12345device Apple M5; IQ2_S/Q2_0, T 256, 64 experts, tile TM 16 TN 32 TK 32 SG 4right-input layout: runs of 4 kplan: 190 tiles of 16, 3040 padded rows for 2560 entriesparity: plan ok; y vs fp16-operand reference rel L2 1.60e-05, vs exact float reference 4.38e-04 PASS