The MPU_CLUSTER consists of:
- Two CPUs (dual-core configuration), including:
- Arm version 7 ISA: Standard Arm instruction set plus Thumb®-2 ,
Jazelle® RCT, and Jazelle DBX
Java™ accelerator
- Neon SIMD coprocessor and VFPv4
- 12-stage in-order MPU core pipeline
- 128-bit-wide instruction fetch allows fetching up to four instructions/cycle
- 32 KiB/32 KiB instruction and data cache for each MPU core
- Complex Execution Unit (FPU) per MPU core
- 32-entry fully-associative micro-TLB each for instruction and data per MPU core
- 512-entry 4-way set-associative unified TLB per MPU core
- Unified L2 cache control including tags
- Interrupt controller (MPU_INTC)
- Supports 160 hardware interrupts. For more
information about interrupt mapping, see Interrupt
Controllers.
- One timer and watchdog timer per MPU core
- Internal APB bridge that connects to the APB port of each MPU core
The major interfaces of the MPU_CLUSTER are:
- Single AXI master supporting 128-bit interface
- System coherency supported through the AXI4 ACE interface
- An ATB port (for processor trace)
- An APB port
- Interrupt request lines
Note: The following hardware restrictions/limitations are applied to the Cortex-A15 MPCore in this device:
- Not implemented features:
- ACP port to support hardware coherency with external master
- ECC/parity support for the L1 or L2 caches
- Supported ACE configurations:
- AXI3 mode (default). In this configuration, barrier transactions are not issued on the AMBA4 interface.
- ACE non-coherent, no L3 mode. In this configuration, barrier transactions are issued on the AMBA4 interface.
For more information about AMBA4 interface configuration, see Section 4.3.8, MPU Subsystem AMBA Interface Configuration.
For more information about Cortex-A15 MPCore, see the Arm Cortex-A15 MPCore Technical Reference Manual (available at infocenter.arm.com/help/index.jsp).