# PA-RISC Architecture ## Datapath SLPARISCCPU -> instruction fetch -> decoder -> register file -> integer/shift units -> branch/control -> load/store -> FPU/MAX -> MMU/TLB -> cache/interconnect -> trap/retirement ## PA-RISC 1.x Early implementations can be represented as scalar or limited superscalar pipelines. PA-7100 and later 1.1 processors introduced superscalar execution. citeturn0search1 ## PA-RISC 2.0 PA-8000 introduced a substantially wider out-of-order design with ten functional units and an Instruction Reorder Buffer. citeturn0search3 The generic model therefore exposes: - issue/dispatch; - dependency tracking; - execution resources; - load/store scheduling; - reorder/retirement state. Concrete widths remain processor-specific. ## Traps and interrupts Precise interrupt/trap state is represented explicitly. Fast interrupt/shadow state belongs to the selected architectural profile. ## Load/store Only memory operations access ordinary memory; arithmetic and logical operations operate on registers. citeturn0search1