FPGA AI Suite: IP Reference Manual

Download PDF
ID 768974
Date 3/29/2024
Public
Document Table of Contents
Document Table of Contents x
1. FPGA AI Suite IP Reference Manual 2. About the FPGA AI Suite IP 3. FPGA AI Suite IP Generation Utility 4. FPGA AI Suite Ahead-of-Time Splitter Utility 5. CSR Map and Descriptor Queue A. FPGA AI Suite IP Reference Manual Archives B. FPGA AI Suite IP Reference Manual Document Revision History
2. About the FPGA AI Suite IP x
2.1. Supported Models 2.2. Model Performance 2.3. FPGA AI Suite Layer / Primitive Ranges 2.4. FPGA AI Suite IP Block Configuration 2.5. IP Block Interfaces
2.1. Supported Models x
2.1.1. MobileNet V2 differences between Caffe and TensorFlow models
2.2. Model Performance x
2.2.1. Throughput on the MobileNetV1 model (and other very fast models)
2.4. FPGA AI Suite IP Block Configuration x
2.4.1. Architecture Description File Format for Instance Parameterization 2.4.2. Architecture Description File Parameters
2.4.2. Architecture Description File Parameters x
2.4.2.1. Parameter Group: Global Parameters 2.4.2.2. Parameter Group: activation 2.4.2.3. Parameter Group: pe_array 2.4.2.4. Parameter Group: pool 2.4.2.5. Parameter Group: depthwise 2.4.2.6. Module: softmax 2.4.2.7. Parameter Group: dma 2.4.2.8. Parameter Group: xbar 2.4.2.9. Parameter Group: filter_scratchpad 2.4.2.10. Parameter Group: config_network 2.4.2.11. (Early Access) Parameter Group: layout_transform_params
2.5. IP Block Interfaces x
2.5.1. Clock and Reset 2.5.2. AXI Interfaces 2.5.3. AXI Interface Clock and Reset 2.5.4. Input Feature Tensor In-Memory Format 2.5.5. Output Tensor In-Memory Format
2.5.4. Input Feature Tensor In-Memory Format x
2.5.4.1. Multiple Input Graphs 2.5.4.2. Input Folding 2.5.4.3. Input Scale and Shift 2.5.4.4. Input Transform Mapping 2.5.4.5. Input Layout Transform Hardware
3. FPGA AI Suite IP Generation Utility x
3.1. IP Generation Utility Execution Flows 3.2. IP Generation Utility Inputs 3.3. IP Generation Utility Outputs 3.4. IP Generation Utility Command Line Options
3.4. IP Generation Utility Command Line Options x
3.4.1. The --flow create_ip Flow 3.4.2. The --flow add_arch Flow 3.4.3. The --flow list Flow 3.4.4. The --flow remove_arch Flow
4. FPGA AI Suite Ahead-of-Time Splitter Utility x
4.1. Files Generated by the FPGA AI Suite Ahead-of-Time (AOT) Splitter Utility 4.2. Building the FPGA AI Suite Ahead-of-Time (AOT) Splitter Utility 4.3. Running the FPGA AI Suite Ahead-of-Time (AOT) Splitter Utility 4.4. FPGA AI Suite Ahead-of-Time (AOT) Splitter Utility Example Application
5. CSR Map and Descriptor Queue x
5.1. Discovery ROM 5.2. Interrupt Control 5.3. DMA Descriptor Queue 5.4. DMA Control Registers 5.5. Performance Registers 5.6. Debug Network Registers 5.7. DMA License Register
1. FPGA AI Suite IP Reference Manual
2. About the FPGA AI Suite IP
3. FPGA AI Suite IP Generation Utility
4. FPGA AI Suite Ahead-of-Time Splitter Utility
5. CSR Map and Descriptor Queue
A. FPGA AI Suite IP Reference Manual Archives
B. FPGA AI Suite IP Reference Manual Document Revision History

5.3. DMA Descriptor Queue

The DMA contains a single descriptor FIFO for enqueuing inference requests. Descriptors potentially require multiple register writes and are added to the queue upon writing to the desc_input_output_base_addr register.

The desc_cfg_filter_base_addr and desc_cfg_num_words are registers that hold their value.

If you already enqueued a DMA descriptor and want to enqueue another descriptor with the same values for the desc_cfg_filter_base_addr and desc_cfg_num_words registers, then write to the desc_input_output_base_addr register.

If you want to change the desc_cfg_filter_base_addr and desc_cfg_num_words registers for the next descriptor, then you must set new values before writing to the desc_input_output_base_addr register.

Table 13.  DMA Descriptor Queue Registers

Register

Offset

Attribute

Description

desc_cfg_filter_base_addr

0x000

RW

Base address pointer for the configuration buffer and for the filter buffer.

The filters are located at desc_cfg_filter_base_addr + desc_cfg_num_words, which is encoded in the address provided to the filter reader as configuration data.

Must be aligned to a multiple of the DDR word size.

desc_cfg_num_words - 2

0x004

RW

Length of the configuration buffer - 2, in config words (64 bits – 32 for instruction, 32 for data)

desc_input_output_base_addr

0x008

RW

Base address pointer for the input feature data to be used for inference, and also the base address to place the output inference results into.

Must be aligned to a multiple of the DDR word size.

Writing to this register enqueues a descriptor into the internal DMA descriptor queue.

desc_diagnostics

0x00C

RO

This register is useful for debugging. Production software should not need to read from this.

Bit 0: Asserts if the descriptor queue overflows; this is a sticky bit which only clears after reset.

Bit 1: Descriptor queue is full or almost full.

Bit 2: Asserts if the inference limit for an unlicensed IP is reached. When asserted, inference requests are rejected.

All other bits are reserved.
Level Two Title