Timeline for Does memoization skew benchmarks?
Current License: CC BY-SA 4.0
3 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 5, 2024 at 21:40 | comment | added | Peter Cordes | The question is whether the better (micro-)benchmark results are realistic. If the whole-application real-world performance is obviously better then yes, good, it's a real speedup, not only in a benchmark with repetitive inputs and no other cache-footprint pressure. But there's still a question of how to micro-benchmark if you want to tune even further than your first attempt at memoization. | |
| Dec 5, 2024 at 13:33 | comment | added | Greg Burghardt | This was my thought too. If you aren't expecting the cache to get hit, why memoize anything? It's faster with memoization? Mission accomplished. But still, is the real environment going to run under similar situations? | |
| Dec 5, 2024 at 3:40 | history | answered | gnasher729 | CC BY-SA 4.0 |