Use mul_mat_transpose at axpy op for large batch
#51
| Job | Run time |
|---|---|
| 1m 46s | |
| 21s | |
| 1m 30s | |
| 1m 18s | |
| 4m 4s | |
| 3m 29s | |
| 3m 10s | |
| 7m 53s | |
| 5m 1s | |
| 3m 17s | |
| 1m 7s | |
| 3m 23s | |
| 1m 12s | |
| 27s | |
| 40s | |
| 1m 27s | |
| 1m 6s | |
| 1m 34s | |
| 45s | |
| 1m 49s | |
| 1m 28s | |
| 1m 31s | |
| 13m 23s | |
| 13m 21s | |
| 52s | |
| 0s | |
| 1h 15m 54s |