Next: Performance
Up: Motivation: an example
Previous: Architectural details
- L1 Instruction cache: 16 KB, 4-Way, 32 Byte/Line, LRU
- L1 Data cache: 16 KB, 4-way, 32 byte/line,
non-blocking, dual-ported, write allocate, LRU
- L2 unified cache:
256 KB, 8-Way, 32 Byte/Line, non-blocking
- Note that the matrix occupies
2MB.
- Each row of the matrix occupies
4KB.