Dependent integer recurrence¶

Generated workload catalogue · Execution contract

Purpose¶

A compact legacy forwarding/dependency loop; unlike dependency, it has no configurable interleaving distance.

Setup¶

Property	Value
Workload ID	`dependent`
Selection	`--workload dependent`
Generator	`legacyProgram() in RV32IWorkloads.cpp`
End condition	EBREAK after the loop

x1 = iterations, x2 = 0, x3 = 0. No input data or ABI stack.

N means the requested iteration count. Registers start from reset; the loop's x2 is not an ABI stack pointer.

Implementation¶

Each outer iteration increments x2, doubles that value into x3, then adds x3 back into x2. Decrement x1 and branch until zero.

x2 = 3 * (x2 + 1);  // uint32_t wraparound
x1 -= 1;

This is a behavioral summary of the generated assembly, not the host program being timed.

Expected result¶

After N iterations, x2 follows v[0]=0, v[k+1]=3*(v[k]+1) modulo 2^32. x3 is twice the last incremented value; x1=0. For N=2: x2=12, x3=8.

The runner validates the executed architectural state with the instruction oracle. This legacy generator does not publish a separate expected_result checksum; the equations above describe its loop.

Measurement and running¶

Observe dependence stalls/forwarding and loop overhead. This is the older recurrence fixture, not the 64-operation distance sweep.

build/rv32i_experiment --core advanced --workload dependent \
  --iterations 2  \
  --fidelity behavioral --period 4000 \
  --allocation-limit-mib 512 --max-cycles 4096

The explicit period is for a bounded functional check, not a hardware-speed result. For comparisons, use the serial campaign driver, identical case inputs, and a certified per-target clock.

Source and tests¶

src/analysis/RV32IWorkloads.cpp: emitted instructions.
tests/checks/RV32IExperimentCLITest.py: legacy loops across all three cores.
tests/checks/RV32IBenchmarkCLITest.py and tests/checks/BenchmarkCampaignCLITest.py: shared execution/campaign checks.