7
PC7457/47 [Preliminary]
5345B–HIREL–02/04
–
–
–
–
–
–
MESI data cache coherency maintained in hardware
Separate copy of data cache tags for efficient snooping
Parity support on cache and tags
No snooping of instruction cache except for icbi instruction
Data cache supports AltiVec LRU and transient instructions
Critical double- and/or quad-word forwarding is performed as needed.
Critical quad-word forwarding is used for AltiVec loads and instruction
fetches. Other accesses use critical double-word forwarding
Level 2 (L2) cache interface
–
On-chip, 512 Kbyte, eight-way set-associative unified instruction and data
cache
–
Fully pipelined to provide 32 bytes per clock cycle to the L1 caches
–
A total nine-cycle load latency for an L1 data cache miss that hits in L2
–
PLRU replacement algorithm
–
Cache write-back or write-through operation programmable on a per-page or
per-block basis
–
64-byte, two-sectored line size
–
Parity support on cache
Level 3 (L3) cache interface (not implemented on PC7447)
–
Provides critical double-word forwarding to the requesting unit
–
Internal L3 cache controller and tags
–
External data SRAMs
–
Support for 1, 2, and 4M bytes (MB) total SRAM space
–
Support for 1 or 2 MB of cache space
–
Cache write-back or write-through operation programmable on a per-page or
per-block basis
–
64-byte (1 MB) or 128-byte (2 MB) sectored line size
–
Private memory capability for half (1 MB minimum) or all of the L3 SRAM
space for a total of 1-, 2-, or 4-MB of private memory
–
Supports MSUG2 dual data rate (DDR) synchronous Burst SRAMs, PB2
pipelined synchronous Burst SRAMs, and pipelined (register-register) Late
Write synchronous Burst SRAMs
–
Supports parity on cache and tags
–
Configurable core-to-L3 frequency divisors
–
64-bit external L3 data bus sustains 64-bit per L3 clock cycle
Separate memory management units (MMUs) for Instructions and data
–
52-bit virtual address; 32- or 36-bit physical address
–
Address translation for 4 Kbyte pages, variable-sized blocks, and
256M bytes segments
–
Memory
programmable
as
inhibited/caching-allowed, and memory coherency enforced/memory
coherency not enforced on a page or block basis
–
Separate IBATs and DBATs (eight each) also defined as SPRs
write-back/write-through,
caching-