Large batches (e.g. ViT-S/16 LNC=1 batch 32, ViT-B/L/H+ LNC=1 batch 16) failed to compile with NCC_IBIR229.
Shapes that compiled before keep the same schedule: bit-identical output and the same speed (within 0.12%).
Bitwise-identical to v1 at LNC=1 and LNC=2. Packed fc1 bias loaded in one DMA on the
Scalar HW-DGE queue; row-block-0 x transposed on the Tensor engine before the weights;
res_layernorm uses dge_mode.none with all token tiles SBUF-resident. Based on an external
contribution by jlongen-aws (PR #2 on jburtoft/vjepa2-neuron-kernels), re-implemented with
the LNC-parametric path preserved. Verified on trn2.3xlarge: real-weight V-JEPA2 output
bitwise identical to v1, -4.0% ms/clip.
26 days ago
v1: NKI kernels for RoPE ViT encoders (DINOv3, V-JEPA2) on Trainium2
