@Mark-Simulacrum I want to ensure I'm reading the report correctly, as different measures give conflicting results. #2305
Description
@Mark-Simulacrum I want to ensure I'm reading the report correctly, as different measures give conflicting results.
| measure | before (total) | after (total) | change |
|---|---|---|---|
| cpu-clock | 372185.29 | 369019.14 | -0.85% |
| cycles:u | 1421498864061 | 1409937857765 | -0.81% |
| faults | 1511256 | 1510959 | -0.02% |
| instructions:u | 1699836885596 | 1723213735874 | +1.38% |
| max-rss | 16735048 | 16739924 | +0.03% |
| task-clock | 370849.03 | 367449.54 | -0.92% |
| wall-time | 202.36 | 203.21 | +0.42% |
If we judge by "instructions", there is a significant regression, but if we judge by "cpu-clock" or "cycles", there is a signficant improvement.
Originally posted by @kennytm in rust-lang/rust#45282 (comment)