Andrew Kerr
|
877bdcace6
|
Cutlass 1.3 Release (#42)
CUTLASS 1.3 Release
- Efficient GEMM kernel targeting Volta Tensor Cores via mma.sync instruction added in CUDA 10.1.
|
2019-03-20 10:49:17 -07:00 |
|
akerr
|
822b0952cd
|
Resolved issue for incorrect SGEMM on Maxwell architecture.
|
2018-12-19 15:07:16 -08:00 |
|
akerr
|
74df0331f2
|
CUTLASS 1.2
|
2018-10-26 14:38:46 -07:00 |
|
akerr
|
cfe4b933ef
|
CUDA 9 lacks host-side conversions from float=>half. Instead, we must reinterpret_cast<> from cutlass::half_t => half.
|
2018-09-29 15:04:20 -07:00 |
|
akerr
|
461f417b9d
|
Checkpointing CUTLASS 1.1 release.
|
2018-09-18 16:58:03 -07:00 |
|