Skip to content

docs(hardware): record TTSKY26b withdrawal in silicon_anchor §3 #52

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
dmitrii-f-t27 wants to merge 1 commit into
base: main
Choose a base branch
from
Open

docs(hardware): record TTSKY26b withdrawal in silicon_anchor §3 #52

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion docs/hardware/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,7 +25,7 @@ root [`README.md`](../../README.md) §"Why $TRI Is Mined Only on TTSKY26b" and
|-------|--------|-------------------|
| GF16 format specified | `verified` | `gf16_spec.md` §2.1–2.4 |
| GF16 phi-distance < f16 | `verified` | `gf16_spec.md` §2.5 |
| FPGA 323 MHz, 35/35 RTL tests | `verified` | `gf16_benchmarks.json` |
| FPGA 323 MHz, 35/35 RTL tests | `verified` | **EVIDENCE MISSING** — `gf16_benchmarks.json` does not exist in this repo. The only benchmark file present, `gf16_bench_results.json`, is a *software* benchmark (Zig `src/bench_*.zig`: MSE, ns/op latency) and contains no FPGA synthesis frequency and no RTL test count. |
| BPB (bits-per-byte) advantage | `empirical_fit` | `bpb_benchmark.py` — synthetic data only |
| φ-structured step (0.694-bit) | `verified` | `gf16_mathematics.md` §3 |
| Optimal field size GF(2⁴) | `verified` | `gf16_mathematics.md` §2 |
Expand Down
27 changes: 19 additions & 8 deletions docs/hardware/silicon_anchor.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -141,16 +141,25 @@ GF16 draft. It carries the basic `gf16_mul` / `gf16_add` / `gf16_dot4` core.

| Crown | Repo | Module | PRs in sky-26b | Status |
|-------|------|--------|----------------|--------|
| **Phi** (Nano) | `gHashTag/tt-trinity-phi` | `tt_um_trinity_nano` | 9 closed | **Submitted** |
| **Euler** (Compact) | `gHashTag/tt-trinity-euler` | `tt_um_ghtag_trinity_gf16` | 7 closed | **Submitted** |
| **Gamma** (Full) | `gHashTag/tt-trinity-gamma` | `tt_um_trinity_max_true` | 5 closed | **Submitted** |
| **Phi** (Nano) | `gHashTag/tt-trinity-phi` | `tt_um_trinity_nano` | 9 closed | Submitted 2026-05-17 — **WITHDRAWN** |
| **Euler** (Compact) | `gHashTag/tt-trinity-euler` | `tt_um_ghtag_trinity_gf16` | 7 closed | Submitted 2026-05-17 — **WITHDRAWN** |
| **Gamma** (Full) | `gHashTag/tt-trinity-gamma` | `tt_um_trinity_max_true` | 5 closed | Submitted 2026-05-17 — **WITHDRAWN** |

**Total: 21 closed PRs** in [TinyTapeout/tinytapeout-sky-26b](https://github.com/TinyTapeout/tinytapeout-sky-26b/pulls?q=is%3Apr+is%3Aclosed+gHashTag),
all opened by TinyTapeoutBot, updating the three Trinity repos.

### 3.3 Post-Silicon Validation Plan
> [!IMPORTANT]
> **WITHDRAWN.** The TTSKY26b submission (2026-05-17) was cancelled and
> refunded, so no die will be returned. No Trinity die exists and none is
> scheduled; no silicon route is currently selected. This matches the root
> [`README.md`](../../README.md) ("Hardware exclusivity: WITHDRAWN") and is
> recorded in `trinity-papers-ru` PR #18 (merged 2026-07-31).

When chips return from Sky130 fab (expected ~Nov 2026 for TTSKY26b):
### 3.3 Post-Silicon Validation Plan (not executable — no silicon route)

This plan cannot be executed: the TTSKY26b submission was withdrawn and
refunded, no die exists, and no silicon route is currently selected. It is
kept only as the procedure that *would* apply if a die were ever returned:

1. **Logic analyzer check:** Apply reset, verify `uo_out=0x47`, `uio_out=0xC0`
2. **Clock sweep:** Increase clk from 1 MHz to 100 MHz, measure GF16 dot4
Expand All @@ -165,9 +174,11 @@ When chips return from Sky130 fab (expected ~Nov 2026 for TTSKY26b):
- **TTSKY26b** carries the **Three Crowns** (`tt-trinity-phi`, `tt-trinity-euler`,
`tt-trinity-gamma`) — a multi-chip family with shared phi-arithmetic DNA.

Both shuttles were submitted in May 2026. The README "TTSKY26b" claim is
**correct** for the Three Crowns. The earlier "26a only" wording in this repo
was stale and has been corrected.
Both shuttles were submitted in May 2026, but the TTSKY26b submission
(2026-05-17) was withdrawn and refunded, so no Three Crowns die will be
returned. The README status for TTSKY26b — **WITHDRAWN**, no silicon route
currently selected — is the correct one; any wording in this repo that still
presents TTSKY26b as awaiting silicon is stale.

---

Expand Down