FPGA AI Suite: Compiler Reference Manual

Download PDF
ID 768972
Date 3/29/2024
Public
Document Table of Contents
Document Table of Contents x
1. FPGA AI Suite Compiler Reference Manual 2. About the FPGA AI Suite Compiler 3. FPGA AI Suite Compiler Use Modes 4. FPGA AI Suite Compiler Command Line Options A. FPGA AI Suite Compiler Reference Manual Archives B. FPGA AI Suite Compiler Reference Manual Document Revision History
3. FPGA AI Suite Compiler Use Modes x
3.1. Compiling a Graph 3.2. Estimating the Performance of a Graph 3.3. Estimating the Performance of a Partition of a Graph 3.4. Estimating the Area of an Architecture 3.5. Generating an Optimized Architecture
3.5. Generating an Optimized Architecture x
3.5.1. Generating an Architecture for Highest Performance 3.5.2. Generating an Architecture Optimized for a Frame Rate Target Value
4. FPGA AI Suite Compiler Command Line Options x
4.1. Inputs (dla_compiler Command Options) 4.2. Outputs (dla_compiler Command Options) 4.3. Reporting (dla_compiler Command Options) 4.4. Compilation Options (dla_compiler Command Options) 4.5. Architecture Options (dla_compiler Command Options) 4.6. Architecture Optimizer Options (dla_compiler Command Options) 4.7. Analyzer Tool Options (dla_compiler Command Options) 4.8. Miscellaneous Options (dla_compiler Command Options) 4.9. Input File Formats for dla_compiler Command 4.10. Generation of Output File Formats 4.11. The Partitioning Table Report
1. FPGA AI Suite Compiler Reference Manual
2. About the FPGA AI Suite Compiler
3. FPGA AI Suite Compiler Use Modes
4. FPGA AI Suite Compiler Command Line Options
A. FPGA AI Suite Compiler Reference Manual Archives
B. FPGA AI Suite Compiler Reference Manual Document Revision History

4.3. Reporting (dla_compiler Command Options)

When a graph is compiled, the following CSV files are created by default to provide additional information about the input and output tensors and how they map to FPGA AI Suite IP tensors: You can disable the generation of these reports with the --fno-transform-tables command option.

input_transform_dump Report

The input_transform_dump file is a CSV file with input tensor dimension, striding, padding, and precision. The input_transform_dump report is provided in a file named input_transform_dump_<graph_name>.csv

An example of a simple output CSV file created from an AlexNet graph is as follows:

Table 2.  (input_transform_dump_AlexNet.csv)
 

depth

height

width

height stride

width stride

top pad

left pad

Bot pad

right pad

precision

Input num 0

48

57

57

4

4

0

0

1

1

fp16

Additional information about the initial scale and shift performed on the tensor is included separately.

input_transform_mapping Report

The input_transform_mapping file is a CSV file that shows the element-wise mapping of the logical input tensor elements to the FPGA AI Suite input tensor format. The input_transform_mapping report is provided in a file named input_transform_mapping_<graph_name>.csv.

The transform mapping report has columns corresponding to the offset and subscript indices for the logical input tensor elements, and the corresponding elements in the transformed FPGA AI Suite input tensor.

For more information about contents of the input transform mapping report, refer to "Input Transform Mapping" in the FPGA AI Suite IP Reference Manual .

output_transform_dump Report

The output_transform_dump file is a CSV file with output tensor offsets and dimension (batch, depth, height, width). Output transforms tables are named output_transform_dump_<Graph_name>.csv

An example of a simple output CSV file created from an AlexNet graph is as follows:

Table 3.  (output_transform_dump_AlexNet.csv)
 

batch dim

depth dim

height dim

width dim

batch offset

depth offset

height offset

width offset

output 0 tensor 0

1

1000

1

1

0

0

0

0

output_transform_mapping Report

The report is a CSV file that show the element-wise mapping of the FPGA AI Suite output tensor format to the logical output tensor. The output_transform_mapping report is provided in a file named output_transform_mapping_<graph_name>.csv.

This mapping in the output_transform_mapping report is the inverse of the input mapping described in the input_transform_mapping report.

For more information about the output tensor format, refer to "Output Tensor In-Memory Format" in the FPGA AI Suite IP Reference Manual .

--fno-transform-tables Option

Syntax
--fno-transform-tables
Description
This option disables the creation of the CSV report files.
Parsed As
(bool)
Level Two Title