Please note as of Wednesday, August 15th, 2018 this wiki has been set to read only. If you are a TI Employee and require Edit ability please contact x0211426 from the company directory.

DM814x AM387x Power Estimation

From Texas Instruments Wiki
Jump to: navigation, search

Power Estimation Spreadsheet

Abstract

This article discusses the power consumption of the Texas Instruments TMS320DM814x DaVinci™ digital media processors and AM387x Sitara™ ARM® microprocessors. For the rest of the document, the word "device" represents the device chosen by the user (DM814x or AM387x).

Power consumption on the device is highly application-dependent, therefore a spreadsheet is provided to model power consumption for a user’s application and to present some measured scenarios. Version 1.0 of the spreadsheet supports configurability of device core modules such as the ARM Cortex-A8, C674x and most peripherals. The data in the accompanying spreadsheet are indicative of the expected maximums of power consumption for production units. Thus, the spreadsheet values may be used for board thermal analysis and power supply design as a maximum long-term average.

The data presented in the Version 1.0 power estimation spreadsheet are based on measurements performed on DM8148 revision 2.1 silicon.

The Power Estimation Spreadsheet

The spreadsheet applies to DM814x and AM387x devices. In the document, the word "device" represents the device chosen by the user (DM814x or AM387x). The power spreadsheet is available here: File:DM814x AM387x PowerSpreadSheet RevE0.zip - Updated 08/31/12

Modeling Power for an Active Scenario

Power consumption estimates in the spreadsheet are separated by the major modules of the device, so that their contribution can be gauged independent of each other. The options in the spreadsheet are intended to provide some flexibility in customizing the worst case power consumption estimate for a specific application.

Power consumption in the device consists of:

  • static power – due to leakage current, and dependent on temperature; and
  • active power – due to transistor switching, and independent of temperature.

Version 1.0 of the DM814x Power Estimation Spreadsheet models temperature effects on device power consumption. The user is allowed to specify case temperature of the device to be used in the power estimate and also the Core OPP(can choose from 100,120) at which the user wants to operate the device. This is used to scale the baseline power estimate in the spreadsheet, which consists of device static power and power consumption from interconnects and clock trees.

Active power consumption is the power that is consumed by portions of the device that are performing some processing. Active power is independent of temperature, but dependent on voltage and module activity levels. Version 1.0 of the DM814x Power Estimation Spreadsheet allows configuration of activity for some modules in the DM814x device. For other modules, measured scenarios are provided for the module, and the user is expected to select the usage scenario most closely aligned to the intended application.

Using the Power Estimation Spreadsheet

Note:

The power spreadsheet uses circular references. To avoid any error, ensure that the power spreadsheet is the first spreadsheet in an "Excel session". In other words, the MS Excel should not be open before you open the power spreadsheet.

Using the power estimation spreadsheet involves entering appropriate usage parameters for modules. Cells that are designed for user input are white in color. To use the spreadsheet, simply configure the white cells to a value most closely aligned with your intended scenario. The spreadsheet will display the details of power consumption for the configuration selected.

The following sections shall explain the usage of the spreadsheet, specifically the cells that need user inputs.

Device level entries

This section explains the entries that are not module-dependent.

1. First-of-all, the user needs to select the device intended to be used in the end application. This can be done by selecting the appropriate button for the cell marked "DEVICE".

2. Enter the case temperature of the device in the cell marked "Case Temp".

3. Enter the Core OPP of the device in the cell marked “Core OPP”. User can choose the OPP as follows.

Device

CVDD_ARM

CVDD

VDD_IVA

VDD_GEM

OPP 100 1.1V 1.1V 1.1V 1.1V
OPP 120 1.2V 1.2V 1.2V 1.2V
OPP 166 1.35V 1.35V 1.35V 1.35V

4.. Based on the above inputs, a baseline power value is calculated which is the static power of Active power domain.

5. The device selection buttons at top (Jacinto 5, Auto Vision, Stand-by, Deep Sleep) can be used to quickly generate the power estimate for typical use cases of these devices.

Module level entries

This section illustrates the various entries expected for the modules. The spreadsheet presents information for the full-featured DM814x device in the DaVinci family. However, since the spreadsheet breaks out the power consumption by each module in the DM814x, estimates for the AM387x can be obtained by choosing the "Off" option in status for the modules not present in the device under consideration. This will ensure active and static power from the non-applicable module is not included in the power tally. The calculations in the power spreadsheet assumes that with exception of Always-on power domain, the rest of power domains are switched off when all the modules corresponding to that power domain are not used. For more details on Power domains, please refer to the Power Management section of the device TRM.

The end user needs to take care of the pin muxing while enabling the modules in the spreadsheet. If two modules are exclusive of each other (due to pin muxing), only one of them needs to be enabled in the spreadsheet, based on the end use case.

The frequency for some of the modules is dependent on the device option chosen. The current maximum device frequencies (in MHz) supported are mentioned in the table below. Please note that in some of the instances the module frequency is quoted higher than the frequency mentioned in device datasheet. This enables the end user to get a feel of power increase with higher frequency of module operation. However, the module functionality is guaranteed only till the frequencies mentioned in the datasheet.


Device

ARM

C674x

HDVICP2

HDVPSS

ISS

SGX

Max Voltage

FT Bin

Max OPP

DDR

Min Temp

Max Temp

TMS320DM8148BCYE0 600 500 266 200 400 200 1V2 0 120 400 -40 105


TMS320DM8148BCYE1 1000 750 410 200 480 280 1V35 1 166 533 -40 105
TMS320DM8148BCYE2 1000 850 500 200 560 280 1V35 2 166 533 -40 105
TMS320DM8147BCYE0 600 500 266 200 400 0 1V2 3 120 400 -40 105
TMS320DM8147BCYE1 1000 750 410 200 480 0 1V35 4 166 533 -40 105
TMS320DM8147BCYE2 1000 850 500 200 560 0 1V35 5 166 533 -40 105
TMS320DM8127BCYE0 600 500 266 200 400 0 1V2 6 120 400 -40 105
TMS320DM8127BCYE1 720 750 410 200 480 0 1V35 7 166 533 -40 105
TMS320DM8127BCYE2 1000 850 500 200 560 0 1V35 8 166 533 -40 105


Cortex A8

1. Status: If Cortex A8 is used, select "Active", else if A8 is in WFI (stand-by) state, select "Idle"

2. Frequency: Enter the frequency of Cortex A8

3. Utilization %: Enter the percentage of Cortex A8 CPU utilization in your use case

4. Neon State: Enter the state of Neon. Choose “Idle” if Neon is not used, otherwise choose “Active“.

C674x

1. Status: If C674x is used, choose "Active" else choose "Off"

2. Frequency: Enter the frequency of C674x

3. Utilization %: Enter the percentage of C674x CPU utilization in your use case

Media Controller

The power of Media Controller is included in the power numbers of HDVICP and HDVPSS.

HDVICP2

1. Status: If HDVICP2 is used, choose "Active", else choose "Off"

2. Frequency: Enter the frequency of HDVICP2
While reducing the frequency of HDVICP, ensure the resolution and/or frame rate is changed accordingly.

3. Resolution: Enter the resolution of encode/decode being performed in HDVICP2

4. Enc/Dec: Select HDVICP2 is using H.264 Enc or H.264 Dec

5. Number of channels: Enter the number of channels/instances of Encoder/Decoder executed on HDVICP2

6. FPS per channel: Number of frames per second per channel/instance of Encoder/Decoder

7. Max FPS* No. Channel: This value can not be changed by the user. It indicates the upper limit for FPS * No. channels based on the resolution and Encoder/Decoder being used in the HDVICP2.

Note:

When the same HDVICP is used for encode and decode operation, it is recommended to use Encode power for sum of encoder and decoder channels. For example, if an application uses X instances of encode and Y instances of decode on the same HDVICP, the maximum power can be obtained by configuring the HDVICP to encode X+Y channels/instances

HDVPSS

1. Status: If HDVPSS is used, choose "Active", else choose "Off"

2. Scenario: The following use cases are available to be chosen from. For use case not listed below, add the power of all the components of the HDVPSS use case.

  • Display: Onchip HDMI  : Only on-chip HDMI is used for Display (capture and other processing is not used)
  • Display: Offchip HDMI  : Only off-chip HDMI (DVO) is used for Display (capture and other processing is not used)
  • Display: SDVENC  : Only SDVENC Display is used (capture and other processing is not used)
  • Tridisplay (without GPX): Three displays (on-chip, offchip HDMI and SD VENC) without Graphics
  • Tridisplay (with GPX)  : Three displays (on-chip, offchip HDMI and SD VENC) with Graphics
  • Capture: 16 D1  : 16 channel D1 capture is used (display and other processing is not used)
  • Capture: 8 D1  : 8 channel D1 capture is used (display and other processing is not used)
  • Capture: 4 D1  : 4 channel D1 capture is used (display and other processing is not used)
  • Capture: 2 1080 I  : 2 1080 I channel capture is used (display and other processing is not used)
  • Capture: 1 1080p  : 1 1080p channel capture is used (display and other processing is not used)
  • DEI_H  : DEI Primary path with scalar bypassed
  • DEI_M  : DEI Auxillary path is enabled
  • NF  : Only Noise filter is enabled
  • Scalar  : Only 1 scalar is enabled
  • 1 ch DVR  : Typical capture, processing and display for 1 channel DVR scenario
  • 4 ch DVR  : Typical capture, processing and display for 4 channel DVR scenario
  • 8 ch DVR  : Typical capture, processing and display for 8 channel DVR scenario

SGX530

1. Status: If SGX530 is used, choose "Active", else choose "Off"

2. Frequency: Enter the frequency for SGX530

3. Frames per Sec (1080p): Enter the number of 1080p frames processed by SGX. If any other resolution is used, the total processing need to be converted to appropriate 1080p frames per second

EDMA

1. Status: If EDMA is used, choose "Active", else choose "Off"

Note:
Since the power of EDMA is low, it is not split further into the number of EDMA channels being used.

EMIF_0, EMIF_1

1. Status: If EMIF is used, choose "Active", else choose "Off". Please note that EMIF_1 can not be independently marked Active/Off. It follows the status of EMIF_0. Similarly, the frequency of EMIF_1 is same as EMIF_0.

2. Frequency: Enter the frequency for EMIF operation. The upper frequency limit for DDR3 is 533 MHz and for DDR2 is 400 MHz.

3. DDR2/DDR3: Select the memory type (DDR2/DDR3) being used.

4. Transfer Rate (MBps): Specify the transfer rate for each EMIF in terms of Mega Bytes per Second. The upper limit for the transfer rate is fixed at 2240 MBps at 400MHz (assuming 70% efficiency)

5. Write %: Specify the percentage of data being written to DDR memory with respect to the Transfer Rate mentioned above. (Rest of the data is assumed to be read from DDR memory)

6. Power Save Mode:- Choose On if you want to Enable the EMIF power saving mode.

The unused DDR interface can be put into least power consuming state by the configuring it as follows

  • Enable the PHY Clock gating for unused PHY, by programming the EMIF_CLK_GATE register in the chip control module.
  • Disable VTP calibration for the unused DDR interface. This can be achieved by clearing the ENABLE bit in the DDR_VTP_CTRL_n register the chip control module.
  • Enable the dynamic power down for the IOs of the unused DDR interface. This can be done by setting the DYN_PWRDN_EN bit in the corresponding DDRPHYCR register.
  • Disable ODT for the unused DDR interface by selecting the following configurations in the corresponding DDRPHYCR register.
    • IDLE_LOCAL_ODT = 0
    • RD_LOCAL_ODT =0
  • Configure the output impedance of the unused DDR IOs to be the highest (Or lowest Drive strength) by selecting the following configurations in the DDRn_IO_CTRL register in the chip control module.
    • DATA_IMPEDENCE=0
    • COMMAND_IMPEDENCE=0
    • CS_IMPEDENCE=0
Note:

For the same frequency of operation and data transfer rate, DDR3 consumes much less power than DDR2.

PCIe

1. Status: If PCIe is used, choose "Active", else choose "Off"

2. Mode: Select the mode of operations for PCIe (Gen 1 , Gen 2(Full Duplex: 1X, Full Duplex: 2X))

SATA

1. Status: If SATA is used, choose "Active", else choose "Off" 2. Mode: Select the mode of operations for SATA (Gen 1 , Gen 2)

USB

1. Status: If USB is used, choose "Active", else choose "Off"

2. No. of Instances: Select the number of USB Instances used

EMAC

1. Status: If EMAC is used, choose "Active", else choose "Off"

2. No. of Instances: Select the number of EMAC Instances used

Mc ASP

1. Status: If Mc ASP is used, choose "On", else choose "Off"

2. No. of Instances: Select the number of Mc ASP instances used(McASP0,1,2,3,4,5)

DCAN

1. Status: If DCAN is used, choose "On", else choose "Off"

UART

1. Status: If UART is used, choose "On", else choose "Off"

VCP

1. Status: If VCP is used, choose "On", else choose "Off"

MLB

1. Status: If MLB is used, choose "On", else choose "Off"

GPMC

1. Status: GPMC is used, choose "On", else choose "Off"

MMC

1. Status: If MMC is used, choose "On", else choose "Off"

2. No. of Instances: Select the number of MMC Instances used

SPI

1. Status: If SPI is used, choose "On", else choose "Off"

2. No. of Instances: Select the number of SPI Instances used

PATA

1. Status: If PATA is used, choose "On", else choose "Off"

FAQ

Q: Why do I get circular reference error when I open the power spreadsheet?
A: The power spreadsheet uses circular references. To avoid any error, ensure that the power spreadsheet is the first spreadsheet in an "Excel session". In other words, the MS Excel should not be open before you open the power spreadsheet.

Q: Why I don't get value update after changing the parameter(s) in the power spreadsheet?
A: Please make sure the MS Excel Macros has been enabled and the formulas has Iterative calculation enabled with Maximum Iterations (1000) and Maximum Change (0.5).

Q: What is the estimated current draw for 0.9V (USB Logic)? It is not specified in the datasheet nor in the power estimation spreadsheet.
A: It is very small (just a few mA) compared to the other power supplies. Hence. it was not included as a separate item.

Q: Can you provide information about the Power Management IC solutions suggested for use with these devices?
A: Please contact your local analog power FAE for the latest information on suggested PMIC solutions.

Q: Do the DDR2/DDR3 power estimates include the power consumption of the DDR devices or just the interface on the processor?
A: The power spreadsheet DOES NOT include the power consumed by DDR memories, only the processor.

Q: Why is case temperature used in the power spreadsheet and not junction temperature as in the datasheet?
A: It is easier for the end user to measure the case temperature of the device.

Q: The datasheet indicates that some kind of heat dissipative solution is required for this device. What heat dissipative solution was used for the power measurements in this spreadsheet?
A: Thermal inducing systems were used for regulating the case temperature during power measurement. This is needed only for power characterization at given temperature. However, the end user needs to use the heat dissipating solution while using the device on their board.

Q: Is SmartReflex (AVS) implemented in the power spreadsheet?
A: No, not in the current version. The option to include AVS functionality will be included in a future version.


Q: Is the “Power Save Mode” in the power spreadsheet EMIF section for ODT on/off/partial? How is this set up?
A: The power save options for EMIF are

1. Dynamic receiver power down enable
2. Disable ODT – This is not a generic recommendation though. Choose this mode only after properly running a board simulation with no terminations on DQ/DQS/DM signals and analyzing the signal integrity.
3. Reduce the DDR IO drive strength. The caution mentioned in #2 applies here as well.

Also, see comments from the power app note at: EMIF_0, EMIF_1 sect 6 on this wiki page.