[Codegen] Use CUDA's half2 and nv_bfloat162 intrinsics for vector fp16/bf16 data types - #15190

Closed

yzh119 wants to merge 38 commits into

apache:mainfrom

yzh119:vec-fp16-bf16

Commits on Jun 3, 2023

upd
yzh119
committed
a941951
wip
yzh119
committed
3284151

Commits on Jun 10, 2023

upd
yzh119
committed
e753640

Commits on Jun 29, 2023

upd
yzh119
committed
7f31636
upd
yzh119
committed
0f6b858
rm redundant files
yzh119
committed
edc938a
upd
yzh119
committed
4ac20f1
Merge branch 'main' into native-bf16
yzh119
authored
a971ee0
upd
yzh119
committed
de6c5b2
do not change nvcc
yzh119
committed
9483524
fix
yzh119
committed
fd0ff79
remove empty line
yzh119
committed
a0525b9
fix
yzh119
committed
9c6d639
lint
yzh119
committed
bdc3382

Commits on Jun 30, 2023

c++ lint
yzh119
committed
a15b600
use ml_dtypes for llvm codegen test
yzh119
committed
c33423a

Commits on Jul 1, 2023

add ml_dtypes to ci-constraints
yzh119
committed
6d3b6c3
alphabetical
yzh119
committed
40f088c
pylint
yzh119
committed
259f278
lint
yzh119
committed
aff8f61
upd
yzh119
committed
d883e33
improve comments
yzh119
committed
487c15f
improved code comment
yzh119
committed
73d6361
upd
yzh119
committed
0d7d8ba
bugfix
yzh119
committed
4f4e8a6

Commits on Jul 2, 2023

bugfix
yzh119
committed
2162eba
lint
yzh119
committed
31355fa

Commits on Jul 3, 2023

refactor buildprocess
yzh119
committed
daaae71
remove unused functions
yzh119
committed
af72e1e
pylint
yzh119
committed
98cb6e4
import error
yzh119
committed
aae8208
upd
yzh119
committed
0a4a67a
fix
yzh119
committed
61ed5cc
wip
yzh119
committed
1cbd2f5
wip
yzh119
committed
617496a
upd
yzh119
committed
e962032
complete unary
yzh119
committed
c2bd9ee
bugfix
yzh119
committed
3887ff6