2024-11-14 10:31:32 +01:00
|
|
|
## Benchmarks
|
|
|
|
|
|
|
|
|
|
In here we show the preliminary benchmarks of codex storage proofs circuits.
|
|
|
|
|
|
|
|
|
|
## Running Benchmarks
|
|
|
|
|
|
2025-03-10 14:49:13 +01:00
|
|
|
**Runtime Benchmarks**
|
2024-11-14 10:31:32 +01:00
|
|
|
|
2025-03-10 14:49:13 +01:00
|
|
|
To run the benchmarks for sampling circuit and aggregation, you can use the following commands:
|
2024-11-14 10:31:32 +01:00
|
|
|
|
|
|
|
|
```bash
|
2025-03-10 14:49:13 +01:00
|
|
|
cd bench_scripts
|
|
|
|
|
bash ./bench_runtime
|
2024-11-14 10:31:32 +01:00
|
|
|
```
|
|
|
|
|
|
2025-03-10 14:49:13 +01:00
|
|
|
The following operations were benchmarked for sampling and recursion circuits (both single-thread and multi-thread):
|
2024-11-14 10:31:32 +01:00
|
|
|
|
|
|
|
|
- **Build Circuit**: Time taken to construct the circuit for the specified params.
|
|
|
|
|
- **Prove Circuit**: Time taken to generate a proof for the constructed circuit.
|
|
|
|
|
- **Verify Circuit**: Time taken to verify the generated proof.
|
|
|
|
|
|
2025-03-10 14:49:13 +01:00
|
|
|
**Memory Benchmarks**
|
|
|
|
|
To run the memory benchmarks for sampling circuit and aggregation, you can use the following commands:
|
2024-11-14 10:31:32 +01:00
|
|
|
|
2024-11-14 13:23:34 +01:00
|
|
|
```bash
|
2025-03-10 14:49:13 +01:00
|
|
|
cd bench_scripts
|
|
|
|
|
bash ./bench_memory
|
2024-11-14 13:23:34 +01:00
|
|
|
```
|
2025-03-10 14:49:13 +01:00
|
|
|
Note: The memory usage is quite difficult to replicate as it depends on the operating system,
|
|
|
|
|
but on macOS look for the "maximum resident set size" to get the peak memory usage.
|
2024-11-14 10:31:32 +01:00
|
|
|
|
2025-03-10 14:49:13 +01:00
|
|
|
## Bench Results
|
|
|
|
|
See this [document](https://hackmd.io/@NQdG6IOmQE6astjwhJ6ACw/Bk-SopNj1g) for full benchmark results.
|