PowerPC 403GCX
32-Bit RISC
Embedded Controller
Features
•
PowerPC
™
RISC CPU and instruction set
architecture
•
Glueless interfaces to DRAM, SRAM,
ROM, and peripherals, including byte and
half-word devices
•
16KB instruction cache and 8KB write-
back data cache, two-way set-associative
•
Memory management unit
–64-entry, fully associative TLB array
–Variable page size (1KB-16MB)
–Flexible TLB management
•
Individually programmable on-chip
controllers for:
–Four DMA channels
–DRAM, SRAM, and ROM banks
–External interrupts
•
DRAM controller supports EDO DRAM
•
Flexible interface to external bus masters
•
CPU core can run at 2X the external bus
speed
Applications
•
Set-top boxes and network computers
•
Consumer electronics and video games
•
Telecommunications and networking
•
Office automation (printers, copiers, fax)
Specifications
•
CPU core frequencies of 50, 66, and 80
MHz, I/Os to 25, 33, and 40 MHz
•
Interfaces to both 3V and 5V technologies
•
Low-power 3.3V operation with built-in
power management and stand-by mode
•
Low-cost 160 lead PQFP package
•
Small footprint 160 PBGA package
•
0.45
µ
m triple-level-metal CMOS
Data
Sheet
Overview
The PowerPC 403GCX 32-bit RISC embedded
controller offers high performance and functional
integration with low power consumption. The
403GCX RISC CPU executes at sustained
speeds approaching one cycle per instruction.
On-chip caches and integrated DRAM and
SRAM control functions reduce chip count and
design complexity in systems, while improving
system throughput.
External I/O devices or SRAM/DRAM memory
banks can be directly attached to the 403GCX
bus interface unit (BIU). Interfaces for up to eight
memory banks and I/O devices, including a max-
imum of four DRAM banks, can be configured
individually, allowing the BIU to manage devices
or memory banks with differing control, timing, or
bus width requirements.
Interrupt
Controller
JTAG
Port
Serial
Port
4-Channel
DMA
Controller
(Address
and
Control)
Timers
RISC Execution Unit
Memory Management Unit
Instruction
Cache Unit
Data
Cache Unit
On-chip
Peripheral
Bus
Bus Interface Unit
DRAM Controller
I/O Controller
Data Address
Bus Bus
DRAM
Controls
SRAM, ROM, I/O
Controls
IBM PowerPC 403GCX
The 403GCX RISC controller consists of a pipe-
lined RISC processor core and several peripheral
interface units: BIU, DMA controller, asynchro-
nous interrupt controller, serial port, and JTAG
debug port.
The RISC processor core includes the internal
16KB instruction cache and 8KB data cache,
reducing overhead for data transfers to or from
external memory. The instruction queue logic
manages branch prediction, folding of branch
and condition register logical instructions, and
instruction prefetching to minimize pipeline
stalls.The integrated memory management unit
provides robust memory management and pro-
tection functions, optimized for embedded envi-
ronments.
RISC CPU
the address for the data read or write to the
BIU.When noncacheable operands are being
transferred, data can pass directly between the
EXU and the BIU, which interfaces to the exter-
nal memory being accessed.
Special Purpose Registers
Special purpose registers are used to control
debug facilities, timers, interrupts, the protection
mechanism, memory cacheability, and other
architected processor resources. SPRs are
accessed using move to/from special purpose
register (mtspr/mfspr) instructions, which move
operands between GPRs and SPRs.
Supervisory programs can write the appropriate
SPRs to configure the operating and interface
modes of the execution unit. The condition regis-
ter (CR) and machine state register (MSR) are
written by internal control logic with program exe-
cution status and machine state, respectively.
Status of external interrupts is maintained in the
external interrupt status register (EXISR). Fixed-
point arithmetic exception status is available from
the exception register (XER).
Device Control Registers
The RISC core comprises four tightly coupled
functional units: the execution unit (EXU), the
memory management unit (MMU), the data
cache unit (DCU), and the instruction cache unit
(ICU). Each cache unit consists of a data array,
tag array, and control logic for cache manage-
ment and addressing. The execution unit con-
sists of general purpose registers (GPR), special
purpose registers (SPR), ALU, multiplier, divider,
barrel shifter, and the control logic required to
manage data flow and instruction execution
within the EXU. The 403GCX core can operate at
either 1X or 2X the speed of the external buses,
which run at the SysClk input rate.
The EXU handles instruction decoding and exe-
cution, queue management, branch prediction,
and branch folding. The instruction cache unit
passes instructions to the queue in the EXU or, in
the event of a cache miss, requests a fetch from
external memory through the bus interface unit.
The MMU provides translation and memory pro-
tection for instruction and data accesses, using a
unified 64-entry, fully associative TLB array.
General Purpose Registers
Device control registers (DCR) are used to man-
age I/O interfaces, DMA channels, SRAM and
DRAM memory configurations and timing, and
status/address information regarding bus errors.
DCRs are accessed using move to/from device
control register (mtdcr/mfdcr) instructions, which
move operands between GPRs and DCRs.
Instruction Set
Table 1 summarizes the 403GCX instruction set
by categories of operations. Most instructions
execute in a single cycle, with the exceptions of
load/store multiple, load/store string, multiply,
and divide instructions.
Bus Interface Unit
The bus interface unit integrates the functional
controls for data transfers and address opera-
tions other than those which the DMA controller
handles. DMA transfers use the address logic in
the BIU to output the memory addresses being
accessed.
Control functions for direct-connect I/O devices
Data transfers to and from the EXU are handled
through the bank of 32 GPRs, each 32 bits wide.
Load and store instructions move data operands
between the GPRs and the data cache unit,
except in the cases of noncacheable data or
cache misses. In such cases the DCU passes
2
IBM PowerPC 403GCX
and for DRAM, SRAM, or ROM banks are pro-
vided by the BIU. Burst access for SRAM, ROM,
and page-mode DRAM devices is supported for
cache fill and flush operations.
The BIU controls the transfer of data between the
external bus and the instruction cache, the data
cache, or registers internal to the processor core.
The BIU also arbitrates among external bus mas-
ter and DMA transfers, the internal buses to the
cache units and the register banks, and the serial
port on the on-chip peripheral bus (OPB).
ble. For each SRAM/ROM bank, the bank size,
bank location, number of wait states, and timings
of chip selects, byte enables, and output enables
are all user-programmable.
Memory Management Unit
The memory management unit (MMU) supports
address translation and protection functions for
embedded applications. When used with appro-
priate system level software, the MMU provides
the following functions: translation of 4GB logical
address space into physical addresses, indepen-
dent enabling of instruction and data translation/
protection, page level cacheability and access
control via the translation mechanism, software
control of page replacement strategy, and addi-
tional control over protection via zones.
The fully associative 64-entry TLB array handles
both instruction and data accesses. The transla-
tion for any virtual address can be placed in any
one of the 64 entries, allowing maximum flexibil-
ity by TLB management software. Each TLB
entry contains a translation for a page that can
be any one of eight sizes from 1KB to 16MB,
incrementing by powers of 4.
The TLB can simultaneously contain any mix of
page sizes. This feature enables the use of small
pages when maximum granularity is required,
reducing the amount of wasted memory when
compared to the more common fixed 4KB page
Memory Addressing Regions
The 403GCX can address an effective range of
four gigabytes, mapped to 3.5GB (256MB for
SRAM/ROM or other I/O, 256MB DRAM, and
3GB OPB/reserved) of physical address space
containing twenty-eight 128MB regions. Cache-
ability with respect to the instruction or data
cache is programmed via the instruction and data
cache control registers, respectively.
Within the DRAM and SRAM/ROM regions, a
total of eight banks of devices are supported.
Each bank can be configured for 8-, 16-, or 32-bit
devices.
For individual DRAM banks, the number of wait
states, bank size,
RAS
-to-
CAS
timing, use of an
external address multiplexer (for external bus
masters), and refresh rate are user-programma-
Table 1. 403GCX Instructions by Category
Category
Data Movement
Arithmetic / Logical
Comparison
Branch
Condition
Rotate/Shift
Cache Control
Interrupt Control
Processor Management
load, store
add, subtract, negate, multiply, divide, and, or, xor, nand, nor, xnor, sign
extension, count leading zeros
compare, compare logical, compare immediate
branch, branch conditional
condition register logical
rotate, rotate and mask, shift left, shift right
invalidate, touch, zero, flush, store
write to external interrupt enable bit, move to/from machine state register,
return from interrupt, return from critical interrupt
system call, synchronize, move to/from device control registers, move to/
from special purpose registers
Base Instructions
3
IBM PowerPC 403GCX
size.
Instruction Cache Unit
The instruction cache unit (ICU) is a two-way set-
associative 16KB cache memory unit with
enhancements to support branch prediction and
folding. The ICU is organized as 512 sets of 2
lines, each line containing 16 bytes. A separate
bypass path is available to handle cache-inhib-
ited instructions and to improve performance dur-
ing line fill operations.
The cache can send two cached instructions per
cycle to the execution unit, allowing instructions
to be folded out of the queue without interrupting
normal instruction flow. When a branch instruc-
tion is folded and executed in parallel with
another instruction, the ICU provides two more
instructions to replace both of the instructions
just executed so that bandwidth is balanced
between the ICU and the execution unit.
at the requested fullword, continuing to the end of
the block and then wrapping around to fill the
remaining fullwords at the beginning of the block.
DMA Controller
The four-channel DMA controller manages block
data transfers in buffered, fly-by and memory-to-
memory transfer modes with options for burst-
mode operation. In fly-by and buffered modes,
the DMA controller supports transactions
between memory and peripheral devices.
Each DMA channel provides a control register, a
source address register, a destination address
register, a transfer count register, and a chained
count register. Peripheral set-up cycles, wait
cycles, and hold cycles can be programmed into
each DMA channel control register. Each chan-
nel supports chaining operations. The DMA sta-
tus register holds the status of all four channels.
Exception Handling
Table 2 summarizes the 403GCX exception prior-
ities, types, and classes. Exceptions are gener-
ated by interrupts from internal and external
peripherals, instructions, the internal timer facility,
debug events or error conditions. Six external
interrupt signals are provided on the 403GCX:
one critical and five general-purpose, all individu-
ally maskable.
All exceptions fall into three basic classes: asyn-
chronous imprecise exceptions, synchronous
precise exceptions, and asynchronous precise
exceptions. Asynchronous exceptions are
caused by events external to processor execu-
tion, while synchronous exceptions are caused
by instructions.
Except for a system reset or machine check, all
403GCX exceptions are handled precisely. Pre-
cise handling implies that the address of the
excepting instruction (synchronous exceptions
other than system call) or the address of the next
sequential instruction (asynchronous exceptions
and system call) is passed to the exception han-
dling routine. Precise handling also implies that
all instructions prior to the excepting instruction
have completed execution and have written back
their results.
Data Cache Unit
The data cache unit is provided to minimize the
access time of frequently used data items in main
store. The 8KB cache is organized as a two-way
set associative cache. There are 256 sets of 2
lines, each line containing 16 bytes of data. The
cache features byte-writeability to improve the
performance of byte and halfword store opera-
tions.
Cache operations are performed using a write-
back strategy. A write-back cache only updates
locations in main storage that corresponds to
changed locations in the cache. Data is flushed
from the cache to main storage whenever
changed data needs to be removed from the
cache to make room for other data.
The data cache may be disabled for a 128MB
memory region via control bits in the data cache
control register or on a per-page basis if the
MMU is enabled for data translation. A separate
bypass path is available to handle cache-inhib-
ited data operations and to improve performance
during line fill operations.
Cache flushing and filling are triggered by load,
store, and cache control instructions executed by
the processor. Cache blocks are loaded starting
4
IBM
PowerPC 403GCX
Asynchronous imprecise exceptions include sys-
tem resets and machine checks. Synchronous
precise exceptions include most debug excep-
tions, program exceptions, data storage viola-
tions, TLB misses, system calls, and alignment
error exceptions. Asynchronous precise excep-
tions include the critical interrupt exception,
external interrupts, and internal timer facility
exceptions and some debug events.
Only one exception is handled at a time. If multi-
ple exceptions occur simultaneously, they are
handled in priority order.
The 403GCX processes exceptions as reset, crit-
ical, or noncritical. Four exceptions are defined
as critical: machine check exceptions, debug
exceptions, exceptions caused by an active level
on the critical interrupt pin, and the first time-out
from the watchdog timer.
When a noncritical exception is taken, special
purpose register Save/Restore 0 (SRR0) is
loaded with the address of the excepting instruc-
tion (synchronous exceptions other than system
call) or the next sequential instruction to be pro-
cessed (asynchronous exceptions and system
call). If the 403GCX is executing a multicycle
instruction (load/store multiple, load/store string,
multiply or divide), the instruction is terminated
and its address stored in SRR0. Save/Restore
Register 1 (SRR1) is loaded with the contents of
the machine state register. The MSR is then
updated to reflect the new context of the
machine. The new MSR contents take effect
beginning with the first instruction of the excep-
tion handling routine.
At the end of the exception handling routine, exe-
cution of a return from interrupt (rfi) instruction
forces the contents of SRR0 and SRR1 to be
loaded into the program counter and the MSR,
respectively. Execution then begins at the
address in the program counter.
The four critical exceptions are processed in a
similar manner. When a critical exception is
taken, SRR2 and SRR3 hold the next sequential
address to be processed when returning from the
exception and the contents of the machine state
register, respectively. After the critical exception
handling routine, return from critical interrupt
(rfci) forces the contents of SRR2 and SRR3 to
be loaded into the program counter and the
MSR, respectively.
Timers
The 403GCX contains four timer functions: a
time base, a programmable interval timer (PIT), a
fixed interval timer (FIT), and a watchdog timer.
The time base is a 64-bit counter incremented at
the timer clock rate. The timer clock may be
driven by either an internal signal equal to the
processor clock rate or by a separate external
timer clock pin. No interrupts are generated when
the time base rolls over.
Table 2. 403GCX Exception Priorities, Types and Classes
Priority
1
2
3
4
5
6
7
8
9
10
Exception Type
System Reset
Machine Check
Debug
Critical Interrupt
WatchdogTimer Time-out
Exception Class
Asynchronous imprecise
Asynchronous imprecise
Synchronous precise
(except UDE and EXC)
Asynchronous precise
Asynchronous precise
Program Exception, Data Storage Exception,TLB Miss, and Synchronous precise
System Calls
Alignment Exceptions
External Interrupts
Fixed Interval Timer
Programmable Interval Timer
Synchronous precise
Asynchronous precise
Asynchronous precise
Asynchronous precise
5
