Dependent pointer chasing¶

Generated workload catalogue · Execution contract

Purpose¶

Force a subsequent memory address to depend on a prior load.

Setup¶

Property	Value
Workload ID	`pointer-chase`
Selection	`--workload pointer-chase`
Generator	`buildRV32IBenchmark() in RV32IBenchmarks.cpp`
End condition	EBREAK after the loop and result store

n nodes at 0x1000; each is {next address, 32-bit payload}. A seeded shuffle links all nodes into one ring. Traversal starts at node zero.

Use the shared address map and parameter bounds. N includes any training iterations; n means size.

Implementation¶

Load the next-node address into the address register, then load that node's payload. Make n steps per outer iteration.

for (size_t i = 0; i < n; ++i) { node = node->next; sum += node->payload; }

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

Expected result¶

Each complete ring traversal visits every payload once. Result=N*sum(payloads) modulo 2^32. The one-node case points back to itself.

The runner checks the generator's expected result in addition to retirement-by-retirement oracle validation. Output-array assertions, where described, check more than a checksum.

Measurement and running¶

Compare with stream while acknowledging different instruction sequences. On current FF memory there is no modeled DRAM/cache locality penalty; the dependent-address chain is still real.

build/rv32i_experiment --core advanced --workload pointer-chase \
  --iterations 2 --size 16 --seed 1  \
  --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/RV32IBenchmarks.cpp: emitted instructions and independent expected results.
tests/cpp/RV32IBenchmarkTest.cpp: generator results, parameter bounds, and edge cases.
tests/checks/RV32IBenchmarkCLITest.py and tests/checks/BenchmarkCampaignCLITest.py: shared execution/campaign checks.