# 2026-11-06: rtl version 2, on the EMIO link and on the AXI link

Raw output as the board wrote it. Board-level, one board, one meter. The `*.txt` files at the top
are the predictions and pre-registrations, each committed before the measurement it covers.

| file | what |
|---|---|
| `PREDICTION.txt` | the class predictions for rtl version 2 |
| `first-attempt/` | 08:01 UTC on the EMIO build: the run/halt step and the known-answer vehicle; then the reload refused (the session check matched its own parent process, since fixed) |
| `emio-classes/` | 12:39 to 14:55 UTC on the EMIO build (sha256 `feea5015...`): step, null, settle, sampling, idle, the settle job at 50 MHz |
| `emio-checks/` | the same build afterwards: conformance, agreement, latency, throughput, the processor's port widths |
| `AXI-PREDICTION.txt` | committed at 13:06 UTC, during the class campaign and before its null had finished, and before the AXI build was metered |
| `NULL-REPEAT.txt` | committed at 15:03 UTC, after the class campaign ended and its null had failed, before the AXI build was metered |
| `axi-first/` | the first AXI build (sha256 `3d84dbd3...`, results read back by register reads): checks, then `power/` (the settle check and the repeated null) |
| `JOINT-PREREG.txt` | the joint-statistics test, registered before its first board run (`axi-first/joint.txt`) |
| `AXI-FINAL-PREDICTION.txt`, `axi-final/` | the final AXI build (sha256 `bb825055...`, results written back by DMA): checks, then `power/` (the settle check) |
| `build.txt` | the builds: tool and block versions, sizes, timing, each bitstream's sha256, and the testbenches' results |

Summarise a meter log with `wai-quantum-landscape-fpga analyze <file>` or
`python3 scripts/power_tables.py <files>`.