA IP Core References x
8.1. ALTMULT_COMPLEX Intel FPGA IP Release Information 8.2. Features 8.3. Complex Multiplication 8.4. Parameters 8.5. Signals
9. LPM_MULT Intel® FPGA IP Core References x
9.1. LPM_MULT Intel FPGA IP Release Information 9.2. Features 9.3. Parameters 9.4. Signals
9.3. Parameters x
9.3.1. General Tab 9.3.2. General 2 Tab 9.3.3. Pipelining Tab
10. Native Floating Point DSP Stratix® 10 FPGA IP References x
10.1. Native Floating Point DSP Stratix® 10 FPGA IP Release Information 10.2. Native Floating Point DSP Stratix® 10 FPGA IP Core Supported Operational Modes 10.3. Parameterizing the Native Floating Point DSP Stratix® 10 FPGA IP 10.4. Native Floating Point DSP Stratix® 10 FPGA IP Core Signals
10.3. Parameterizing the Native Floating Point DSP Stratix® 10 FPGA IP x
10.3.1. Native Floating Point DSP Stratix® 10 FPGA IP Parameters
11. LPM_DIVIDE (Divider) Intel FPGA IP Core x
11.1. LPM_DIVIDE Intel® FPGA IP Release Information 11.2. Features 11.3. Verilog HDL Prototype 11.4. VHDL Component Declaration 11.5. VHDL LIBRARY_USE Declaration 11.6. Ports 11.7. Parameters
11.7. Parameters x
11.7.1. General Tab 11.7.2. General1 Tab
1. Stratix® 10 Variable Precision DSP Blocks Overview
2. Block Architecture Overview
3. Operational Mode Descriptions
4. Design Considerations
5. Stratix® 10 Variable Precision DSP Blocks Implementation Guide
6. Native Fixed Point DSP Stratix® 10 FPGA IP Core References
7. Multiply Adder IP Core References
8. ALTMULT_COMPLEX Intel® FPGA IP Core References
9. LPM_MULT Intel® FPGA IP Core References
10. Native Floating Point DSP Stratix® 10 FPGA IP References
11. LPM_DIVIDE (Divider) Intel FPGA IP Core
12. Stratix® 10 Variable Precision DSP Blocks User Guide Document Archives
13. Document Revision History for the Stratix® 10 Variable Precision DSP Blocks User Guide

7.2.2. Systolic Delay Register

In a systolic architecture, the input data is fed into a cascade of registers acting as a data buffer. Each register delivers an input sample to a multiplier where it is multiplied by the respective coefficient. The chain adder stores the gradually combined results from the multiplier and the previously registered result from the chainin[] input port to form the final result. Each multiply-add element must be delayed by a single cycle so that the results synchronize appropriately when added together. Each successive delay is used to address both the coefficient memory and the data buffer of their respective multiply-add elements. For example, a single delay for the second multiply add element, two delays for the third multiply-add element, and so on.

Figure 38. Systolic Registers


x(t) represents the results from a continuous stream of input samples and y(t) represents the summation of a set of input samples, and in time, multiplied by their respective coefficients. Both the input and output results flow from left to right. The c(0) to c(N-1) denotes the coefficients. The systolic delay registers are denoted by S-1, whereas the –1 represents a single clock delay. Systolic delay registers are added at the inputs and outputs for pipelining in a way that ensures the results from the multiplier operand and the accumulated sums stay in sync. This processing element is replicated to form a circuit that computes the filtering function. This function is expressed in the following equation.



N represents the number of cycles of data that has entered into the accumulator, y(t) represents the output at time t, A(t) represents the input at time t, and B(i) are the coefficients. The t and i in the equation correspond to a particular instant in time, so to compute the output sample y(t) at time t, a group of input samples at N different points in time, or A(n), A(n-1), A(n-2), … A(n-N+1) is required. The group of N input samples are multiplied by N coefficients and summed together to form the final result y.

The systolic register architecture is available only for sum-of-2 and sum-of-4 modes.

The following figure shows the systolic delay register implementation of 2 multipliers.

Figure 39. Systolic Delay Register Implementation of 2 Multipliers


The sum of two multipliers is expressed in the following equation.



The following figure shows the systolic delay register implementation of 4 multipliers.

Figure 40. Systolic Delay Register Implementation of 4 Multipliers


The sum of four multipliers is expressed in the following equation.



The following lists the advantages of systolic register implementation:

  • Reduces DSP resource usage
  • Enables efficient mapping in the DSP block using the chain adder structure
Related Information
Level Two Title