# ARM Cache ARM architecture supports implementation-specific cache hierarchies rather than one universal cache size or topology. SLeeLa exposes: - L1 instruction cache; - L1 data cache; - optional unified cache; - L2; - L3; - TLBs; - translation-table walk caches; - coherency/interconnect coupling. Arm's AMBA AXI documentation separates processor/cache masters from memory/interconnect infrastructure, while Cortex-A documentation describes concrete cache and memory-system implementations. citeturn0search22turn0search23 No cache size, associativity, line size, replacement policy, or coherency protocol is invented for generic ARM.