|
MV78100 Datasheet, PDF (3/124 Pages) – | |||
|
◁ |
MV78100
Discovery⢠Innovation Series CPU Family
Hardware Specifications
PRODUCT OVERVIEW
Building upon the Marvell® high-performance Sheevaâ¢
CPU core, the MV78100 is part of the Discoveryâ¢
Innovation series CPU family.
The MV78100 is optimally designed for a broad range of
applications ranging from sophisticated routers,
switches, and wireless-base stations to high-volume
storage and laser printer applications.
The MV78100 incorporates a fully ARMv5TE-compliant
dual-issue CPU core with a double-precision, IEEE
compliant Floating-point Unit (FPU), and 512 KB of
L2 cache.
Its innovative crossbar architecture, advanced
communications peripherals, and performance-tuned
interfaces, make it a perfect, high-performance solution
for embedded applications such as:
⢠Printers
⢠Core and edge routers
⢠Cellular base stations
⢠Ethernet switch management
⢠Storage arrays
⢠Network Attached Storage (NAS) devices
 The MV78100 Includes
⢠High-performance Sheeva⢠dual-issue CPU with
IEEE compliant FPU support
⢠512 KB L2 cache
⢠High bandwidth DDR II memory interface (32/64-bit
DDR2â800 MHz data rate with an 8-bit ECC
option)
⢠8/16/32-bit device bus with up to five chip selects,
and with NAND and NOR Flash support
⢠Two x4 wide PCI Express ports with integrated
PHY; each one can also act as four x1 ports
⢠Two Gigabit Ethernet MAC controllers
⢠Three USB 2.0 ports with integrated PHYs
⢠Two SATA 2.0 ports with integrated 3 Gbps SATA II
PHY
⢠Security Cryptographic engine
⢠Four 16550 compatible UARTs
⢠Two channels SLIC/Codec TDM interface
⢠Four IDMA engines
⢠Integrated Storage Accelerator engine (two XOR
DMA or iSCSI CRC engines)
⢠Timers
⢠Interrupt controller
 Sheeva⢠Dual-Issue CPU with FPU support
⢠Up to 1.2 GHz
⢠Super-scalar, dual-issue CPU
⢠Single-precision and double-precision FPU support
⢠32-bit and 16-bit RISC architecture
⢠Compliant with v5TE architecture, published in the
ARM Architect Reference Manual, Second Edition
⢠Supports 32-bit instruction set for performance and
flexibility
⢠Supports 16-bit Thumb instruction set for code
density
⢠Supports DSP instructions to boost performance
for signal processing applications
⢠Includes MMU to support virtual memory features
⢠MPU can be used instead of MMU
⢠32-KB I-Cache and 32-KB D-Cache, parity
protected
⢠512-KB unified L2 cache, ECC protected
⢠64-bit internal data bus
⢠Variable pipeline stagesâsix to nine stages
⢠Out-of-order execution for increased performance
⢠In-order retire via Reordering Buffer (ROB)
⢠Branch Prediction Unit
⢠Supports JTAG/ARM-compatible ICE
⢠Supports both Big and Little Endian modes
 DDR2 SDRAM controller
⢠40/72-bit interface (32/64-bit data + 8-bit ECC)
⢠Up to 400 MHz clock frequency (DDR2â800 MHz
data rate)
⢠DDR SDRAM with a clock ratio of 1:N and 2:N
(up to 1:5) between the DDR SDRAM and the
Sheeva⢠core, respectively
⢠SSTL 1.8V I/Os
⢠Auto-calibration of I/Os output impedance
⢠Supports four DRAM banks
⢠Supports all DDR devices, densities up to 2 Gb
⢠Up to 4 GB address space
Copyright © 2008 Marvell
December 6, 2008, Preliminary
Document Classification: Proprietary Information
MV-S104552-U0 Rev. D
Page 3
|
▷ |