Add --target-cuda argument for selecting CUDA architecture
#7368
| Job | Run time |
|---|---|
| 23m 58s | |
| 16m 4s | |
| 22m 9s | |
| 16m 2s | |
| 27m 16s | |
| 15m 49s | |
| 15m 39s | |
| 23m 5s | |
| 1h 13m 46s | |
| 13m 19s | |
| 4m 6s | |
| 13m 12s | |
| 13m 37s | |
| 12m 38s | |
| 25m 32s | |
| 12m 36s | |
| 12m 53s | |
| 0s | |
| -1s | |
| 5h 41m 40s |