perf: improve default/fallback backend implementation for blockwise quantization ops #291
| Job | Run time |
|---|---|
| 17s | |
| 12s | |
| 10s | |
| 11s | |
| 3m 33s | |
| 12s | |
| 4m 8s | |
| 17s | |
| 16s | |
| 21s | |
| 4m 42s | |
| 3m 59s | |
| 5m 16s | |
| 3m 43s | |
| 3m 55s | |
| 3m 34s | |
| 4m 21s | |
| 4m 51s | |
| 19m 35s | |
| 24m 32s | |
| 23m 19s | |
| 15m 29s | |
| 15m 10s | |
| 22m 22s | |
| 18m 9s | |
| 14m 20s | |
| 9m 22s | |
| 7m 2s | |
| 9m 48s | |
| 9m 3s | |
| 8m 31s | |
| 46m 54s | |
| 8m 42s | |
| 10m 7s | |
| 8m 56s | |
| 7m 22s | |
| 5h 22m 41s |