Intel® Agilex™ Variable Precision DSP Blocks User Guide

Download PDF
ID 683037
Date 11/17/2022
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents
Document Table of Contents x
1. Intel® Agilex™ Variable Precision DSP Blocks Overview 2. Intel® Agilex™ Variable Precision DSP Blocks Architecture 3. Intel® Agilex™ Variable Precision DSP Blocks Operational Modes 4. Intel® Agilex™ Variable Precision DSP Blocks Design Considerations 5. Native Fixed Point DSP Intel® Agilex™ FPGA IP Core References 6. Multiply Adder Intel® FPGA IP Core References 7. ALTMULT_COMPLEX Intel® FPGA IP Core Reference 8. LPM_MULT Intel® FPGA IP Core References 9. LPM_DIVIDE (Divider) Intel FPGA IP Core 10. Native Floating Point DSP Intel® Agilex™ FPGA IP References 11. Intel® Agilex™ Variable Precision DSP Blocks User Guide Archives 12. Document Revision History for the Intel® Agilex™ Variable Precision DSP Blocks User Guide
1. Intel® Agilex™ Variable Precision DSP Blocks Overview x
1.1. Features 1.2. Supported Operational Modes in Intel® Agilex™ Devices
1.2. Supported Operational Modes in Intel® Agilex™ Devices x
1.2.1. Fixed-point Arithmetic 1.2.2. Floating-point Arithmetic
2. Intel® Agilex™ Variable Precision DSP Blocks Architecture x
2.1. Fixed-point Arithmetic 2.2. Floating-point Arithmetic
2.1. Fixed-point Arithmetic x
2.1.1. Input Register Bank for Fixed-point Arithmetic 2.1.2. Pipeline Registers for Fixed-point Arithmetic 2.1.3. Pre-adder for Fixed-point Arithmetic 2.1.4. Internal Coefficient for Fixed-point Arithmetic 2.1.5. Multipliers for Fixed-point Arithmetic 2.1.6. Adder or Subtractor for Fixed-point Arithmetic 2.1.7. Accumulator, Chainout Adder, and Preload Constant for Fixed-point Arithmetic 2.1.8. Systolic Register for Fixed-point Arithmetic 2.1.9. Double Accumulation Register for Fixed-point Arithmetic 2.1.10. Output Register Bank for Fixed-point Arithmetic
2.1.7. Accumulator, Chainout Adder, and Preload Constant for Fixed-point Arithmetic x
2.1.7.1. Dynamic Chainout
2.2. Floating-point Arithmetic x
2.2.1. Input Register Bank for Floating-point Arithmetic 2.2.2. Pipeline Registers for Floating-point Arithmetic 2.2.3. Multipliers for Floating-point Arithmetic 2.2.4. Adder or Subtractor for Floating-point Arithmetic 2.2.5. Output Register Bank for Floating-point Arithmetic 2.2.6. Exception Handling for Floating-point Arithmetic
3. Intel® Agilex™ Variable Precision DSP Blocks Operational Modes 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. 8 x 8 (Unsigned) or 9 x 9 (Signed) Sum of 4 Mode 3.1.3. Multiplier Adder Sum Mode 3.1.4. Independent Complex Multiplier 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
3.1.4. Independent Complex Multiplier x
3.1.4.1. 18 × 19 Multiplication Summed with 36-Bit Input Mode
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. FP32 Single-precision Floating-point Arithmetic Functions 3.2.2. FP16 Half-precision Floating-point Arithmetic Functions 3.2.3. Multiple Floating-point Variable DSP Blocks Functions
3.2.1. FP32 Single-precision Floating-point Arithmetic Functions x
3.2.1.1. FP32 Multiplication Mode 3.2.1.2. Adder or Subtract Mode 3.2.1.3. Multiply Accumulate Mode 3.2.1.4. FP32 Vector One Mode 3.2.1.5. FP32 Vector Two Mode
3.2.2. FP16 Half-precision Floating-point Arithmetic Functions x
3.2.2.1. FP16 Supported Precision Formats 3.2.2.2. Sum of Two FP16 Multiplication Mode 3.2.2.3. Sum of Two FP16 Multiplication with FP32 Addition Mode 3.2.2.4. Sum of Two FP16 Multiplication with Accumulation Mode 3.2.2.5. FP16 Vector One Mode 3.2.2.6. FP16 Vector Two Mode 3.2.2.7. FP16 Vector Three Mode
3.2.3. Multiple Floating-point Variable DSP Blocks Functions x
3.2.3.1. Multiply-Add or Multiply-Subtract Mode 3.2.3.2. Direct Vector Dot Product 3.2.3.3. Complex Multiplication
4. Intel® Agilex™ Variable Precision DSP Blocks Design Considerations x
4.1. Fixed-point Arithmetic 4.2. Floating-point Arithmetic 4.3. DSP Block Cascade Limit in Intel® Agilex™ Devices
4.1. Fixed-point Arithmetic x
4.1.1. Configurations for Input, Pipeline, and Output Registers 4.1.2. Internal Coefficient and Pre-Adder for Fixed-point Arithmetic 4.1.3. Accumulator for Fixed-point Arithmetic 4.1.4. Input Cascade for Fixed-point Arithmetic 4.1.5. Chainout Adder
4.1.1. Configurations for Input, Pipeline, and Output Registers x
4.1.1.1. Restrictions for Input Registers 4.1.1.2. Restrictions for Pipeline Registers 4.1.1.3. Supported Register Configurations per Operation Modes
4.1.4. Input Cascade for Fixed-point Arithmetic x
4.1.4.1. Dynamic Scanin
4.2. Floating-point Arithmetic x
4.2.1. Configurations for Input, Pipeline, and Output Registers 4.2.2. Chainout Adder
4.2.1. Configurations for Input, Pipeline, and Output Registers x
4.2.1.1. FP32 Operation Modes Supported Register Configurations 4.2.1.2. FP16 Operation Mode Supported Register Configurations
5. Native Fixed Point DSP Intel® Agilex™ FPGA IP Core References x
5.1. Native Fixed Point DSP Intel® Agilex™ FPGA IP Release Information 5.2. Supported Operational Modes 5.3. Maximum Input Data Width for Fixed-point Arithmetic 5.4. Maximum Output Data Width for Fixed-point Arithmetic 5.5. Parameterizing Native Fixed Point DSP IP 5.6. Native Fixed Point DSP Intel® Agilex™ FPGA IP Signals
5.3. Maximum Input Data Width for Fixed-point Arithmetic x
5.3.1. Using Less Than 36-Bit Operand In 18 x 18 Plus 36 Mode Example
5.5. Parameterizing Native Fixed Point DSP IP x
5.5.1. Operation Mode Tab 5.5.2. Input Cascade Tab 5.5.3. Pre-adder Tab 5.5.4. Internal Coefficient Tab 5.5.5. Accumulator/Output Chaining 5.5.6. Pipelining 5.5.7. Clear Signal
5.6. Native Fixed Point DSP Intel® Agilex™ FPGA IP Signals x
5.6.1. 9 × 9 Sum of 4 Mode Signals 5.6.2. 18 × 18 Full Mode Signals 5.6.3. 18 × 18 Sum of Two Mode Signals 5.6.4. 18 × 18 Plus 36 Mode Signals 5.6.5. 18 × 18 Systolic Mode Signals 5.6.6. 27 × 27 Mode Signals
6. Multiply Adder Intel® FPGA IP Core References x
6.1. Multiply Adder Intel® FPGA IP Release Information 6.2. Features 6.3. Parameters 6.4. Signals
6.2. Features x
6.2.1. Pre-adder 6.2.2. Systolic Delay Register 6.2.3. Pre-load Constant 6.2.4. Double Accumulator
6.2.1. Pre-adder x
6.2.1.1. Pre-adder Simple Mode 6.2.1.2. Pre-adder Coefficient Mode 6.2.1.3. Pre-adder Input Mode 6.2.1.4. Pre-adder Square Mode 6.2.1.5. Pre-adder Constant Mode
6.3. Parameters x
6.3.1. General Tab 6.3.2. Extra Modes 6.3.3. Multipliers Tab 6.3.4. Preadder Tab 6.3.5. Accumulator Tab 6.3.6. Systolic/Chainout Tab 6.3.7. Pipelining Tab
7. ALTMULT_COMPLEX Intel® FPGA IP Core Reference x
7.1. ALTMULT_COMPLEX Intel® FPGA IP Release Information 7.2. Features 7.3. Complex Multiplication 7.4. Parameters 7.5. Signals
8. LPM_MULT Intel® FPGA IP Core References x
8.1. LPM_MULT Intel® FPGA IP Release Information 8.2. Features 8.3. Parameters 8.4. Signals
8.3. Parameters x
8.3.1. General Tab 8.3.2. General 2 Tab 8.3.3. Pipelining Tab
9. LPM_DIVIDE (Divider) Intel FPGA IP Core x
9.1. LPM_DIVIDE Intel® FPGA IP Release Information 9.2. Features 9.3. Verilog HDL Prototype 9.4. VHDL Component Declaration 9.5. VHDL LIBRARY_USE Declaration 9.6. Ports 9.7. Parameters
9.7. Parameters x
9.7.1. General Tab 9.7.2. General1 Tab
10. Native Floating Point DSP Intel® Agilex™ FPGA IP References x
10.1. Native Floating Point DSP Intel® Agilex™ FPGA IP Release Information 10.2. Native Floating Point DSP Intel® Agilex™ FPGA IP Core Supported Operational Modes 10.3. Parameterizing the Native Floating Point DSP Intel® Agilex™ FPGA IP 10.4. Native Floating Point DSP Intel® Agilex™ FPGA IP Core Signals
10.3. Parameterizing the Native Floating Point DSP Intel® Agilex™ FPGA IP x
10.3.1. General Tab 10.3.2. Registers Tab
10.4. Native Floating Point DSP Intel® Agilex™ FPGA IP Core Signals x
10.4.1. FP32 Multiplication Mode Signals 10.4.2. FP32 Addition or Subtraction Mode Signals 10.4.3. FP32 Multiplication with Addition or Subtraction Mode Signals 10.4.4. FP32 Multiplication with Accumulation Mode Signals 10.4.5. FP32 Vector One and Vector Two Modes Signals 10.4.6. Sum of Two FP16 Multiplication Mode Signals 10.4.7. Sum of Two FP16 Multiplication with FP32 Addition Mode Signals 10.4.8. Sum of Two FP16 Multiplication with Accumulation Mode Signals 10.4.9. FP16 Vector One and Vector Two Modes Signals 10.4.10. FP16 Vector Three Mode Signals
1. Intel® Agilex™ Variable Precision DSP Blocks Overview
2. Intel® Agilex™ Variable Precision DSP Blocks Architecture
3. Intel® Agilex™ Variable Precision DSP Blocks Operational Modes
4. Intel® Agilex™ Variable Precision DSP Blocks Design Considerations
5. Native Fixed Point DSP Intel® Agilex™ FPGA IP Core References
6. Multiply Adder Intel® FPGA IP Core References
7. ALTMULT_COMPLEX Intel® FPGA IP Core Reference
8. LPM_MULT Intel® FPGA IP Core References
9. LPM_DIVIDE (Divider) Intel FPGA IP Core
10. Native Floating Point DSP Intel® Agilex™ FPGA IP References
11. Intel® Agilex™ Variable Precision DSP Blocks User Guide Archives
12. Document Revision History for the Intel® Agilex™ Variable Precision DSP Blocks User Guide

10.2. Native Floating Point DSP Intel® Agilex™ FPGA IP Core Supported Operational Modes

Table 107.  Operational Modes Supported by Native Floating Point DSP Intel® Agilex™ FPGA IP Core
Operational Modes Description Supported Exception Flags
FP32 multiplication mode

This mode performs single precision multiplication operation.

This mode applies the following equation:
  • fp32_result = fp32_result = fp32_mult_a*fp32_mult_b
  • fp32_mult_overflow
  • fp32_mult_underflow
  • fp32_mult_inexact
  • fp32_mult_invalid
FP32 addition or subtraction mode This mode performs single precision addition or subtraction operation.
This mode applies the following equations:
  • fp32_result = fp32_adder_b+fp32_adder_a
  • fp32_result = fp32_adder_b-fp32_adder_a
  • fp32_adder_overflow
  • fp32_adder_underflow
  • fp32_adder_inexact
  • fp32_adder_invalid
FP32 multiplication with addition or subtraction mode

This mode performs single precision multiplication, followed by addition or subtraction operations.

This mode applies the following equations:
  • When chainin feature is enabled:
    • fp32_result = (fp32_mult_a*fp32_mult_b) + fp32_chainin
    • fp32_result = (fp32_mult_a*fp32_mult_b) - fp32_chainin
  • When chainin feature is disabled:
    • fp32_result = (fp32_mult_a*fp32_mult_b) + fp32_adder_a
    • fp32_result = (fp32_mult_a*fp32_mult_b) - fp32_adder_a
  • fp32_mult_overflow
  • fp32_mult_underflow
  • fp32_mult_inexact
  • fp32_mult_invalid
  • fp32_adder_overflow
  • fp32_adder_underflow
  • fp32_adder_inexact
  • fp32_adder_invalid
FP32 multiplication with accumulation mode

This mode performs floating-point multiplication followed by floating-point addition or subtraction with the previous multiplication result.

This mode applies the following equations:
  • When accumulate signal is driven high:
    • fp32_result(t) = [fp32_mult_a(t)*fp32_mult_b(t)] + fp32_result(t-1)
    • fp32_result(t) = [fp32_mult_a(t)*fp32_mult_b(t) - fp32_result(t-1)
  • When accumulate signal is driven low:
    • fp32_result = fp32_mult_a*fp32_mult_b.
FP32 vector one mode

This mode performs floating-point multiplication followed by floating-point addition or subtraction with the chainin input from the previous variable DSP Block.

This mode applies the following equations:
  • When chainin feature is enabled:
    • fp32_result = (fp32_mult_a * fp32_mult_b) + fp32_chainin, fp32_chainout = fp32_adder_a
    • fp32_result = (fp32_mult_a * fp32_mult_b) - fp32_chainin, fp32_chainout = fp32_adder_a
  • When chainin feature is disabled:
    • fp32_result = fp32_mult_a * fp32_mult_b, fp32_chainout = fp32_adder_a
FP32 vector two mode This mode performs floating-point multiplication where the multiplication result is directly fed to chainout. The chainin input from the previous variable DSP Block is then added or subtracted from input Ax as the output result.

This mode applies the following equations:

  • When chainin feature is enabled:
    • fp32_result = fp32_adder_a + fp32_chainin, fp32_chainout = fp32_mult_a * fp32_mult_b
    • fp32_result = fp32_adder_a - fp32_chainin, fp32_chainout = fp32_mult_a * fp32_mult_b
  • When chainin feature is disabled:
    • fp32_result = fp32_adder_a, fp32_chainout = fp32_mult_a * fp32_mult_b
Sum of two FP16 multiplication mode

This mode performs a summation of two half-precision multiplication and provide a single-precision result.

This mode applies the following equations:
  • fp32_result = (fp16_mult_top_a*fp16_mult_top_b) + (fp16_mult_bot_a*fp16_mult_bot_b)
  • fp32_result = (fp16_mult_top_a*fp16_mult_top_b) - (fp16_mult_bot_a*fp16_mult_bot_b)
Exception flags supported in flushed and bfloat16 formats:
  • fp16_mult_top_invalid
  • fp16_mult_top_inexact
  • fp16_mult_top_overflow
  • fp16_mult_top_underflow
  • fp16_mult_bot_invalid
  • fp16_mult_bot_inexact
  • fp16_mult_bot_overflow
  • fp16_mult_bot_underflow
  • fp16_adder_invalid
  • fp16_adder_inexact
  • fp16_adder_overflow
  • fp16_adder_underflow
Exception flags supported in extended format:
  • fp16_mult_top_invalid
  • fp16_mult_top_inexact
  • fp16_mult_top_infinite
  • fp16_mult_top_zero
  • fp16_mult_bot_invalid
  • fp16_mult_bot_inexact
  • fp16_mult_bot_infinite
  • fp16_mult_bot_zero
  • fp16_adder_invalid
  • fp16_adder_inexact
  • fp16_adder_infinite
  • fp16_adder_zero
Sum of two FP16 multiplication with FP32 addition mode

This mode performs a summation of two half-precision multiplication and provide a single-precision result.

This mode applies the following equations:
  • fp32_result = (fp16_mult_top_a*fp16_mult_top_b) + (fp16_mult_bot_a*fp16_mult_bot_b) - fp32_adder_a
  • fp32_result = (fp16_mult_top_a*fp16_mult_top_b) - (fp16_mult_bot_a*fp16_mult_bot_b) - fp32_adder_a
  • fp32_result = (fp16_mult_top_a*fp16_mult_top_b) + (fp16_mult_bot_a*fp16_mult_bot_b) + fp32_adder_a
  • fp32_result = (fp16_mult_top_a*fp16_mult_top_b) - (fp16_mult_bot_a*fp16_mult_bot_b) + fp32_adder_a
Exception flags supported in flushed and bfloat16 formats:
  • fp16_mult_top_invalid
  • fp16_mult_top_inexact
  • fp16_mult_top_overflow
  • fp16_mult_top_underflow
  • fp16_mult_bot_invalid
  • fp16_mult_bot_inexact
  • fp16_mult_bot_overflow
  • fp16_mult_bot_underflow
  • fp16_adder_invalid
  • fp16_adder_inexact
  • fp16_adder_overflow
  • fp16_adder_underflow
  • fp32_adder_invalid
  • fp32_adder_inexact
  • fp32_adder_overflow
  • fp32_adder_underflow
Exception flags supported in extended format:
  • fp16_mult_top_invalid
  • fp16_mult_top_inexact
  • fp16_mult_top_infinite
  • fp16_mult_top_zero
  • fp16_mult_bot_invalid
  • fp16_mult_bot_inexact
  • fp16_mult_bot_infinite
  • fp16_mult_bot_zero
  • fp16_adder_invalid
  • fp16_adder_inexact
  • fp16_adder_infinite
  • fp16_adder_zero
  • fp32_adder_invalid
  • fp32_adder_inexact
  • fp32_adder_overflow
  • fp32_adder_underflow
Sum of two FP16 multiplication with accumulation mode

This mode performs a summation of two half-precision multiplication and accumulate the value into single-precision format.

This mode applies the following equations:
  • When accumulate signal is driven high:
    • fp32_result (t) = [fp16_mult_top_a(t) * fp16_mult_top_b(t)] + [fp16_mult_bot_a(t) * fp16_mult_bot_b(t)] + fp32_result(t-1)
    • fp32_result (t) = [fp16_mult_top_a(t) * fp16_mult_top_b(t)] - [fp16_mult_bot_a(t) * fp16_mult_bot_b(t)] + fp32_result(t-1)
    • fp32_result (t) = [fp16_mult_top_a(t) * fp16_mult_top_b(t)] + [fp16_mult_bot_a(t) * fp16_mult_bot_b(t)] - fp32_result(t-1)
    • fp32_result (t) = [fp16_mult_top_a(t) * fp16_mult_top_b(t)] - [fp16_mult_bot_a(t) * fp16_mult_bot_b(t)] - fp32_result(t-1)
  • When accumulate signal is driven low:
    • fp32_result = [fp16_mult_top_a * fp16_mult_top_b] + [fp16_mult_bot_a * fp16_mult_bot_b]
    • fp32_result = [fp16_mult_top_a * fp16_mult_top_b] - [fp16_mult_bot_a * fp16_mult_bot_b]
Exception flags supported in flushed and bfloat16 formats:
  • fp16_mult_top_invalid
  • fp16_mult_top_inexact
  • fp16_mult_top_overflow
  • fp16_mult_top_underflow
  • fp16_mult_bot_invalid
  • fp16_mult_bot_inexact
  • fp16_mult_bot_overflow
  • fp16_mult_bot_underflow
  • fp16_adder_invalid
  • fp16_adder_inexact
  • fp16_adder_overflow
  • fp16_adder_underflow
  • fp32_adder_invalid
  • fp32_adder_inexact
  • fp32_adder_overflow
  • fp32_adder_underflow
Exception flags supported in extended format:
  • fp16_mult_top_invalid
  • fp16_mult_top_inexact
  • fp16_mult_top_infinite
  • fp16_mult_top_zero
  • fp16_mult_bot_invalid
  • fp16_mult_bot_inexact
  • fp16_mult_bot_infinite
  • fp16_mult_bot_zero
  • fp16_adder_invalid
  • fp16_adder_inexact
  • fp16_adder_infinite
  • fp16_adder_zero
  • fp32_adder_invalid
  • fp32_adder_inexact
  • fp32_adder_overflow
  • fp32_adder_underflow
FP16 vector one mode

This mode performs a summation of two half-precision multiplications with the chainin input from the previous variable DSP Block. The output is a single-precision floating-point value which is fed into chainout.

This mode applies the following equation:
  • When chainin feature is enabled:
    • fp32_result = (fp16_mult_top_a * fp16_mult_top_b) + (fp16_mult_bot_a * fp16_mult_bot_b) + fp32_chainin, fp32_chainout = fp32_adder_a
    • fp32_result = (fp16_mult_top_a * fp16_mult_top_b) - (fp16_mult_bot_a * fp16_mult_bot_b) + fp32_chainin, fp32_chainout = fp32_adder_a
    • fp32_result = (fp16_mult_top_a * fp16_mult_top_b) + (fp16_mult_bot_a * fp16_mult_bot_b) - fp32_chainin, fp32_chainout = fp32_adder_a
    • fp32_result = (fp16_mult_top_a * fp16_mult_top_b) - (fp16_mult_bot_a * fp16_mult_bot_b) - fp32_chainin, fp32_chainout = fp32_adder_a
  • When chainin feature is disabled:
    • fp32_result = (fp16_mult_top_a * fp16_mult_top_b) + (fp16_mult_bot_a * fp16_mult_bot_b), fp32_chainout = fp32_adder_a
    • fp32_result = (fp16_mult_top_a * fp16_mult_top_b) - (fp16_mult_bot_a * fp16_mult_bot_b), fp32_chainout = fp32_adder_a
Exception flags supported in flushed and bfloat16 formats:
  • fp16_mult_top_invalid
  • fp16_mult_top_inexact
  • fp16_mult_top_overflow
  • fp16_mult_top_underflow
  • fp16_mult_bot_invalid
  • fp16_mult_bot_inexact
  • fp16_mult_bot_overflow
  • fp16_mult_bot_underflow
  • fp16_adder_invalid
  • fp16_adder_inexact
  • fp16_adder_overflow
  • fp16_adder_underflow
  • fp32_adder_invalid
  • fp32_adder_inexact
  • fp32_adder_overflow
  • fp32_adder_underflow
Exception flags supported in extended format:
  • fp16_mult_top_invalid
  • fp16_mult_top_inexact
  • fp16_mult_top_infinite
  • fp16_mult_top_zero
  • fp16_mult_bot_invalid
  • fp16_mult_bot_inexact
  • fp16_mult_bot_infinite
  • fp16_mult_bot_zero
  • fp16_adder_invalid
  • fp16_adder_inexact
  • fp16_adder_infinite
  • fp16_adder_zero
  • fp32_adder_invalid
  • fp32_adder_inexact
  • fp32_adder_overflow
  • fp32_adder_underflow
FP16 vector two mode

This mode performs a summation of two half precision multiplication and fed to chainout. The chainin input from the previous variable DSP Block is then added or subtracted from input fp32_adder_a as the output result.

This mode applies the following equation:
  • When chainin feature is enabled:
    • fp32_result = fp32_adder_a + fp32_chainin, fp32_chainout = (fp16_mult_top_a * fp16_mult_top_b) + (fp16_mult_bot_a * fp16_mult_bot_b)
    • fp32_result = fp32_adder_a - fp32_chainin, fp32_chainout = (fp16_mult_top_a * fp16_mult_top_b) + (fp16_mult_bot_a * fp16_mult_bot_b)
    • fp32_result = fp32_adder_a + fp32_chainin, fp32_chainout = (fp16_mult_top_a * fp16_mult_top_b) - (fp16_mult_bot_a * fp16_mult_bot_b)
    • fp32_result = fp32_adder_a - fp32_chainin, fp32_chainout = (fp16_mult_top_a * fp16_mult_top_b) - (fp16_mult_bot_a * fp16_mult_bot_b)
  • When chainin feature is disabled:
    • fp32_result = fp32_adder_a, fp32_chainout = (fp16_mult_top_a * fp16_mult_top_b) + (fp16_mult_bot_a * fp16_mult_bot_b)
    • fp32_result = fp32_adder_a, fp32_chainout = (fp16_mult_top_a * fp16_mult_top_b) - (fp16_mult_bot_a * fp16_mult_bot_b)
Exception flags supported in flushed and bfloat16 formats:
  • fp16_mult_top_invalid
  • fp16_mult_top_inexact
  • fp16_mult_top_overflow
  • fp16_mult_top_underflow
  • fp16_mult_bot_invalid
  • fp16_mult_bot_inexact
  • fp16_mult_bot_overflow
  • fp16_mult_bot_underflow
  • fp16_adder_invalid
  • fp16_adder_inexact
  • fp16_adder_overflow
  • fp16_adder_underflow
  • fp32_adder_invalid
  • fp32_adder_inexact
  • fp32_adder_overflow
  • fp32_adder_underflow
Exception flags supported in extended format:
  • fp16_mult_top_invalid
  • fp16_mult_top_inexact
  • fp16_mult_top_infinite
  • fp16_mult_top_zero
  • fp16_mult_bot_invalid
  • fp16_mult_bot_inexact
  • fp16_mult_bot_infinite
  • fp16_mult_bot_zero
  • fp16_adder_invalid
  • fp16_adder_inexact
  • fp16_adder_infinite
  • fp16_adder_zero
  • fp32_adder_invalid
  • fp32_adder_inexact
  • fp32_adder_overflow
  • fp32_adder_underflow
FP16 vector three

This mode performs a single-precision accumulation and a summation of two half-precision multiplications.

This mode applies the following equation:
  • When accumulate is driven high:

    • fp32_result(t) = fp32_adder_a(t) + fp32_result(t-1), fp32_chainout = {fp16_mult_top_a * fp16_mult_top_b} + {fp16_mult_bot_a * fp16_mult_bot_b}

    • fp32_result(t) = fp32_adder_a(t) - fp32_result(t-1), fp32_chainout = {fp16_mult_top_a * fp16_mult_top_b} + {fp16_mult_bot_a * fp16_mult_bot_b}

    • fp32_result(t) = fp32_adder_a(t) + fp32_result(t-1), fp32_chainout = {fp16_mult_top_a * fp16_mult_top_b} - {fp16_mult_bot_a * fp16_mult_bot_b}

    • fp32_result(t) = fp32_adder_a(t) - fp32_result(t-1), fp32_chainout = {fp16_mult_top_a * fp16_mult_top_b} - {fp16_mult_bot_a * fp16_mult_bot_b}
  • When accumulate is driven low:
    • fp32_result = fp32_adder_a, fp32_chainout = {fp16_mult_top_a * fp16_mult_top_b} + {fp16_mult_bot_a * fp16_mult_bot_b}
    • fp32_result = fp32_adder_a, fp32_chainout = {fp16_mult_top_a * fp16_mult_top_b} - {fp16_mult_bot_a * fp16_mult_bot_b}
Exception flags supported in flushed and bfloat16 formats:
  • fp16_mult_top_invalid
  • fp16_mult_top_inexact
  • fp16_mult_top_overflow
  • fp16_mult_top_underflow
  • fp16_mult_bot_invalid
  • fp16_mult_bot_inexact
  • fp16_mult_bot_overflow
  • fp16_mult_bot_underflow
  • fp16_adder_invalid
  • fp16_adder_inexact
  • fp16_adder_overflow
  • fp16_adder_underflow
  • fp32_adder_invalid
  • fp32_adder_inexact
  • fp32_adder_overflow
  • fp32_adder_underflow
Exception flags supported in extended format:
  • fp16_mult_top_invalid
  • fp16_mult_top_inexact
  • fp16_mult_top_infinite
  • fp16_mult_top_zero
  • fp16_mult_bot_invalid
  • fp16_mult_bot_inexact
  • fp16_mult_bot_infinite
  • fp16_mult_bot_zero
  • fp16_adder_invalid
  • fp16_adder_inexact
  • fp16_adder_infinite
  • fp16_adder_zero
  • fp32_adder_invalid
  • fp32_adder_inexact
  • fp32_adder_overflow
  • fp32_adder_underflow
Level Two Title