Newer
Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
.. _beaglebone-ai-specs:
BeagleBone AI High Level Specification
###########################################
This section provides the high level specification of BeagleBone® AI
Block Diagram
******************
The figure below is the high level block diagram of BeagleBone® AI. For
detailed layout information please check the schematics.
.. image:: images/BB_AI_Blockdiagram_1000px.jpg
:alt: beaglebone ai component placement
AM572x Sitara™ Processor
**************************
The Texas Instruments AM572x Sitara™ processor family of SOC devices
brings high processing performance through the maximum flexibility of a
fully integrated mixed processor solution. The devices also combine
programmable video processing with a highly integrated peripheral set
ideal for AI applications. The AM5729 used on BeagleBone® AI is the
super-set device of the family.
Programmability is provided by dual-core ARM® Cortex®-A15 RISC CPUs with
Arm® Neon™ extension, and two TI C66x VLIW floating-point DSP core, and
Vision AccelerationPac (with 4x EVEs). The Arm allows developers to keep
control functions separate from other algorithms programmed on the DSPs
and coprocessors, thus reducing the complexity of the system software.
Texas Instruments AM572x Sitara™ Processor Family Block Diagram\*
.. image:: images/BB_AI_AM5729_blockdiagram.jpg
:alt: beaglebone ai component placement
**MPU Subsystem** The Dual Cortex-A15 MPU subsystem integrates the
following submodules:
- ARM Cortex-A15 MPCore
- Two central processing units (CPUs)
- ARM Version 7 ISA: Standard ARM instruction set plus Thumb®-2,
Jazelle® RCT Java™ accelerator, hardware virtualization support,
and large physical address extensions (LPAE)
- Neon™ SIMD coprocessor and VFPv4 per CPU
- Interrupt controller with up to 160 interrupt requests
- One general-purpose timer and one watchdog timer per CPU – Debug
and trace features
- 32-KiB instruction and 32-KiB data level 1 (L1) cache per CPU
- Shared 2-MiB level 2 (L2) cache
- 48-KiB bootable ROM
- Local power, reset, and clock management (PRCM) module
- Emulation features
- Digital phase-locked loop (DPLL)
**DSP Subsystems** There are two DSP subsystems in the device. Each DSP
subsystem contains the following submodules:
- TMS320C66x™ Floating-Point VLIW DSP core for audio processing, and
general-purpose imaging and video processing. It extends the
performance of existing C64x+™ and C647x™ DSPs through enhancements
and new features.
- 32-KiB L1D and 32-KiB L1P cache or addressable SRAM
- 288-KiB L2 cache
- 256-KiB configurable as cache or SRAM
- 32-KiB SRAM
- Enhanced direct memory access (EDMA) engine for video and audio data
transfer
- Memory management units (MMU) for address management.
- Interrupt controller (INTC)
- Emulation capabilities
- Supported by OpenCL
**EVE Subsystems**
- 4 Embedded Vision Engines (EVEs) supported by TIDL machine learning
library
.. image:: images/BB_AI_EVEmodule.jpg
:alt: BeagleBone AI component placement
The Embedded Vision Engine (EVE) module is a programmable imaging and
vision processing engine. Software support for the EVE module is
available through OpenCL Custom Device model with fixed set of
functions. More information is available
http://www.ti.com/lit/wp/spry251/spry251.pdf
**PRU-ICSS Subsystems**
- 2x Dual-Core Programmable Real-Time Unit (PRU) subsystems (4 PRUs
total) for ultra low-latency control and software generated
peripherals. Access to these powerful subsystems is available through
through the P8 and P9 headers. These are detailed in Section 7.
**IPU Subsystems** There are two Dual Cortex-M4 IPU subsystems in the
device available for general purpose usage, particularly real-time
control. Each IPU subsystem includes the following components:
- Two Cortex-M4 CPUs
- ARMv7E-M and Thumb-2 instruction set architectures
- Hardware division and single-cycle multiplication acceleration
- Dedicated INTC with up to 63 physical interrupt events with 16-level
priority
- Two-level memory subsystem hierarchy
- L1 (32-KiB shared cache memory)
- L2 ROM + RAM
- 64-KiB RAM
- 16-KiB bootable ROM
- MMU for address translation
- Integrated power management
- Emulation feature embedded in the Cortex-M4
**IVA-HD Subsystem**
- IVA-HD subsystem with support for 4K @ 15fps H.264 encode/decode and
other codecs @ 1080p60 The IVA-HD subsystem is a set of video encoder
and decoder hardware accelerators. The list of supported codecs can
be found in the software development kit (SDK) documentation.
**BB2D Graphics Accelerator Subsystem** The Vivante® GC320 2D graphics
accelerator is the 2D BitBlt (BB2D) graphics accelerator subsystem on
the device with the following features:
- API support:
- OpenWF™, DirectFB
- GDI/DirectDraw
- BB2D architecture:
- BitBlt and StretchBlt
- DirectFB hardware acceleration
- ROP2, ROP3, ROP4 full alpha blending and transparency
- Clipping rectangle support
- Alpha blending includes Java 2 Porter-Duff compositing rules
- 90-, 180-, 270-degree rotation on every primitive
- YUV-to-RGB color space conversion
- Programmable display format conversion with 14 source and 7
destination formats
- High-quality, 9-tap, 32-phase filter for image and video scaling
at 1080p
- Monochrome expansion for text rendering
- 32K × 32K coordinate system
**Dual-Core PowerVR® SGX544™ 3D GPU** The 3D graphics processing unit
(GPU) subsystem is based on POWERVR® SGX544 subsystem from Imagination
Technologies. It supports general embedded applications. The GPU can
process different data types simultaneously, such as: pixel data, vertex
data, video data, and general-purpose data. The GPU subsystem has the
following features:
- Multicore GPU architecture: two SGX544 cores.
- Shared system level cache of 128 KiB
- Tile-based deferred rendering architecture
- Second-generation universal scalable shader engines (USSE2),
multithreaded engines incorporating pixel and vertex shader
functionality
- Present and texture load accelerators
- Enables to move, rotate, twiddle, and scale texture surfaces.
- Supports RGB, ARGB, YUV422, and YUV420 surface formats.
- Supports bilinear upscale.
- Supports source colorkey.
- Fine-grained task switching, load balancing, and power management
- Programmable high-quality image antialiasing
- Bilinear, trilinear, anisotropic texture filtering
- Advanced geometry DMA driven operation for minimum CPU interaction
- Fully virtualized memory addressing for OS operation in a unified
memory architecture (MMU)
Memory
************
.. __5_3_1_1gb_ddr3l:
1GB DDR3L
============
Dual 256M x 16 DDR3L memory devices are used, one on each side of the
board, for a total of 1 GB. They will each operate at a clock frequency
of up to 533 MHz yielding an effective rate of 1066Mb/s on the DDR3L bus
allowing for 4GB/s of DDR3L memory bandwidth.
16GB Embedded MMC
===================
A single 16GB embedded MMC (eMMC) device is on the board.
microSD Connector
==================
The board is equipped with a single microSD connector to act as a
secondary boot source for the board and, if selected as such, can be the
primary booth source. The connector will support larger capacity microSD
cards. The microSD card is not provided with the board.
Boot Modes
*************
Power Management
******************
Connectivity
******************
TODO: Add WiFi/Bluetooth/Ethernet
BeagleBone® AI supports the majority of the functions of the AM5729 SOC
through connectors or expansion header pin accessibility. See section 7
for more information on expansion header pinouts. There are a few
functions that are not accessible which are: (TBD)
TODO: This text needs to go somewhere.
On-board I2C Devices
***********************
+--------+--------+-----------------+
| A | Iden | Description |
| ddress | tifier | |
+========+========+=================+
| 0x12 | U3 | TPS6590379 PMIC |
| | | DVS |
+--------+--------+-----------------+
| 0x41 | U78 | STMPE811Q ADC |
| | | and GPIO |
| | | expander |
+--------+--------+-----------------+
| 0x47 | U13 | HD3SS3220 USB |
| | | Type-C DRP port |
| | | controller |
+--------+--------+-----------------+
| 0x50 | U9 | 24LC32 board ID |
| | | EEPROM |
+--------+--------+-----------------+
| 0x58 | U3 | TPS6590379 PMIC |
| | | power registers |
+--------+--------+-----------------+
| 0x5a | U3 | TPS6590379 PMIC |
| | | interfaces and |
| | | auxilaries |
+--------+--------+-----------------+
| 0x5c | U3 | TPS6590379 PMIC |
| | | trimming and |
| | | test |
+--------+--------+-----------------+
| 0x5e | U3 | TPS6590379 PMIC |
| | | OTP |
+--------+--------+-----------------+