Professional Documents
Culture Documents
• AM57x Processors
• AM57x Peripherals
Software Compatible
MHz MHz MHz MHz (Quad Core)
Pin Compatible
Video Acceleration
PRU (Quad Core)
+ - * = IVA HD 1080p Video, VPE
ARM
C66x DSP
ARM
Cortex-A15 + - * = Graphics Acceleration Industrial
Cortex-A15 ARM M4 32K/32K L1
32K/32K L1
C66x DSP
288KB L2
3D GPU BB2D Communication
32K/32K L1 ARM M4 32K/32K L1 2x SGX544 GC320
288KB L2 Subsystem (ICSS)
EtherCAT®,PROFINET®,
1MB L2
L2 32KB L1
2MB
w/ECC 64KB RAM Display Subsystem EtherNET/IP™,
PROFIBUS,
3 LCD HDMI 1.4a POWERLINK, SERCOS 3
512KB L3 Shared RAM w/ECC
1080p
Blend/Scale/ Convert
32b DDR3/3L
32b DDR3/3L
w/ ECC Security Acceleration
Video Input Ports 2 AES, 2 SHA2MD5,
2x24b, 2x8b 2x24b, 2x8b 2x16b DES3DES, RNG
System Services
EDMA 13 Mailbox RTC SDMA Secure WDT WDT Spinlock 16 Timer KBD
Video Acceleration
PRU (Quad Core)
IVA HD 1080p Video, VPE
ARM + - * = Graphics Acceleration Industrial
Cortex-A15 ARM M4 C66x DSP 3D GPU BB2D Communication
32K/32K L1 ARM M4 32K/32K L1 SGX544 GC320
288KB L2 Subsystem (ICSS)
EtherCAT®,PROFINET®,
1MB
1MB L2
L2 32KB L1 Display Subsystem
w/ ECC
w/ECC 64KB RAM EtherNET/IP™,
3 LCD HDMI 1.4a PROFIBUS,
512KB
512KBL3L3Shared RAMw/ECC
Shared RAM w/ ECC 1080p POWERLINK, SERCOS 3
Blend/Scale/ Convert
System Services
EDMA 13 Mailbox RTC SDMA Secure WDT WDT Spinlock 16 Timer KBD
ABC Package:
23 x 23mm, 0.8 pitch, 760-pin BGA package
AM57x Processors
• 64-byte (vs 32) cache line – 2nd stage MMU for • 1.4x higher instructions per
Hypervisor control of cycle (IPC) from enhancements
• Dual load/store (vs one or other)
guest OS memory • 1.4x faster floating point
• Improved branch prediction:
– CP15 trapping • 10-15% higher clock in same
– Higher capacity
• Debug/trace support: process due to design
– Support for indirect branches
– Integrated trace • 10-100x faster fully-pipelined
• More out-of-order instructions cache maintenance support
• Physically-indexed/tagged L1 cache – Virtualization support
• Improved system-level support
• Tighter integration with NEON/VFP: • AMBA4 bus supports:
to support new architecture
– Faster interworking with ARM – System coherency needs: Larger memory,
code – MMU coherency virtualization, system
– Dual-issue (vs single) coherency…
• Improved memory performance:
– Tightly-coupled L2 cache to
reduce latency 23 to 14 clocks
– Enhanced auto-prefetch
– More requests buffering
Cortex-A15 offers substantial enhancements and new features to dramatically increase performance
and system-level support.
Processors and Memory: M4, DSP, & L3
Dual (AM572x) / Single (AM571x) ARM Cortex-A15
• Up to 1.5 GHz, r2p2 revision core(s), ARMv7-A instructions set
• Out-of-order instruction dispatch and completion
• Backward-compatible with code for previous ARM processors
• Integrated NEON™ processing engine and VFPv4 compatible hardware
• Five execution units handle simple instructions, branch instructions,
NEON and floating point instructions, multiply instructions, & load and
store instructions AM5728 diagram shown
• AM572x has 2MB L2 memory (no ECC), while AM571x has 1MB L2
with ECC
Dual-ARM Cortex M4s
• Up to 213 MHz, ARMv7-M and Thumb®-2 ISAs w/ ARMv6 SIMD & DSP
extensions
Dual (AM572x) / Single (AM571x) C66x DSPs
• Up to 750MHz, fixed- and floating-point ISA
• Object code compatible with C64x+ and C674x DSPs
• Advanced VLIW architecture w/ two multiplier units and six arithmetic
logic units operating in parallel
L3 Memory with ECC
• AM572x has 2.5MB of L3 memory with ECC
• AM571x has 512 KB of L3 memory with ECC
Processors and Memory: PRU-ICSS
Programmable Real-Time Unit – Industrial Communication Subsystem (PRU-ICSS) (2)
• Dual 32-bit RISC cores
• 12KB program RAM, 8 KB data RAM per CPU Programming Tools
• 32KB shared RAM •PRU C-compiler for PRU firmware
• Interrupt controller •ARM Linux remoteproc + rpmsg driver
• Fast IO interface •PRU debugger in CCS
• Peripherals:
– One Ethernet MII_RT module with two MII ports *
– One MDIO port *
– One Industrial Ethernet peripheral *
and Industrial Ethernet timer
– 1 x 16550-compatible UART
– 1 x eCAP
• Capable of supporting master and/or slave modes
of protocols such as:
– Profinet, Ethernet IP, Profibus,
Ethercat, Powerlink, and Sercos 3
* Use of these ports is only supported via TI Processor SDK RTOS.
PRU-ICSS Feature Comparison
AM18x/
AM335x AM437x AM571x AM572x (SR1.1)
Features OMAPL138
PRUSS PRU-ICSS1 PRU-ICSS1 PRU-ICSS0 2 x PRU-ICSS 2 x PRU-ICSS
PRU core version 1 3 3 3 3 3
Number of PRU cores 2 2 2 2 2 2
Max frequency CPU freq / 2 200 MHz 200 MHz 200 MHz 200 MHz 200 MHz
IRAM size (per PRU core) 4 KB 8 KB 12 KB 4 KB 12 KB 12 KB
DRAM size (per PRU core) 512 B 8 KB 8 KB 4 KB 8 KB 8 KB
Shared DRAM size -- 12 KB 32 KB -- 32KB 32KB
Direct; or 16-bit Direct; or 16-bit Direct; or 16-bit
Direct; or 16-bit parallel capture; or parallel capture; or parallel capture; or Direct; or 16-bit
General purpose input
Direct parallel capture; 28-bit shift; or 3ch 28-bit shift; or 3ch 28-bit shift; or 3ch parallel capture; or
(per PRU core)
or 28-bit shift EnDat 2.2; or 9ch EnDat 2.2; or 9ch EnDat 2.2; or 9ch 28-bit shift;
Sigma Delta Sigma Delta Sigma Delta
General purpose output
Direct Direct; or Shift out Direct; or Shift out Direct; or Shift out Direct; or Shift out Direct; or Shift out
(per PRU core)
GPI Pins (PRU0, PRU1) 30, 30 17, 17 13, 0 20, 20 21*, 21 21, 21
GPO Pins (PRU0, PRU1) 32, 32 16, 16 12, 0 20, 20 21*, 21 21, 21
MPY/MAC N Y Y Y Y Y
Scratchpad N Y (3 banks) Y (3 banks) N Y (3 banks) Y (3 banks)
CRC16/32 0 0 2 2 2 0
INTC 1 1 1 1 1 1
Peripherals n/a Y Y Y Y Y
UART 0 1 1 1 1 1
eCAP 0 1 1 not pinned out 1 1
IEP 0 1 1 not pinned out 1 1
MII_RT 0 2 2 not pinned out 2 2
MDIO 0 1 1 not pinned out 1 1
Simultaneous protocols 1 1 2** 2
* PRU-ICSS2 only. PRU-ICSS1 does not pin out the PRU0 core GPIs/GPOs.
** 2nd protocol limited to EnDAT/Profibus/BISS/HIperphase DSL or serial based protocol
AM57x Peripherals
• QSPI
– Master-only interface primarily intended for fast booting from Serial IO
quad-SPI flash memories 2 8
– Supports single, dual or quad reads. Only single writes are supported. DCAN McASP
• Clocking:
– MII/RMII: Supports both internal and external
50MHz reference clock
– RGMII: Internal clock reference only
• No support for:
– GMII interface
– 2.5V Signaling
• Single Port SATA host controller supporting 1.5-Gbps and 3-Gbps speeds (SATA-1 and SATA-2)
• Supports multiple drives with a port multiplier: command-based switching only (Issuing commands
to only one drive at a time)
MMC1 4-bit SD 192 MHz (via DLL) UHS1 Master & Slave
MMC2 8-bit eMMC 192 MHz (via DLL) LVCMOS Master & Slave
MMC3 8-bit SDIO / SD 96 MHz LVCMOS Slave
MMC4 4-bit SDIO / SD 48 MHz LVCMOS Slave
* AM572x errata – MMC1/2/3 write speed limited.
• Full compliance with standards: Refer to device errata for impacted silicon revisions.
– JC64 MMC/eMMC standard specification, v4.5
– SD Physical Layer specification v3.01
– SD part E1 specification v3.00 (SDIO)
– SD card specification Part A2 v3.00
• No support for:
– MMC POW output pin: Must utilize chip level GPIO for this function, if desired.
Storage I/Os: GPMC & ELM
• General Purpose Memory Controller (GPMC) GPMC
– Used for accessing SRAM, NOR, NAND, etc.
– 8/16-bit data at up to 88MHz
– Non-muxed, Address-Data muxed, and
Address-Address-Data muxed modes
– Async mode with read page access
– Sync mode with burst access and wrap
capability
– 8 chip-selects covering 512MB of address
space
• Error Locator Module (ELM)
– Used when interfacing GPMC to a NAND
device
– Provides 4-, 8- or 16-bit error location over a
512-Byte block based on BCH algorithms
– Allows detected errors and their locations in
the NAND block to be retrieved by the
processor when the PAGE_VALID interrupt is
generated
Storage I/Os: EMIF & DMM
• External Memory Interface (EMIF)
Number EMIF ECC Addressable SDRAM size Chip Selects
Controllers
AM572x Dual EMIF1 only Up to 2GB per controller 1 per controller
AM571x Single Yes Up to 4 GB * 2
* if using stacked die package
– 16-/32-bit DDR3/ DDR3L support
– Speeds up to 533MHz (DDR-1066)
– Hardware-leveling support
– Class of Service and burst priority counter
– MPU MA has 128-bit direct path with
optimized latency to each of EMIF0/1
System Services
EDMA 13 Mailbox RTC SDMA Secure WDT WDT Spinlock 16 Timer KBD
AM57x System Architecture
• The AM57x Data Manual lists the pad configuration requirements to achieve
the various timing modes of operation.
• The TRM Pad Configuration Section provides the details on implementing the
pad configuration requirements.
Virtual IO Timing Modes
• Virtual IO Timing Modes are pre-defined IO timing settings that are coded in the
Device ROM.
• Selection of Virtual Modes is done via the Pad Configuration Registers. This is
described in AM57x TRM section Virtual IO Timing Modes.
• Below is an example of how Virtual IO Timing Modes are defined in the AM57x Data
Manual:
Manual IO Timing Modes
• Manual IO Timing Modes are IO timing settings that must be calculated and
programmed by system software based on seed values in the datasheet.
• Application of Manual IO Timing Modes is done via the CFG_x_IN, CFG_x_OEN, and
CFG_x_OUT registers in the IODELAYCONFIG Module. This is described in the
AM57x TRM section Manual IO Timing Modes.
• Below is an example of how Manual IO Timing Modes seed values are defined in the
AM57x Data Manual:
IO Isolation Mode
• Any changes to the Pad Configuration Registers or IODELAYCONFIG registers
can potentially result in an undesirable state (i.e., output state changes or
output enable changes) on the associated IOs.
• To guarantee IO state, device pins should be placed in Isolation Mode when
making any changes to the Pad Configuration Registers or IODELAYCONFIG
Module Registers.
• Run-time (non-isolated) changes are only supported for MMC.
• See TRM section “Isolation Requirements” for details.
Customer Software Implications
• All I/O timing modes and pinmuxing shall be set by software …
– At boot-time
– While under protection of isolation
– While executing code from OCMCRAM (since DDR cannot be accessed while IOs
are in isolation)
– The Secondary Boot Loaders (just after ROM bootloader) execute from OCMC
RAM. The MLO in the Linux eco-system operates from the OCMC RAM.
– Therefore the MLO/ SBL is the ideal place for pad configurations.
• Run-time changes are needed for MMC where dynamic configurations can
not be avoided. TI has validated this use case.
COMING SOON
TPS659037 TPS659037 configured for the TPS65916 Power Generic TPS65916 configuration
Power Supply AM571x IDK Supply for AM571x
SMPS8 vdd_iva
SMPS9 3.3V
Thermal Considerations
• Thermal management ensures that every silicon device on the board works within its allowable
operating junction temperature. Failure to maintain a junction temperature within the range
specified reduces operating lifetime, reliability, and performance.
• The product design cycle should include thermal analysis to verify the operating junction
temperature of the device is within functional limits. If the temperature is too high, component-
or system-level thermal enhancements are required to dissipate the heat from the system.
AM572/1x Layout Compatibility
• AM572x & AM571x are designed as layout compatible solutions.
• Example differences:
– Same multiplexings for all common features
– Features/signals removed or added in AM571x
– New added pin muxings in AM571x
– IVA/DSP must run at same OPP when AM571x populated
(unless special population hooks)
– Some special powers swap, must treat them the same
(same filtering/handling)
For More Information
• For questions about this training, refer to the E2E Sitara Processors Forum:
https://e2e.ti.com/support/arm/sitara_arm