[NE16] Add GAP9_w_NE16 platform (NE16 accelerator on GAP9)#1

Open

runwangdl wants to merge 24 commits into

develfrom

Commits on Apr 14, 2026

Commits on Apr 18, 2026

[NE16] add CNN_Libraries_HWPE include path for GAP9 SDK NE16 kernels
runwangdl
committed

Commits on Apr 19, 2026

Commits on May 14, 2026

[NE16] enable MobileNetV1 (MLPerf VisualWakeWords) as L2 single-buffer model test
runwangdl
committed
[NE16] engine-aware DW NHWC layout: fix mixed-engine graphs with --enable-3x3
runwangdl
committed
[NE16] enable --enableStrides for model tests: stride-2 convs on NE16
runwangdl
committed
[NE16] enable L2 double-buffer model tests + fix Dense WeightCube rank
runwangdl
committed
[CI] add GAP9_w_NE16 model jobs (singlebuffer + doublebuffer L2)
runwangdl
committed
[CI] gap9-w-ne16-tiled: mark workspace safe (match other runner workflows)
runwangdl
committed
[GAP9] gate NE16AdjustGEMMWeightLayoutPass on engine == "NE16"
runwangdl
committed
[GAP9] re-enable PULPAddRequantMergePass; harden RQSSplitPass int32 cast
runwangdl
committed
[GAP9] unmask FP32 GEMM and SkipConnection paths
runwangdl
committed
[GAP9] lint: drop unused MatMulAddMergePass import + yapf format
runwangdl
committed