Embench: CRC32¶
Compiled workload catalogue · Execution contract
Purpose¶
Compute CRCs over a deterministic pseudo-random input stream.
Setup¶
| Property | Value |
|---|---|
| Source workload | crc32 |
| Pinned source | src/crc32/crc_32.c |
| Revision | Embench 1.0, commit 0466a18e4f6b47e19598d7c6ba72916d54b68f65 |
| Build selection | bash tools/rv32i-benchmarks/build-embench.sh crc32 |
| Resulting image | build/rv32i-embench/artifacts/crc32/crc32.elf |
| Completion | Upstream verifier, main return, then committed tohost verdict |
These are the upstream fixed inputs and repetition rules, not the native generator's size/iterations options. A CRC32 artifact exists in the inspected checkout; historical execution evidence is not a current-core rerun. See port limitations.
Implementation¶
Each body repetition reseeds the upstream pseudo-random generator with zero and runs the original CRC routine. The upstream local repetition factor is 170 at CPU_MHZ=1.
Expected result¶
verify_benchmark requires the returned value to equal 11433.
The shared upstream main converts verifier failure into a nonzero return. The CircuitSim startup bridge publishes tohost=1 on success or 3 on failure. A checked instruction prefix cannot establish this final result.
Measurement and running¶
Bit manipulation, loop control, and software arithmetic. This is not the shorter project-generated --workload crc32 kernel.
Build this source explicitly, inspect its ELF size/instruction requirements, then follow the bounded prefix and campaign procedure. Full-run timing includes startup, warmup, benchmark body, verifier, and verdict unless a separate valid measurement window is selected.
Source and tests¶
The pinned upstream source contains the benchmark body and verifier. Other files in that source directory may provide helper implementations. Local port files are under tools/rv32i-benchmarks/embench/.
The ordinary repository regression suite tests the ELF loader/runner using small fixtures; it does not download and run this workload automatically. Keep its external verifier result, runner/oracle result, and any structural/behavioral comparison as separate recorded checks.