From e398de4cf68ffd0f43e0b1692b08b2726457fad4 Mon Sep 17 00:00:00 2001 From: Arthur Hanson Rasmusson Date: Sun, 18 Oct 2026 15:03:17 -0400 Subject: cuda-ab: ctaid_probe (CTA-coordinate + A-row reader) + GEO geometry print in transformer_mt The vGPU GR engine's failing chains are correct for the by=0 band and replicate one tile for rows >= 16 with the QMD geometry and the input recs proven clean, so the remaining hypothesis is engine-side CTA-coordinate behaviour. ctaid_probe reads it directly in two phases on the exact (16,4,1)/(16,16,1) geometry: Phase 1 coord_map writes by*10000+bx*100+ty*10+tx per element (int32 64x256) and reports per-band row0 plus a bad count - a collapsed or wrong CTAID.Y shows instantly. Phase 2 arow_read writes A[(by*16+ty)*256+(bx*16+tx)] with A[i]=i, so a correct engine reads each band's own rows; a collapsed row register reads one band's rows everywhere. transformer_mt prints its per-chain geometry (GEO lines) for the host-side GRID-WIT comparison. Runtime <10 s. Built with nvcc -O2 --cudart=static -arch=sm_75 (deterministic static runtime). --- tools/cuda-ab/run/ctaid_probe | Bin 0 -> 829280 bytes tools/cuda-ab/run/transformer_mt | Bin 845968 -> 846064 bytes 2 files changed, 0 insertions(+), 0 deletions(-) create mode 100755 tools/cuda-ab/run/ctaid_probe (limited to 'tools/cuda-ab/run') diff --git a/tools/cuda-ab/run/ctaid_probe b/tools/cuda-ab/run/ctaid_probe new file mode 100755 index 0000000..a675ed1 Binary files /dev/null and b/tools/cuda-ab/run/ctaid_probe differ diff --git a/tools/cuda-ab/run/transformer_mt b/tools/cuda-ab/run/transformer_mt index 32d159d..da03273 100755 Binary files a/tools/cuda-ab/run/transformer_mt and b/tools/cuda-ab/run/transformer_mt differ -- cgit v1.2.3