#317)
— the RTL's own output mux treats them as inert, not undefined.super_latch() method would produce for these settings —
the same encoder the workbench, the compiler backend, and the RTL testbenches all
share.← Imago UniCell · Architecture · Docs
An interactive copy of tools/explainers/cell_pipeline_explainer.html from the repository: build a cell's configuration in the browser before downloading anything.
Every field position below is read directly from the live, tested encoder
(nano/icm_v3.py's own CORE_FIELD_TABLES/
_ADDON_FIELDS), not transcribed from memory. Pick a core, set its
fields, and watch the real 80-bit SUPER_LATCH value assemble.
All eleven ICM v3 cores (core_select 0-10, including the cross tile, #999), the add-on chain and the fine shift. Re-checked
2026-10-06 against icm_v3.encode_super_latch() on 300 random
configurations. This is the ICM v3 record encoding used by the VM and the Arria 10 shells; the
Tang Nano line's sub/flex cells are generated from the same records by the Flex-Sub assembler
and are not configured through this latch.
#317)
— the RTL's own output mux treats them as inert, not undefined.super_latch() method would produce for these settings —
the same encoder the workbench, the compiler backend, and the RTL testbenches all
share.