The README and tests/chips.rs report node, transistor, name and polygon counts
for all three chips. Only the 6502's are checked for value — those figures
are known independently in tinymachines/6502 and pinned in a test, which is
what stops a parser change silently altering every chip at once.
The 6800 and Z80 numbers are checked for shape only: more than 1000
transistors, more than 500 nodes, rails distinct and actually rails, and every
polygon belonging to a node the netlist agrees exists.
|
nodes |
transistors |
names |
polygons |
| 6800 |
2944 |
3995 |
1144 |
9805 |
| Z80 |
3597 |
6813 |
511 |
14604 |
They should not be quoted as authoritative, and the README says so.
What would fix it
Any independent count — the upstream visual6502 project's own reporting for
those dies, a published die analysis, or a second implementation reading the
same files. Then pin them the way the 6502's are pinned.
Until then this is a known limit, not a defect.
The README and
tests/chips.rsreport node, transistor, name and polygon countsfor all three chips. Only the 6502's are checked for value — those figures
are known independently in
tinymachines/6502and pinned in a test, which iswhat stops a parser change silently altering every chip at once.
The 6800 and Z80 numbers are checked for shape only: more than 1000
transistors, more than 500 nodes, rails distinct and actually rails, and every
polygon belonging to a node the netlist agrees exists.
They should not be quoted as authoritative, and the README says so.
What would fix it
Any independent count — the upstream visual6502 project's own reporting for
those dies, a published die analysis, or a second implementation reading the
same files. Then pin them the way the 6502's are pinned.
Until then this is a known limit, not a defect.