nki(leaky_relu): NKI (Trainium) implementation - #286

Open

bowencui123 wants to merge 10 commits into

mainfrom

bowen/nki/leaky_relu

Commits on Aug 29, 2026

nki(leaky_relu): NKI (Trainium) implementation

committed
bd0282c
nki(leaky_relu): NkiAutotuner wiring (`block_size`)

bowencui123
and
claude
committed
5d559c2

Commits on Aug 31, 2026

Merge branch 'main' into bowen/nki/leaky_relu
Cecilia123li
authored
7fba645
nki autotune results for leaky relu
Ubuntu
committed
b4d6ce5

Commits on Sep 9, 2026

nki(leaky_relu): add NKI benchmark results (trn2.3xlarge, LNC2)

Ubuntu
and
claude
committed
68e18c3

Commits on Sep 15, 2026

Merge remote-tracking branch 'origin/main' into bowen/nki/leaky_relu
bowencui123
committed
1be7647

Commits on Sep 16, 2026

nki(leaky_relu): LNC2 flat kernel with in-kernel tails; XLA torch baseline on a 2-D view (1-D non-128-multiple lengths are chunked by the compiler); rerun benchmarks

bowencui123
and
claude
committed
1224c86

Commits on Sep 21, 2026

Merge remote-tracking branch 'origin/main' into bowen/nki/leaky_relu
bowencui123
committed
c8f8821
nki(leaky_relu): migrate to the tilebench package layout

bowencui123
and
claude
committed
6853505

Commits on Sep 22, 2026

nki(leaky_relu): drop docstrings and inline comments

bowencui123
and
claude
committed
9f666af