This repository collects benchmarks and flame graphs for upstreaming Crapto1 optimizations.
- Standalone mfkey32v2 -O0 (control)
- Standalone mfkey32v2 -O3 (control)
- All identified optimizations (variable)
- No optimizations involving a GPU or over one terabyte of storage
- 20 test nonces, mix of unique keys (12) 1 and consecutive real keys (8) 2
- Observed time is the median time across 50 runs on an unloaded system
- System specifications:
- Ubuntu 23.04
- Dual core (dedicated) Intel(R) Xeon(R) Platinum 8280 CPU @ 2.70GHz
- 4 GB RAM, ~3 GB free
- 23 GB free disk
Times will be published when iteration2.c is complete
| Name | Description | Time (s) |
|---|---|---|
mfkey32v2 -O0 |
Control | N/A |
mfkey32v2 -O3 |
Control | N/A |
| iteration1.c | Base speedup using compiler flags, minimized implementation, loading all of the nonces at once | N/A |
| iteration2.c | Eliminating statelist, return early from recovery function | N/A |