CUDA: fix MMQ stream-k rounding if ne00 % 128 != 0 (#8311) #14187
Job | Run time |
---|---|
12m 6s | |
13m 46s | |
49m 37s | |
1h 17m 33s | |
1h 22m 0s | |
1h 50m 0s | |
37m 48s | |
38m 47s | |
12m 16s | |
16m 32s | |
7h 30m 25s |
Job | Run time |
---|---|
12m 6s | |
13m 46s | |
49m 37s | |
1h 17m 33s | |
1h 22m 0s | |
1h 50m 0s | |
37m 48s | |
38m 47s | |
12m 16s | |
16m 32s | |
7h 30m 25s |