VisionSDK_UserGuide_TDA3xx Vision SDK User Guide TDA3xx

User Manual:

Open the PDF directly: View PDF .
Page Count: 39

Download
Open PDF In Browser	View PDF

Vision SDK TDA3xx
(v03.03.00)
User Guide

Copyright © 2014 Texas Instruments Incorporated. All rights reserved.
Information in this document is subject to change without notice. Texas Instruments may have pending
patent applications, trademarks, copyrights, or other intellectual property rights covering matter in this
document. The furnishing of this documents is given for usage with Texas Instruments products only and
does not give you any license to the intellectual property that might be contained within this document.
Texas Instruments makes no implied or expressed warranties in this document and is not responsible for
the products based from this document.
Page 1 of 39

IMPORTANT NOTICE
Texas Instruments and its subsidiaries (TI) reserve the right to make changes to their products or to
discontinue any product or service without notice, and advise customers to obtain the latest version of relevant
information to verify, before placing orders, that information being relied on is current and complete. All
products are sold subject to the terms and conditions of sale supplied at the time of order acknowledgment,
including those pertaining to warranty, patent infringement, and limitation of liability.
TI warrants performance of its products to the specifications applicable at the time of sale in accordance with
TI’s standard warranty. Testing and other quality control techniques are utilized to the extent TI deems
necessary to support this warranty. Specific testing of all parameters of each device is not necessarily
performed, except those mandated by government requirements.
Customers are responsible for their applications using TI components.
In order to minimize risks associated with the customer’s applications, adequate design and operating
safeguards ought to be provided by the customer so as to minimize inherent or procedural hazards.
TI assumes no liability for applications assistance or customer product design. TI does not warrant or represent
that any license, either express or implied, is granted under any patent right, copyright, mask work right, or
other intellectual property right of TI covering or relating to any combination, machine, or process in which such
products or services might be or are used. TI’s publication of information regarding any third party’s products or
services does not constitute TI’s approval, license, warranty or endorsement thereof.
Reproduction of information in TI data books or data sheets is permissible only if reproduction is without
alteration and is accompanied by all associated warranties, conditions, limitations and notices. Representation
or reproduction of this information with alteration voids all warranties provided for an associated TI product or
service, is an unfair and deceptive business practice, and TI is neither responsible nor liable for any such use.
Resale of TI’s products or services with statements different from or beyond the parameters stated by TI for
that product or service voids all express and any implied warranties for the associated TI product or service, is
an unfair and deceptive business practice, and TI is not responsible nor liable for any such use.
Also see: Standard Terms and Conditions of Sale for Semiconductor Products.
www.ti.com/sc/docs/stdterms.htm
Mailing Address:
Texas Instruments
Post Office Box 655303
Dallas, Texas 75265

Page 2 of 39

TABLE OF CONTENTS
1
1.1

Introduction ................................................................................................. 4
References ...................................................................................................... 4

2
2.1
2.2
2.3
2.4
2.5
2.6

System Requirements .................................................................................. 5
Windows Installation......................................................................................... 5
Linux Installation ............................................................................................. 6
Hardware Requirements .................................................................................... 7
Required H/W modification / Configurations ....................................................... 10
Supported Sensors ......................................................................................... 12
Software Installation ...................................................................................... 12

3
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12

Build and Run ............................................................................................. 13
Overview of application in release .................................................................... 13
Building the application ................................................................................... 14
UART settings ................................................................................................ 16
Boot Modes ................................................................................................... 18
Load using QSPI............................................................................................. 18
Load using QSPI and SD boot .......................................................................... 22
Load using CCS .............................................................................................. 24
Run the demo ................................................................................................ 29
DCC ............................................................................................................. 32
Fast boot usecase .......................................................................................... 35
Surround View Fast Boot Use case .................................................................... 37
Surround View Use case under 128 MB DDR configuration ................................... 37

Revision History ......................................................................................... 39

Page 3 of 39

Introduction
Vision Software Development Kit (Vision SDK) is a multi-processor software
development package for TI’s family of ADAS SoCs. The software framework allows
users to create different ADAS application data flows involving video capture, video
pre-processing, video analytics algorithms, and video display. The framework has
sample ADAS data flows which exercises different CPUs and HW accelerators in the
ADAS SoC and demonstrates how to effectively use different sub-systems within the
SoC. Frame work is generic enough to plug in application specific algorithms in the
system.
Vision SDK is currently targeted for the TDA2xx and TDA3xx family of SoCs
This document explains the HW/SW setup for TDA3x EVM. Refer
VisionSDK_UserGuide_TDA2xx.pdf for TDA2x EVM related HW/SW setup info.

1.1

References
Refer the below additional documents for more information about Vision SDK
Document

Description

VisionSDK_ReleaseNotes.pdf

Release specific information

VisionSDK_UserGuide_TDA3xx.pdf

This document. Contains install,
build, execution information

VisionSDK_ApiGuide.CHM

User API interface details

VisionSDK_SW_Architecture.pdf

Overview of software architecture

VisionSDK_DevelopmentGuide.pdf

Details how to create data flow (s)
& add new functionality

VisionSDK_DataSheet.pdf

Summary of features supported,
not supported in a release.
Performance
and
benchmark
information.

VisionSDK_FAQs.pdf

Document contains FAQ

Page 4 of 39

System Requirements
This chapter provides a brief description on the system requirements (hardware and
software) and instructions for installing Vision SDK.

2.1

Windows Installation

2.1.1

PC Requirements
Installation of this release needs a windows machine with about 8GB of free disk
space. Building of the SDK is supported on windows environment.

2.1.2

Software Requirements
All software packages required to build and run the Vision SDK are included as part
of the SDK release package except for the ones mentioned below

2.1.2.1

A15 Compiler, Linker

The windows installer for the GCC ARM tools should be downloaded from below link
https://launchpad.net/gcc-arm-embedded/+milestone/4.9-2015-q3-update
The tools need to be installed under
“/ti_components/cg_tools/windows/”.
IMPORTANT NOTE: A15 Compiler and linker MUST be installed before
proceeding else compile will fail. Also make sure the compiler is installed at
the exact path mentioned above

2.1.3

Code Composer Studio
CCS is needed to load, run and debug the software. CCS can be downloaded from
the below link. CCS version 6.0.1.00040 or higher should be installed.

http://processors.wiki.ti.com/index.php/Download_CCS

Page 5 of 39

2.2

Linux Installation

2.2.1

PC Requirements
Installation of this release needs a Linux Ubuntu 14.04 machine.
IMPORTANT NOTE: If you are installing Ubuntu on a virtual machine, ensure it’s a
64 bit Ubuntu.

2.2.2

Software Requirements
All software packages required to build and run the Vision SDK are included as part
of the SDK release package except for the ones mentioned below

2.2.2.1

A15 Compiler, Linker

The Linux installer for the GCC ARM tools should be downloaded from below link
https://launchpad.net/gcc-arm-embedded/+milestone/4.9-2015-q3-update
The tools need to be installed under
“/ti_components/cg_tools/linux/”.
IMPORTANT NOTE: A15 Compiler and linker MUST be installed before initiating the
build else compilation will fail. Also make sure the compiler is installed at the exact
path mentioned above after installation of vision sdk.
Use following steps to install the toolchain
$> cd $INSTALL_DIR/ti_components/cg_tools/linux
$> tar -xvf gcc-arm-none-eabi-4_9-2015q3-20150921-linux.tar.tar
IMPORTANT NOTE: Ensure the toolchain is for 32 / 64 bit machine as per
configuration of installation machine
If your machine is 64 bit and you have downloaded toolchain from link above
Execute following step on installation machine
$>sudo apt-get install ia32-libs lib32stdc++6 lib32z1-dev lib32z1 lib32ncurses5

lib32bz2-1.0

2.2.3

Other software packages for build depending upon OS baseline
Ensure these packages/tools are installed on the installation machine
uname, sed, mkimage, dos2unix, dtrx, mono-complete, git, lib32z1
lib32ncurses5
lib32bz2-1.0
libc6:i386
libc6-i386
libstdc++6:i386
libncurses5:i386 libz1:i386 libc6-dev-i386 device-tree-compiler monocomplete
To install
$>sudo apt-get install

Page 6 of 39

2.3

Hardware Requirements
Hardware setup for different use-cases is described in this section

2.3.1

Single Channel (SC) Use-case Hardware Setup
SC use-case needs the below hardware
1. TDA3xx EVM , power supply (12V 5 AMP)
2. Video Sensors, you would require one of the sensors listed in section 2.5. Please
refer section 2.3.4, it visually shows as to where the sensor should be connected.
3. 1Gbps Ethernet Cable (optional)
4. HDMI 1080p60 capable Display Monitor OR LCD Screen. LCD should be connected
to LCD out connector, as shown in section 2.3.4. 10’’ or 7’’ LCD is supported
WARNING: LI Camera Interface is different from LI Camera CSI2 Interface. Putting
a CSI2 sensor on LI Camera Interface will damage the sensor

2.3.2

VIP multi-channel LVDS capture (SRV) Use-case Hardware Setup
Refer the TDA2x user guide “VisionSDK_UserGuide_TDA2xx.pdf” for the LVDS setup.
To support multichannel LVDS capture on TDA3xx EVM, there are some board
modifications required. Please refer to the “Running VPS Application on TDA3XX
EVM” section of the VPS user guide to get the more details
1. For VIP capture from Multi-deserializer board, the multi-deserializer board should
be configured for 4-channel operation. The CN2, CN3 and CN4 jumper settings
should be set.
2. In case of Multi-deserializer capture through VIP or ISS capture from sensors,
board modification is required in the base board to avoid I2C issues
WARNING: Select the display resolution as HDMI XGA TDM mode . Failing which, only 2 channels of captured video streams are displayed. This
is required as some of the VIP input pins is multiplexed with display output pins.

2.3.3

ISS Multiple Channel (SRV) Use-case Hardware Setup
SRV use-case needs the below hardware
1. TDA3xx EVM, power supply (12V 5 AMP)
2. UB960 Application Board, power supply (12V 5 AMP)
3. 4 TIDA00262 modules (AR0140 sensor) or IMI modules (OV10640 Rev E sensor)
& LVDS cables to connect camera modules to UB960 application board
a. List details of this camera module http://www.ti.com/tool/TIDA00262?keyMatch=TIDA-00262&tisearch=Search-EN-Everything#tiDevice
4. HDMI 1080p60 capable Display Monitor
WARNING: CSI2 Clock: The maximum CSI2 clock could be 750 MHz, please refer
the device data manuals for details. Some of the VisionSDK usecases (UB964 based),
overclocks by 50 MHz (i.e. 800 MHz) and it works as expected. This over clocking is
due the crystal (25 MHz) used in UB964 EVM, by choosing 24 MHz crystal UB964
CSI2 clock can be operated with-in specified limits.

Page 7 of 39

2.3.4

Sensor Interfaces

CSI2 Receiver Connector
Figure 2.3 1 15x15 TDA3x EVM (FRONT SIDE)

Page 8 of 39

Figure 2.3 2 15x15 TDA3x EVM (BACK SIDE)

Figure 2.3 3 TDA3x EVM with UB960 Application Board

NOTE: For Switch settings refer section 2.4.2
2.3.5

EDID Programming for HDMI Capture
EDID information needs to be programmed on the EEPROM present on the EVM. This
is required for the HDMI source to recognize the format and resolution supported by

Page 9 of 39

the receiver (TDA3xx SoC). If this step is not done or if this step fails, then TDA3xx
SoC will not be able to receive data via HDMI.
IMPORTANT NOTE: It’s recommended to program the HDMI receivers EDID.
The default EDID is programmed to receive 1080P60 video streams only. If
stream of different resolution is required (or EDID is corrupted), the EDID
would require an update. Refer the EDID programming points in the section
Running VPS Application on (TDAXXX EVM) documented in VPS User Guide
in PDK.

2.4

Required H/W modification / Configurations

2.4.1

TDA3XX EVM Modifications for SCH use case
I2C transactions for few sensors like OV10640 fail at 400 KHz I2C frequency. To fix
this issue, few resistors need to be updated/changed on the TDA3xx EVM. Please
refer to the “Running VPS Application on TDA3XX EVM” section of the VPS user guide
to get the more details

2.4.2

Changes required on UB960 Application board
1. Configure to supply 9V on the LVDS lines to TIDA00262 modules
a. J14 Short pins 1-2 as show in Error! Reference source
Mode
found.
□
2. Configure UB960 to operate in LVDS in 75 MHz mode
□
a. Dip Switch MODE 3 should be ON
□
3. Refer the picture below
□

ot
4
3
2
1

Page 10 of 39

Mode

Page 11 of 39

2.5

Supported Sensors
Refer to ProcessorSDK_Vision_IssSensor_TestMatrix.xlsx in docs/TestReports for the
supported feature on each of the sensors.

2.6

Software Installation
PROCESSOR_SDK_VISION_03_xx_xx_xx_setupwin.exe is the SDK package installer.
Copy the installer to the path of your choice.
Double click the installer to begin the installation.
Follow the self-guided installer for installation.
IMPORTANT NOTE: On some computers running as administrator is needed. Right
click on the installer and select option of “Run as administrator”. If this is not done
then you may see a message like “This program might not have installed correctly
On
completion
of
installation
a
folder
by
name
PROCESSOR_SDK_VISION_03_xx_xx_xx would have been created in the installation
path.

2.6.1

Uninstall Procedure
To uninstall, double click on uninstall.exe created during installation in the folder
PROCESSOR_SDK_VISION_03_xx_xx_xx.
At the end of uninstall, PROCESSOR_SDK_VISION_03_xx_xx_xx folder still remains.
It is just an empty folder. It can be deleted manually.

Page 12 of 39

Build and Run
This chapter provides a brief overview of the sample application or use case present
in the SDK and procedure to build and run it. For more details about optimized build
process please refer to VisionSDK_UserGuide_BuildSystem.pdf

3.1

Overview of application in release
The Vision SDK supports the following use-cases are grouped under following
categories


Single Camera Use-cases



Multi-Camera LVDS Use-cases



AVB RX Use-cases, (TDA2x & TDA2Ex ONLY)



Dual Display Use-cases, (TDA2x EVM ONLY)



ISS Use-cases, (TDA3x ONLY)



Network RX/TX Use-cases



Fast boot ISS capture + display (TDA3x ONLY)*
* Not listed in Runtime Menu

Refer to VisionSDK_DataSheet.pdf for detailed description of each category.
The demos support devices listed in section 2.4.2 as capture source and HDMI
1080P60 can also be used as a capture source.
The demos support following devices as display devices





LCD 7-inch 800x480@60fps
LCD 10-inch 1280x720@60fps
LCD 10-inch 1920x1200@60fps
HDMI 1080p60 (default)

Use option "s" on the main menu in UART to select different capture and display
devices.

Page 13 of 39

3.2

Building the application
a) On windows command prompt, go inside the directory
PROCESSOR_SDK_VISION_03_xx_xx_xx\vision_sdk\build.
b) Open file Rules.make and
set MAKEAPPNAME=apps and MAKECONFIG=tda3xx_evm_bios_all
c) Build is done by executing gmake. “gmake” is present inside XDC package. For
“gmake” to be available in windows command prompt, the XDC path must be set in
the windows system path.
IMPORTANT NOTE: xdc path is needed to be set in environment variables. If not,
then set it using the set PATH =
/ti_components/os_tools/windows/xdctools_x_xx_xx_xx;%PATH% in
command prompt
IMPORTANT NOTE: A15 Compiler and linker MUST be installed before proceeding
else compile will fail. Also make sure the compiler is installed at the exact path as
mentioned in Directory Structure .
IMPORTANT NOTE: If the installation folder depth is high then windows cmd
prompt fails with error that it cannot find a file, even in file is present in
mentioned path, this is because Windows has a limitation of 8191
characters for the commands that can execute. In such a situation as a
workaround either restrict the folder depth to d:/ or if it cannot be
restricted use git bash (version 2.13) to build. Refer
https://support.microsoft.com/en-in/kb/830473 for more details.
(Always point to xdc path gmake only)
d) Under vision_sdk\build directory
i.

When building first time run the below sequence of commands
> gmake -s -j depend
> gmake -s -j

ii.

When building after the first time or incremental build, run the below command
> gmake -s -j
Executing “gmake -s -j depend “ will build all the necessary components (PDK drivers,
EDMA drivers and sdk dependent files) and “gmake -s -j” will build the Vision SDK
framework and examples.
IMPORTANT NOTE: For incremental build, make sure to do "gmake -s -j
depend" before "gmake -s -j” when below variables specified in
\vision_sdk\$(MAKEAPPNAME)\configs\$(MAKECONFIG)\*cfg.mk are changed







when PROC_$(CPU)_INCLUDE is changed
when DDR_MEM is changed
when PROFILE is changed
when ALG plugin or usecase is enabled or disabled in
\vision_sdk\$(MAKEAPPNAME)\configs\$(MAKECONFIG) \*_cfg.mk
when any .h or .c file in TI component is installed in ti_components is
changed
when any new TI component is installed in ti_components

Page 14 of 39

If "gmake -s -j depend" is not done in these cases then build and/or execution
may fail.
IMPORTANT NOTE: When options (other than those specified above) are
changed in \vision_sdk\$(MAKEAPPNAME)\configs\$(MAKECONFIG)\cfg.mk, a
clean build is recommended for the updated settings to take effect.
e) On a successful build completion, following executables will be generated in the
below path
\vision_sdk\binaries\$(MAKEAPPNAME)\$(MAKECONFIG)\vision_sdk\bin\tda3xxevm
vision_sdk_arp32_1_release.xearp32F
vision_sdk_c66xdsp_1_release.xe66
vision_sdk_c66xdsp_2_release.xe66
vision_sdk_ipu1_0_release.xem4
vision_sdk_ipu1_1_release.xem4

To speed up the incremental builds the following can be done as required.
The number of processors included in the build can be changed by modifying below
values in \vision_sdk\$(MAKEAPPNAME)\configs\$(MAKECONFIG)\cfg.mk. A value of
"no" means CPU not included in build, value of "yes" means CPU included in build.
Make sure to do clean build and then “gmake -s -j depend” before “gmake -s- j”
when number of CPUs included is changed
PROC_DSP1_INCLUDE=yes
PROC_DSP2_INCLUDE=yes
PROC_EVE1_INCLUDE=yes
PROC_IPU1_0_INCLUDE=yes
PROC_IPU1_1_INCLUDE=yes

g) The build configuration that is selected in config file can be confirmed by doing below
> gmake -s showconfig
h) Cleaning the build can be done by following command
> gmake -s clean
Alternatively, below folder can be deleted to delete all
generated files
> rm -rf
..\binaries\$(MAKEAPPNAME)\$(MAKECONFIG)\vision_sdk\bin
> rm –rf ..\links_fw\include\configs

Page 15 of 39

3.3

UART settings
Connect a serial cable to the UART port of the EVM and the other end to the serial
port of the PC (configure the HyperTerminal at 115200 baud rate) to obtain logs and
select demo. EVM it detects 4 UART ports, you need to select the 3rd one.
IMPORTANT NOTE: On some EVMs we were observing that UART terminal does not
work. Updating the USB to UART driver on PC made UART work on the failings PCs.
You can download the drivers from the below link.
http://www.ftdichip.com/Drivers/VCP.htm
http://www.ftdichip.com/Drivers/CDM/CDM%20v2.10.00%20WHQL%20Certified.exe

Page 16 of 39

Page 17 of 39

3.4

Boot Modes

Supported boot modes on TDA3xx ES1.0 and ES2.0 device:
Boot Mode
QSPI_1
QSPI_4
NOR
Debug

EVM Switch Setting
SYSBOOT(SW2)[1:16]
00011000 10000001
10011000 10000001
01011000 10000101
00111000 10000001

EVM Switch Setting
SW8001[1:8]
0100 0001
0100 0001
1100 0001
XXXXXXXX

3.5

Load using QSPI

3.5.1

Steps to generate qspi writer tools
NOTE: SBL qspi image is built from pdk package. To build qspi, run the command gmake -s sbl
from vision_sdk\build directory This generates all required tools and all sbl images under
vision_sdk\binaries\$(MAKEAPPNAME)\$(MAKECONFIG)\sbl directory
1. The flash writer is present in
vision_sdk\binaries\$(MAKEAPPNAME)\$(MAKECONFIG)\sbl\qspi_flash_writer\$(P
LATFORM)\qspi_flash_writer_ipu1_0_release.xem4
2. The SBL images are present in
vision_sdk\binaries\$(MAKEAPPNAME)\$(MAKECONFIG)\sbl\qspi\$(OPP)\$(PLATF
ORM)\sbl_qspi_$(OPP)_ipu1_0_release.tiimage
IMPORTANT NOTE: “gmake -s sbl” requires GCC tools need to be installed in
“/ti_components/cg_tools//gcc-arm-none-eabi-4_9-2015q3”
location. Tool can be downloaded from below link.
https://launchpad.net/gcc-arm-embedded/+milestone/4.9-2015-q3-update

3.5.2



The config options, like CPUs to use, debug or release profile etc, used to
make
the
application
image
will
be
the
values
specified
in

\vision_sdk\$(MAKEAPPNAME)\configs\$(MAKECONFIG)\cfg.mk


3.5.3

The Surround View LUT and Perspective Matrix are flashed at an
offset of 20 MB in the QSPI hence make sure the generated appImage
doesn’t exceed 20 MB in case Surround View use cases are intended
to be run.

Flashing steps
Flashing pin settings: Please refer Boot Modes for pin boot mode pin setting.

Page 18 of 39

NOTE: Image indicates the sysboot position on board not the switch settings

For loading binaries using CCS refer Load using CCS till step 8.
1. Connect M4 (IPU). Do CPU reset
2. Load below image on M4
C:\PROCESSOR_SDK_VISION_03_XX_XX_XX\vision_sdk\binaries\$(MA
KEAPPNAME)\$(MAKECONFIG)\sbl\qspi_flash_writer\$(PLATFORM)\qs
pi_flash_writer_ipu1_0_release.xem4
3. Run the core. You would see below console logs
[Cortex_M4_IPU1_C0]
QSPI Flash writer application
Enter Device type to use
1 - 1 bit read from flash
2 - 4 bit (Quad) read from flash

Erase Options:

Select appropriate Device Type, for TDA3x EVM, press ‘2’.

Enter Erase Option:

--------------0 -> Erase Only Required Region
1 -> Erase Whole Flash
2 -> Skip Erase
1

MID - 1
DID - 18

Load Options:

Enter 0 for Erase-Only (without flashing any image)

-------------

Note : File size should be less than 33554432 Bytes.

0 -> fread using code (RTS Library)

Enter
the
file
path
to
flash:
C:\PROCESSOR_SDK_VISION_03_XX_XX_XX\vision_sdk\binaries\$(M
AKEAPPNAME)\$(MAKECONFIG)\sbl\qspi\$(OPP)\$(PLATFORM)\sbl_qs
pi_$(OPP)_ipu1_0_release.tiimage

Enter Load Option: 0

1 -> load raw using CCS (Scripting console)

Enter the Offset in bytes (HEX) 0x00

Read xxxxxx bytes from [100%] file...Done.
QSPI whole chip erase in progress
QSPI file write started
************QSPI
sucessfully**************

flash

4. Reset the board and Repeat step 1, 2, 3.

Page 19 of 39

completed

[Cortex_M4_IPU1_C0]
QSPI Flash writer application
Enter Device type to use
1 - 1 bit read from flash
2 - 4 bit (Quad) read from flash
Select appropriate Device Type, for TDA3x
EVM, press ‘2’.

Load Options:
------------0 -> fread using code (RTS Library)
1 -> load raw using CCS (Scripting console)
Enter Load Option: 1

MID - 1
DID - 18
Enter the File Name
C:\PROCESSOR_SDK_VISION_03_XX

_XX_XX\vision_sdk\binaries\$(MAK
EAPPNAME)\$(MAKECONFIG)\vision
_sdk\bin\$(SOC)\sbl_boot\AppIma
ge_BE
Enter the Offset in bytes (HEX):

0x80000

Erase Options:
---------------

Open Scripting console window by clicking
“Menu -> View -> Scripting console” and enter
below command on scripting console.
loadRaw(0x80500000,
0,
"C:/VISION_SDK_XX_XX_XX_XX/vision_s
dk/binaries/$(MAKEAPPNAME)/$(MAKECO
NFIG)/vision_sdk/bin/$(SOC)/sbl_boot/A
ppImage_BE", 32, false);
IMPORTANT NOTE: The load address in
loadRaw command could be different based on
the board/SBL size etc. SBL figures out the
address and prints it on CCS console. Use this
address in loadRaw command for copying
AppImage_BE.

0 -> Erase Only Required Region
1 -> Erase Whole Flash
2 -> Skip Erase
Enter Erase Option: 0

In CCS console Enter any alpha-numeric key
once loadraw is complete... as shown in below
image
QSPI file write started
************QSPI
flash
successfully**************

completed

NOTE: If flashing binaries for fast boot use case then the file name and offsets will
be different and this step needs to be done twice, once for UCEarly, once for UCLate.
Refer section section 3.10.3 for steps to generate binaries, binary names, offsets
[Cortex_M4_IPU1_C0]
QSPI Flash writer application
Enter Device type to use
1 - 1 bit read from flash
2 - 4 bit (Quad) read from flash
Select appropriate Device Type, for TDA3x
EVM, press ‘2’.

Load Options:
------------0 -> fread using code (RTS Library)
1 -> load raw using CCS (Scripting console)
Enter Load Option: 1

MID - 1
DID - 18
Enter the File Name
C:\PROCESSOR_SDK_VISION_03_XX

_XX_XX\vision_sdk\binaries\$(MAK
EAPPNAME)\$(MAKECONFIG)\vision
_sdk\bin\$(SOC)\sbl_boot\AppIma
ge_UcEarly_BE

Page 20 of 39

Enter the Offset in bytes (HEX):

0x80000

Erase Options:
--------------0 -> Erase Only Required Region
1 -> Erase Whole Flash
2 -> Skip Erase

IMPORTANT NOTE: The load address in
loadRaw command could be different based on
the board/SBL size etc. SBL figures out the
address and prints it on CCS console. Use this
address in loadRaw command for copying
AppImage_BE.
In CCS console Enter any alpha-numeric key
once loadraw is complete... as shown in below
image

Enter Erase Option: 0
QSPI file write started
************QSPI
flash
successfully**************

[Cortex_M4_IPU1_C0]
QSPI Flash writer application
Enter Device type to use
1 - 1 bit read from flash
2 - 4 bit (Quad) read from flash
Select appropriate Device Type, for TDA3x
EVM, press ‘2’.

completed

Load Options:
------------0 -> fread using code (RTS Library)
1 -> load raw using CCS (Scripting console)
Enter Load Option: 1

MID - 1
DID - 18
Enter the File Name
C:\PROCESSOR_SDK_VISION_03_XX

_XX_XX\vision_sdk\binaries\$(MAK
EAPPNAME)\$(MAKECONFIG)\vision
_sdk\bin\$(SOC)\sbl_boot\AppIma
ge_UcLate_BE
Enter the Offset in bytes (HEX):

0xA80000

Erase Options:
---------------

Open Scripting console window by clicking
“Menu -> View -> Scripting console” and enter
below command on scripting console.
loadRaw(0x80500000,
0,
"C:/VISION_SDK_XX_XX_XX_XX/vision_s
dk/binaries/$(MAKEAPPNAME)/$(MAKECO
NFIG)/vision_sdk/bin/$(SOC)/sbl_boot/A
ppImage_UcLate_BE", 32, false);
IMPORTANT NOTE: The load address in
loadRaw command could be different based on
the board/SBL size etc. SBL figures out the
address and prints it on CCS console. Use this
address in loadRaw command for copying
AppImage_BE.

0 -> Erase Only Required Region
1 -> Erase Whole Flash
2 -> Skip Erase
Enter Erase Option: 0

In CCS console Enter any alpha-numeric key
once loadraw is complete... as shown in below
image
QSPI file write started
************QSPI
flash
successfully**************

completed

5. On completion change the pin setting as shown in Boot Modes table.

Page 21 of 39

3.6

Load using QSPI and SD boot
In this mode SBL boots from QSPI but AppImage boots from SD card. This allows us
to flash SBL once to QSPI and subsequently we can boot new AppImage just by
copying AppImage to SD card.

3.6.1

Steps to generate qspi writer tools
NOTE: SBL image for the qspi sd boot is built from pdk package.To build this image,
Run the command gmake -s sbl from vision_sdk\build dir. This generates all
required tools under vision_sdk\binaries\$(MAKEAPPNAME)\$(MAKECONFIG)\sbl\
1. The flash writer is present in
vision_sdk\binaries\$(MAKEAPPNAME)\$(MAKECONFIG)\sbl\qspi_flash_writer\$(P
LATFORM)\qspi_flash_writer_ipu1_0_release.xem4
2. The SBL images are present in
vision_sdk\binaries\$(MAKEAPPNAME)\$(MAKECONFIG)\sbl\qspi_sd\$(OPP)\$(PL
ATFORM)\sbl_qspi_sd_$(OPP)_ipu1_0_release.tiimage
IMPORTANT NOTE: “gmake -s sbl” requires GCC tools need to be installed in
“/ti_components/cg_tools//gcc-arm-none-eabi-4_9-2015q3”
location. Tool can be downloaded from below link.
https://launchpad.net/gcc-arm-embedded/+milestone/4.9-2015-q3-update

3.6.2

Steps to generate appImage
Following steps need to be followed to generate the application image
1. Make sure the executables are built as shown in Building the application
2. To generate the application image run below command from “vision_sdk\build”
folder
> gmake -s appimage
IMPORTANT NOTE: The config options, like CPUs to use, debug or release profile
etc, used to make the application image will be the values specified in

\vision_sdk\$(MAKEAPPNAME)\configs\$(MAKECONFIG)\cfg.mk
3.6.3

Flashing steps
Flashing pin settings: Please refer Boot Modes for pin boot mode pin setting.

For loading binaries using CCS refer Load using CCS till step 8.

Page 22 of 39

1. Connect M4 (IPU). Do CPU reset
2. Load image on M4
(C:\PROCESSOR_SDK_VISION_03_XX_XX_XX\vision_sdk\binaries\$
(MAKEAPPNAME)\$(MAKECONFIG)\sbl\qspi_flash_writer\$(PLATFO
RM)\qspi_flash_writer_ipu1_0_release.xem4)
3. Run the core.
You should get below logs on console outputs
[Cortex_M4_IPU1_C0]
QSPI Flash writer application
Enter Device type to use
1 - 1 bit read from flash
2 - 4 bit (Quad) read from flash

Enter
the
file
path
to
flash:
C:\PROCESSOR_SDK_VISION_03_XX_XX_XX\vision_sdk\binaries\$(MAK
EAPPNAME)\$(MAKECONFIG)\sbl\qspi_sd\$(OPP)\$(PLATFORM)\sbl_qspi_
sd_$(OPP)_ipu1_0_release.tiimage

Select appropriate Device Type, for TDA3x EVM, press ‘2’.

---------------

Enter the Offset in bytes (HEX) 0x00
Erase Options:

MID - 1

0 -> Erase Only Required Region

DID - 18

1 -> Erase Whole Flash

Enter 0 for Erase-Only (without flashing any image)
Note : File size should be less than 33554432 Bytes.

2 -> Skip Erase
Enter Erase Option:

Load Options:
------------0 -> fread using code (RTS Library)
1 -> load raw using CCS (Scripting console)
Enter Load Option: 0
Read xxxxxx bytes from [100%] file...Done.
QSPI whole chip erase in progress
QSPI file write started
************QSPI flash completed sucessfully**************

NOTE: User needs to copy the AppImage to root folder in SD card and insert SD
card and power on EVM to boot it. SD card should be formatted as FAT32 with 512
bytes per sector.

Page 23 of 39

3.7

Load using CCS
After installing CCS, follow below steps to complete the platform setup,
1. GELs are available in
\ti_components\ccs_csp \auto_device_support_x.x.x.zip
NOTE:
o

GELs are also available at
http://processors.wiki.ti.com/index.php/Device_support_files
Under Automotive pick
Automotive vX.X.X

To install the new GEL versions, you need to extract the zip to
/ccsv6/ccs_base
Change the following GEL files for vision SD as below,

TDA3xx_multicore_reset.gel
o Set VISION_SDK_CONFIG to 1
o 256MB mode not supported

2. CCS Target Configuration creation:
a. Open “Target Configurations” tab, by navigating through the menu “View >Target Configurations”.

b. Create a new Target Configuration (TDA3xx Target Configuration) by
navigating through the menu “File->New->Target Configuration File”.

Page 24 of 39

c. Specify Connections as “Spectrum Digital XDS560V2 STM USB Emulator”.
Specify Board or Device as “TDA3x”.

Connect JTAG to the board.

Page 25 of 39

Reset EVM through the power recycle button.

5. Now launch the previously created TDA3xx Target Configuration.

Page 26 of 39

Connect to core Cortex_M4_IPU1_C0.

On successful connect, the following log appears on CCS console:

Cortex_M4_IPU1_C0: GEL Output: --->>> TDA3xx Target Connect Sequence DONE !!!!!

<<<---

Select Cortex_M4_IPU1_C0, navigate to Scripts->TDA3xx MULTICORE Initialization
TDA3xx_MULTICORE_EnableALLCores

On successful script execution, the following log appears on CCS console:

Cortex_M4_IPU1_C0: GEL Output: --->>> EVESS Initialization is DONE! <<<--10. Now connect the core shown below
ARP32_EVE_1, C66xx_DSP1, C66xx_DSP2 and Cortex_M4_IPU1_C1

Page 27 of 39

11. Once the cores are connected, do CPU Reset for all the cores.

12. On the cores load the binaries as mentioned below

On ARP32_EVE_1, load the binary, “vision_sdk_arp32_1_release.xearp32F”.
On C66xx_DSP2, load the binary, “vision_sdk_c66xdsp_2_release.xe66”.
On C66xx_DSP1, load the binary, “vision_sdk_c66xdsp_1_release.xe66”.
On Cortex_M4_IPU1_C0, load the binary, “vision_sdk_ipu1_0_release.xem4”.
On Cortex_M4_IPU1_C1, load the binary, “vision_sdk_ipu1_1_release.xem4”.
IMPORTANT NOTE: Binary for Cortex_M4_IPU1_C0 MUST be loaded before
Cortex_M4_IPU1_C1 since IPU1-0 does MMU config for the complete IPU1 system. Other
binaries can be loaded in any order.

Page 28 of 39

3.8

Run the demo

3.8.1

Single channel demos with HDMI input
IMPORTANT NOTE: To demonstrate better output all single channel use cases that require
HDMI input should use video clips mentioned in the table below. These clips can be downloaded
from
https://cdds.ext.ti.com/ematrix/common/emxNavigator.jsp?objectId=28670.42872.
30602.25095 .
Use case

Input clip to
be
played
by
HDMI
player

1CH VIP capture + Sparse Optical Flow (EVE1) + Display

Clip2

b: 1CH VIP capture (HDMI) + Lane Detect (DSP1) +
Display

Clip1

c: 1CH VIP capture (HDMI) + SOF (EVE1) + SFM (DSP1)
+ Display

Clip2

d: 1CH VIP capture (HDMI) + Traffic Light Recognition
(TLR) (DSP1) + Display

Clip2

e: 1CH VIP capture (HDMI) + Pedestrian, Traffic Sign,
Vehicle Detect 2 (EVE1 + DSP1) + Display

Clip2

f: 1CH VIP capture (HDMI) + FrontCam Analytics 2
(PD+TSR+VD+LD+TLR+SFM) (DSPx, EVEx) + Display
(HDMI)

Clip3

Use case
No.
“Runtime
Menu”

SFM_POSE.bin - SFM (Usecase ‘c’) and EUNCAP demo (Usecase ‘f’) needs
SFM_POSE.bin on the SD card. It will be available as part of the file downloaded
from the above link.
3.8.2

Steps to run
1. Power-on the Board after loading binaries by (SD, QSPI, NOR or CCS) and follow
Uart settings to setup the console for logs and selecting demo.
2. For HDMI as input select capture source as HDMI “s: System Settings”->
“Capture Settings” -> “2: HDMI Capture 1080P60”
3. Select demo required from the menu by keying in corresponding option from the
uart menu.
IMPORTANT NOTE: Make sure you select SCV (1Ch VIP capture) use-case or ISS use-case
depending on the camera that is connected and supported
After successful initialization of the use-case, you will see video been display on the HDMI as
shown below

Page 29 of 39

a. SCV use-cases:

b. EDGE Detect use-case:

Sparse optical flow usecase

Page 30 of 39

Page 31 of 39

3.9

DCC
Dynamic Camera Configuration (DCC) tool is a PC based tool suit that is primarily
used for offline tuning of raw images obtained from raw camera sensors connected
to ISS hardware. Apart of tuning tool, DCC also contains ISP simulator.
NOTE: DCC tool can be downloaded from the below CDDS link. DCC version 2.1, compatible
with the Vision SDK 3.0 release, should be installed. Please contact local TI FAE to get access to
this CDDS link.
https://cdds.ext.ti.com/ematrix/common/emxNavigator.jsp?objectId=28670.42872.33350.26722
DCC tool is dependent on matlab runtime libraries, refer to the DCC user guide and install
required MatLab runtime.
DCC tuning tool with help of plug-ins generate a set of DCC XML and BIN files. BIN
files contain tuned values in binary file format for various ISP modules. These binary
files for different ISP modules are merged into single binary file and used in the
Vision SDK ISS sensor framework.
The binary file can also be flashed in the QSPI. Network tool command
iss_save_dcc_file can be used to save DCC binary file in the QSPI. Refer to the
network tool documentation (VisionSDK_UserGuide_NetworkTools.pdf) for more
information on this command.
When ISS use case is run in the vision sdk, it reads these binary files, parses them
and applies to the ISP modules. Vision SDK first tries to use DCC binary file from the
QSPI, if it is not available in the QSPI, it will use binary file from the ISS sensor
layer. If the binary file is not available even in ISS sensor layer, it uses ISP default
parameters.
IMPORTANT NOTE: DCC is currently supported only for AR0140, AR0132, IMX224 and
OV10640 Rev E sensors. For AR0140 sensor, the DCC xml files with the tuned ISP parameters
can be found in the path vision_sdk\apps\src\rtos\iss\src\sensor\ar0140\dcc_xml, for AR0132
sensor, they can be found in the path vision_sdk\apps\src\rtos\iss\src\sensor\ar0132\dcc_xml, for
OV10640
Rev
E
driver,
they
can
be
found
in
the
path
vision_sdk\apps\src\rtos\iss\src\sensor\ov10640\dcc_xml and for IMX224 driver, they can be
found in the path vision_sdk\apps\src\rtos\iss\src\sensor\imx224\dcc_xml
DCC tuning tool exposes key parameters for each plugin which controls important
tuning parameters. Users who are imaging and TI ISP experts can control/modify
each parameter of plugin through XML files. Use below steps for updating and
applying new tuned parameters.



Update the DCC xml file for the ISP modules for the given sensor
Convert xml file to binary file using dcc generator tool. This is a windows based
tool to convert xml file to binary file. This tool can be found from the
vision_sdk\apps\tools\dcc_tools\dcc_gen_win.exe. This tool takes name of the
xml file as an argument and generates the binary file from xml file at the same
folder where xml file is stored.

o Usage: dcc_gen_win.exe
 Send the binary file to the target using iss_send_dcc_file command of the
network tool. Please refer to the network tool documentation for getting
information on network tool
IMPORTANT NOTE: DCC xml to bin file convertor, dcc_gen_win.exe, is supported only on
Windows platform. Also this tool is dependent on DCC GUI tool, so make sure that the gui tool is

Page 32 of 39

installed on the computer before using dcc_gen_win.exe executable. Please contact local TI
support to get DCC GUI tool.
Once the tuned parameters are tested and finalized, they can be permanently stored
in the QSPI or in driver.
For storing tuned parameters in the QSPI, run iss_save_dcc_file command of the
network
tool.
Please
refer
to
the
network
tool
documentation
(VisionSDK_UserGuide_NetworkTools.pdf) for more information about this command.
This command saves the binary file at the fixed offset in the QSPI. After saving the
binary file in QSPI, restart the ISS usecase to check the output of the tuned
parameters.
For storing tuned parameters in the sensor driver, go to the dcc_xml folder under
sensor
driver.
For
example,
for
AR0140
sensor,
go
to
the
vision_sdk\apps\src\rtos\iss\src\sensor\ar0140\dcc_xml\. This folder contains all the
xml files that vision sdk is using for this sensor. Copy the updated the xml file under
this folder and run generate_dcc.bat file from the windows. This batch file converts
all xml files into binary files, merges all binary files to single binary file and converts
binary file to header file, which will be used by the driver. After running this batch,
restart the ISS usecase to check the output of the tuned parameters.
Below is list of the DCC plugins supported in vision SDK.
ISIF_CLAMP

DC Offset/Black Level offset in the ISIF
This plugin is currently used only for setting blank level offset.
No other parameters from this plugin are used.

IPIPE_GIC

Green Imbalance Correction Module of IPIPE

IPIPE_NF1

Noise Filter 1 module of the IPIPE

IPIPE_NF2

Noise Filter 2 module of the IPIPE

IPIPE_DPC_OTF

Defect correction OTF module

IPIPE_CFA

Color Filter Array module

IPIPE_Gamma

Gamma Correction module

IPIPE_RGB2RGB1

RGB to RGB color correction module-1
Supports multi photospace*, which means multiple set of
parameters can be defined based on the photospace for this
module.

IPIPE_3D_LUT

3D Lut module

IPIPE_RGB2RGB2

RGB to RGB color correction module-2
Supports multi photospace*

IPIPE_RGB2YUV

RGB to YUV Color Conversion module

IPIPE_EE

Edge Enhancer module

IPIPEIF_SPLIT

VP Decompanding module of the IPIPEIF

IPIPEIF_WDRMERGE

WDR Merge and WDR Companding module of the IPIPEIF
Most of the WDR merge parameters are calculated on the fly
based on the exposure ratio, so the only WDR merge
parameter used from this plugin are enable and black level for
long and short exposure.

GLBCE

GLBCE module

Page 33 of 39

NSF3V

NSF3v module
Supports multi photospace*

CNF

Chroma noise filter module
Supports multi photospace*

AWB_ALG

AWB Calibration Parameters
TI AWB algorithm requires these calibration parameters. If not
provided, it uses default calibration parameters

LDC

Lens Distortion Correction Module

* Photospace is defined by three parameters, exposure time, analog gain and color temperature. A range
of these parameters creates one photospace. Refer the DCC Gui documentation to get more details on
how to create photospace
* Although multiple photospace is supported only for few modules, xml files for almost all modules could
have multiple set of parameters based on the multiple photospace. If the module does not support
multiple photospace and xml file contains multiple set of parameters, only the first parameter set is used
by the parser.

Updating Mesh LDC table in Vision SDK:
For the new fisheye lens, follow below steps to update the mesh LDC table
a. Get the new LDC table for the new Fisheye lens
b. Go to the Vision_sdk\apps\tools\LDC_mesh_table_convert\ directory in the vision
SDK
c. Run the perl script convert.pl as shown below
perl convert.pl input_table.txt imagewidth imageheight downscalefactor
Here inputtable.txt file contains mesh LDC table for the new lens
imagewidth and imageheight are size of the input image
downscalefactor is the down scale factor by which input table is down scaled.
d. It will generate the file input_table.bin file, convert this bin file to header file
using
bin2c
Vision
SDK
utility,
it
can
be
found
under
vision_sdk\apps\tools\dcc_tools\bin2c.exe
e. Replace
this
header
file
in
vision_sdk\apps\src\rtos\iss\src\sensor\iss_tables\iss_tables_ldc_lut_1920x1080.
h
f.

Rebuild vision sdk

Page 34 of 39

3.10

Fast boot usecase
This usecase is mainly targeted for rear view camera systems and mainly
demonstrates how boot time can be optimized to show sensor capture output on
display (preview) first on power on reset and then switch to analytics output shown
on display.
As the execution sequence for this usecase is different than all other usecases, it is
not enlisted in console RunTime Menu.
It is a fixed configuration demo usecase which works when you press reset button on
the TDA3X EVM.

3.10.1 Usecase configuration
It supports following configuration _Only_
1CH ISS Capture + ISP + LDC + Obj detect + Display


Capture - AR0140 Parallel with TDA3x EVM



Display - 10 inch LCD



Boot mode - QSPI

3.10.2 Hardware set up
Refer section 0 “Required H/W modification / Configurations” to understand board
modification needed for TDA3X with and above mentioned usecase configuration. It
is important to have this done before fast boot usecase is tried. H/w mods for
following cannot be skipped.


I2C to run at 400KHz



Support for AR0140 or OV10640 REV E

Figure: TDA3x EVM Fast boot h/w setup

Page 35 of 39

3.10.3 Build
Fast boot is special usecase demonstrating how boot time can be optimized for any
vision_sdk usecase; idea here is to have preview display up in minimum possible time
and then switch to actual usecase.
 The usecase is not enlisted in runtime menu it can be enabled using following
variable in \vision_sdk\apps\configs\tda3xx_evm_bios_all\cfg.mk. By default it is “no”.
# Fast boot usecase is currently supported only for tda3x
FAST_BOOT_INCLUDE=yes
…
Remove DSP2 and IPU1_1 from the \vision_sdk\apps\configs\tda3xx_evm_bios_all\cfg.mk
& define WDR_LDC_INCLUDE to “yes”
PROC_DSP2_INCLUDE=no
PROC_IPU1_1_INCLUDE=no
NDK_PROC_TO_USE=none
WDR_LDC_INCLUDE=yes
Important Note: These can be defined ‘yes’ even in fast boot usecase but they are
not needed for this usecase and can contribute to boot time hence removed from
build config. User may enable these as per their usecases.
Important Note: In order to test the DSP and EVE analytics off and on options one
must make sure to not include IPU1_1 in the build and the PROC_IPU1_1_INCLUDE
should be ‘no’




“gmake -s showconfig” command can be used prior to build to confirm the build
configuration.
To build use
o gmake -s -j depend
o gmake -s –j
SBL also needs to be built for fast boot usecase
- Ensure to remove SBL binaries if SBL was built previously
- Use “gmake -s sbl” to build SBL

3.10.4 Generating and Flashing images
 Refer section 3.5.1 to generate sbl


To generate the application image run below command from “vision_sdk\build” folder
> gmake -s appimage
IMPORTANT NOTE: The config options, like CPUs to use, debug or release profile,
fast boot enable etc, used to make the application image will be the values specified
in \vision_sdk\apps\configs\tda3xx_evm_bios_all\cfg.mk




This command should generate AppImages at \vision_sdk\binaries\
apps\tda3xx_evm_bios_all\vision_sdk\bin\tda3xx-evm\sbl_boot
o

AppImage_UcEarly_BE

AppImage_UcLate_BE

Refer section 3.5.3 to flash following images at given offsets.

Page 36 of 39

Image
sbl_qspi
AppImage_UcEarly_BE
AppImage_UcLate_BE

QSPI offset to be flashed in
0x0
0x80000
0xA80000

Important Note: Ensure images are flashed at given offsets only, order is not
mandatory
3.10.5 Run
Press Power On Reset button on Tda3x EVM. Make sure QSPI boot is selected as
mentioned in section 0
Pass criteria


Display should flash up with preview in 1 sec approx



Use case should switch to Object detect algorithm and Pedestrian / Traffic signs
detection should start as soon as they are in field of view after boot up.



You should see boot time printed on the LCD below the CPU performance bar.



In order to run the analytics ON (option 3) and OFF (option 4) scenario one can
choose to select any of the highlighted menu options. The display shows the
status of the PD+ TSR Object detection above the CPU Performance bar.

1: Save Captured Frame
2: Save SIMCOP Output Frame
3: PD and TSR ON
4: PD and TSR OFF
Important Note: Ensure IPU1_1 image is not a part of the application image when
trying these two options.

3.11

Surround View Fast Boot Use case
The

following

compile

option

\vision_sdk\$(MAKEAPPNAME)\configs\$(MAKECONFIG)\cfg.mk should be used to enable
fast boot for 3D Surround View:
SRV_FAST_BOOT_INCLUDE=yes
This option builds the Application Image involving only ipu1_0 and DSP_1 cores.
Important Note: Ensure the post-calibration tables are present on the SD Card that
is being used to run the usecase. The details of such tables can be found in the
relevant SRV UserGuides.

3.12

Surround View Use case under 128 MB DDR configuration
The following changes are required to be made before building and running the
Surround View use cases (calibration, 2D and 3D) with the 128 MB DDR
configuration:
1. Set ‘DDR_MEM=DDR_MEM_128M’ in the platform ‘cfg.mk’ file
2. Remove EVE1, IPU1_1 and DSP2 from the build in the platform ‘cfg.mk’ file:


PROC_EVE1_INCLUDE=no

Page 37 of 39



PROC_IPU1_1_INCLUDE=no



PROC_DSP2_INCLUDE=no



ECC_FFI_INCLUDE=no

3. Make the following changes to the memory sections
in the
‘\vision_sdk\$(MAKEAPPNAME)\build\tda3xx\mem_segment_definition_12
8mb.xs’ file:


EVE1_VECS_SIZE



EVE1_CODE_SIZE

= 128*KB;



EVE1_DATA_SIZE

= 128*KB;



IPU1_1_CODE_SIZE

= 128*KB;



IPU1_1_DATA_SIZE

= 128*KB;



IPU1_0_CODE_SIZE

10*MB;



IPU1_0_DATA_SIZE

10*MB;



DSP1_CODE_SIZE

1*MB;



DSP1_DATA_SIZE

11*MB;



DSP2_CODE_SIZE

= 128*KB;



DSP2_DATA_SIZE

= 128*KB;



SR1_FRAME_BUFFER_SIZE

90*MB;



SR1_BUFF_ECC_ASIL_SIZE

4*KB;



SR1_BUFF_ECC_QM_SIZE

4*KB;



SR1_BUFF_NON_ECC_ASIL_SIZE

4*KB;

1*KB;

4. Increase the size of SR1 heap to 90 MB in the memory map file (.xs).
5. Exclude the use cases which are out of scope for 128 MB build by setting
‘UC_=no ’ in the platform ‘uc_cfg.mk ’ file as shown below, so
that the sizes of the IPU1_0 and DSP1 code and data fits within the
corresponding
memory
segments
defined
in
the
‘\vision_sdk\$(MAKEAPPNAME)\build\tda3xx\mem_segment_definition_12
8mb.xs’ file.
The below use-cases are validated with 128 MB memory configuration.
UC_saveDisFrame=yes
UC_srv_calibration=yes
UC_iss_mult_capture_isp_simcop_sv_tda3xx=yes
UC_iss_mult_capture_isp_dewarp_3dsv_tda3xx=yes
UC_iss_capture_isp_simcop_display=yes
UC_vip_single_cam_view=yes

Page 38 of 39

UC_vip_single_cam_view_dsswb=yes
UC_vip_single_cam_display_metadata=yes

Revision History
Version

Date

Revision History

1.0

22th August 2014

2.0

14 November 2014

3.0

31 December 2014

4.0

3 March, 2015

5.0

28 June 2015

6.0

16 Oct 2015

7.0

18 March 2016

8.0

18 October 2016

9.0

February 2017
th

10.0

15 Apr 2017

11.0

19 June 2017

12.0

29 June, 2017

13.0

13 October 2017

14.0

20 December 2017

15.0

29 Mar 2018

Initial Version
Added QSPI+SD boot,
GCC and CCSversion
Some minor changes
Added new section DCC
Added fast boot usecase
Updated for release 2.8
Updated for release 2.9
Minor changes and
updated for 2.11
Minor changes and
updates for vision sdk
2.12
Updated for 128M build

Updated linux installer

Updated for Processor
SDK Vision 3.0 release
Updated for Processor
SDK Vision 3.01 release
Update for Processor
SDK vision 3.02 release
Updated 128MB map file
change

««« § »»»

Page 39 of 39

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.5
Linearized                      : No
Page Count                      : 39
Language                        : en-US
Tagged PDF                      : Yes
Title                           : VisionSDK_UserGuide_TDA3xx
Author                          : TI
Creator                         : Microsoft® Word 2010
Create Date                     : 2018:04:03 17:14:23+05:30
Modify Date                     : 2018:04:03 17:14:23+05:30
Producer                        : Microsoft® Word 2010

EXIF Metadata provided by EXIF.tools

VisionSDK_UserGuide_TDA3xx Vision SDK User Guide TDA3xx

Navigation menu

Versions of this User Manual:

Views

Navigation