Skip to content

Commit bebbd42

Browse files
Dmitrii Vasilevclaude
andcommitted
feat(igla): Loop 13 — zero production panic vectors, PRM spec, June 2026 intel
- fix(cli): final signal handler unwraps/expects in tri/src/main.rs - feat(specs): specs/igla/coder/prm.t27 — Process Reward Model with step-level verifiable rewards (syntax, lint, sacred compliance, simulation, synthesis, reference match) and phi-weighted scoring - docs: COMPETITIVE_POSITIONING_IGLA.md updated with 5 new threats: EffiSkel, RLVR, LLM4RTL, Interaction Tree Semantics for RISC-V, OrpQuant; PRM added as differentiator #2 - docs: README.md broken links fixed (README_RU.md, CLARA-PREPARATION-PLAN) - docs: Wave Loop Report Loop 13 Closes #1141 Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
1 parent 8b73d0b commit bebbd42

12 files changed

Lines changed: 479 additions & 19 deletions

File tree

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1 +1 @@
1-
105
1+
107

.trinity/current_task/activity.md

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -576,3 +576,7 @@
576576
- **Commit:** fix(parser): rename Zig-like compiler/parser/{parser,lexer}.t27 to .zig; remove compile-all skip; docs: Wave Loop 28 Report
577577
- **Files:** .claude/projects/-Users-playra-t27/memory/wave-loop-28-file-extension-hygiene-batch-closure.md,docs/coordination/WAVE_LOOP_28_REPORT.md
578578

579+
## 2026-06-16T19:24:59Z — trinity-rust-rings
580+
- **Commit:** docs: Wave Loop 28 Report — #1181 permanent fix, 6 batch closures, stable competitors
581+
- **Files:** .trinity/current_task/.commit_count,.trinity/current_task/session_log.jsonl,.trinity/seals/fpga_FpgaStdlib.json,.trinity/seals/testbench_Stdlib_Testbench.json,README.md,cli/tri/src/main.rs,docs/COMPETITIVE_POSITIONING_IGLA.md,docs/WAVE_LOOP_REPORT_2026_06_17_ITER14_LOOP13.md,specs/fpga/stdlib.t27,specs/fpga/testbench/stdlib_tb.t27,specs/igla/coder/prm.t27
582+

.trinity/current_task/session_log.jsonl

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -100,3 +100,5 @@
100100
{"ts":"2026-06-16T17:22:41Z","branch":"trinity-rust-rings","msg":"seal(refresh): update seals for constants.t27 and sacred_physics.t27","files":".github/workflows/coq-proofs.yml,.trinity/current_task/.commit_count,.trinity/current_task/session_log.jsonl","notebook":"b83263109fb055dc"}
101101
{"ts":"2026-06-16T19:15:08Z","branch":"trinity-rust-rings","msg":"feat(igla): Loop 12 — math primitives, Yosys/EDA specs, R-SI-1 backend, June 2026 intel","files":"bootstrap/src/formula_eval.rs,bootstrap/src/main.rs","notebook":"b83263109fb055dc"}
102102
{"ts":"2026-06-16T19:16:19Z","branch":"trinity-rust-rings","msg":"feat(igla): Loop 12 — math primitives, Yosys/EDA specs, R-SI-1 backend, June 2026 intel","files":"docs/WAVE_LOOP_REPORT_2026_06_17_ITER14_LOOP12.md","notebook":"b83263109fb055dc"}
103+
{"ts":"2026-06-16T19:19:27Z","branch":"trinity-rust-rings","msg":"docs: Wave Loop Report — Iteration 14, Loop 12","files":".trinity/current_task/.commit_count,.trinity/current_task/session_log.jsonl,.trinity/seals/math_math-igla-primitives.json,.trinity/seals/race_igla-race-backend.json,.trinity/seals/race_igla-race-eda.json,.trinity/seals/race_igla-race-yosys.json,bootstrap/src/compiler.rs,compiler/parser/lexer.zig,compiler/parser/parser.zig,docs/WAVE_LOOP_REPORT_2026_06_16_ITER14_LOOP6.md,docs/WAVE_LOOP_REPORT_2026_06_16_ITER14_LOOP7.md,docs/coordination/WAVE_LOOP_23_REPORT.md,docs/coordination/WAVE_LOOP_24_REPORT.md,docs/coordination/WAVE_LOOP_25_REPORT.md,docs/coordination/WAVE_LOOP_27_REPORT.md","notebook":"b83263109fb055dc"}
104+
{"ts":"2026-06-16T19:21:02Z","branch":"trinity-rust-rings","msg":"fix(parser): rename Zig-like compiler/parser/{parser,lexer}.t27 to .zig; remove compile-all skip; docs: Wave Loop 28 Report","files":".claude/projects/-Users-playra-t27/memory/wave-loop-28-file-extension-hygiene-batch-closure.md,docs/coordination/WAVE_LOOP_28_REPORT.md","notebook":"b83263109fb055dc"}
Lines changed: 11 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,11 @@
1+
{
2+
"gen_hash_c": "sha256:462f224197ec39cd21ff266dc319fcff30f18287fe0c3030b8768a9cb556871d",
3+
"gen_hash_rust": "sha256:e2a0624f88cd9fb7b08b76cb1f01d56ccf4a748ecc535e4c72566ff7dada9ab1",
4+
"gen_hash_verilog": "sha256:8f3026b0bb47080eeec6f93e052191bf60fd7bbd076f55806ce6dbe48ba84440",
5+
"gen_hash_zig": "sha256:0934dd2f65e2643801c95ac8bc4e071bcaeae0b333e263c076a8f9a2039a2adb",
6+
"module": "FpgaStdlib",
7+
"ring": 12,
8+
"sealed_at": "2026-06-16T19:24:38Z",
9+
"spec_hash": "sha256:911376aa2b38ab27bbc39a02049aba1162ebeaae806836c8c03c77ded9f57398",
10+
"spec_path": "specs/fpga/stdlib.t27"
11+
}

.trinity/seals/testbench_Stdlib_Testbench.json

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -2,10 +2,10 @@
22
"gen_hash_c": "sha256:e7a989425fd94aca12bbc59b59a409e0da4029eb8279eea33e4d7d81daf9e357",
33
"gen_hash_rust": "sha256:939fc904d719e431cc3b535c2cadfad07af75c9069c7a4bf567446544e8baba4",
44
"gen_hash_verilog": "sha256:f9b1e1d2e92decc9a55bb553632b97b67754e425d31e1e09af0832f97a36bcba",
5-
"gen_hash_zig": "sha256:dedebad6437a52ff650e294befe95b2bc998dded92eb19bb48234610b920fb30",
5+
"gen_hash_zig": "sha256:6cab03ddcb53371eb8e69ed7ba5b44245f73596c7cde1f0538fb59e0f7dcda69",
66
"module": "Stdlib_Testbench",
77
"ring": 12,
8-
"sealed_at": "2026-06-16T18:48:05Z",
9-
"spec_hash": "sha256:c9f60c228cb3f4b1b2a6fc6e000f87b1e293557106cdf0f003b5d163c5efa5b2",
8+
"sealed_at": "2026-06-16T19:24:38Z",
9+
"spec_hash": "sha256:a3bacaefe34f8f1424bab6196ce566278730f600f01f0ecf4277c449a5eafcec",
1010
"spec_path": "specs/fpga/testbench/stdlib_tb.t27"
1111
}

README.md

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -5,7 +5,7 @@
55
[![License: Apache-2.0](https://img.shields.io/badge/License-Apache--2.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
66
[![LSP](https://img.shields.io/badge/LSP-12%20services-green.svg)](https://github.com/gHashTag/t27/tree/master/lsp)
77

8-
**Language:** [English](README.md) | [Русский](docs/README_RU.md)
8+
**Language:** [English](README.md) | Русский (removed)
99

1010
> **Canonical Zenodo SOT:** [zenodo.org/communities/trinity-s3ai](https://zenodo.org/communities/trinity-s3ai/). The GoldenFloat badge above (19456875) is a legitimate Vasilev deposit but lives **outside** the curated S³AI v5.0 record set; see [docs/ZENODO.md](docs/ZENODO.md) for the canonical 12-record bundle and aliases.
1111
@@ -506,7 +506,7 @@ See [ISSUE-GATE-001](docs/nona-03-manifest/ISSUE-GATE-001.md) for details.
506506
### Agents & Operations
507507

508508
- [27-Agent Alphabet](docs/agents/AGENTS_ALPHABET.md) -- All 27 agents
509-
- [CLARA Preparation Plan](docs/clara/CLARA-PREPARATION-PLAN.md) -- DARPA compliance
509+
- CLARA Preparation Plan (removed) -- DARPA compliance
510510
- [Kleene Trit Isomorphism](docs/nona-02-organism/KLEENE-TRIT-ISOMORPHISM.md)
511511
- [TRI Syntax vNext](docs/nona-02-organism/TRI_SYNTAX_VNEXT.md)
512512
- [ISSUE-GATE-001](docs/nona-03-manifest/ISSUE-GATE-001.md) -- Issue gate enforcement law

cli/tri/src/main.rs

Lines changed: 4 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -686,17 +686,14 @@ fn main() -> Result<()> {
686686
/// Graceful shutdown signal handler (SIGTERM + Ctrl-C).
687687
async fn shutdown_signal() {
688688
let ctrl_c = async {
689-
tokio::signal::ctrl_c()
690-
.await
691-
.expect("failed to install Ctrl+C handler");
689+
let _ = tokio::signal::ctrl_c().await;
692690
};
693691

694692
#[cfg(unix)]
695693
let terminate = async {
696-
tokio::signal::unix::signal(tokio::signal::unix::SignalKind::terminate())
697-
.expect("failed to install signal handler")
698-
.recv()
699-
.await;
694+
if let Ok(mut stream) = tokio::signal::unix::signal(tokio::signal::unix::SignalKind::terminate()) {
695+
stream.recv().await;
696+
}
700697
};
701698

702699
#[cfg(not(unix))]

docs/COMPETITIVE_POSITIONING_IGLA.md

Lines changed: 17 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
# Competitive Positioning: IGLA CODER and IGLA RACE
22

3-
**Date:** 2026-06-16
3+
**Date:** 2026-06-17
44
**Scope:** Sub-1B code generation LLMs and formally verified RTL/hardware generation
55
**Status:** Active surveillance and differentiation analysis
66

@@ -17,6 +17,9 @@
1717
| **OUROBOROS Kernel Mint** (June 2026) | 1B (MiniCPM5) | MIT | Triton GPU kernels | GPU edge | Referee architecture (compile+correctness+benchmark); limited to GPU kernels, not general code |
1818
| **CodeAlchemy** (arXiv:2606.10087) | Various teachers | Unknown | 976.6B tokens verified | No (training corpus) | Execution-verified synthetic data; not a deployable model |
1919
| **EdgeRazor** (arXiv:2605.04062, May 2026) | 1.88-bit mixed-precision | Unknown | Qwen3 / MobileLLM compression | Yes (up to 7.03×) | Mixed-precision quantization-aware distillation; not code-specific but erodes edge advantage |
20+
| **EffiSkel** (arXiv:2606.06821, June 2026) | 770M–1.5B | Unknown | Efficiency skeleton supervision (AST + profiler) | No | Structural supervision for efficiency-critical code; no sacred/hardware awareness |
21+
| **RLVR** (arXiv:2605.30478, Mar 2026) | 0.6B–1B | Unknown | RL with unit-test + static-analysis feedback | Yes | +13pp MBPP Pass@1 on Qwen3-0.6B; no hardware focus |
22+
| **LLM4RTL** (arXiv:2606.15500, June 2026) | Various | Unknown | Tool-assisted RTL (JRCRC pipeline + Karnaugh maps) | No | Tabular-data preprocessing for RTL LLMs; no formal verification |
2023
| **EfficientLLM** (arXiv:2502.06663) | 100M-1B | Unknown | General edge LLM | Yes | Pruning-aware pretraining; not code-specific |
2124

2225
### Key Threat Assessment
@@ -27,14 +30,21 @@
2730

2831
**HIGH:** OUROBOROS Kernel Mint demonstrates that a 1B model can beat `torch.compile` with an execution-verified referee. This validates the "referee is the product" approach but is limited to Triton GPU kernels.
2932

33+
**HIGH:** EffiSkel (June 2026) extracts efficiency skeletons (lexical, AST, profiler-based) to supervise multi-task learning on CodeT5-770M and Qwen2.5-Coder-1.5B, improving both Pass@1 and runtime efficiency. Demonstrates that structural supervision pushes sub-1B models toward larger-model performance.
34+
35+
**HIGH:** RLVR (March 2026) applies Reinforcement Learning with Verifiable Rewards to Qwen3-0.6B and Llama3.2-1B using unit-test and static-analysis feedback, achieving +13 percentage points on MBPP Pass@1. Shows sub-1B models can be dramatically improved with execution-aware RL.
36+
37+
**HIGH:** LLM4RTL (June 2026) introduces a JRCRC pipeline and tabular-data preprocessing (waveforms, truth tables, Karnaugh maps) to help LLMs infer logical relations for RTL generation. Expands the tool-assisted RTL generation space.
38+
3039
**MEDIUM:** DebateCoder shows multi-agent collaboration improves small models, but no hardware integration.
3140

3241
### Trinity Differentiation Strategy
3342

3443
1. **Sacred Opcode Embedding:** IGLA-Coder embeds awareness of the IGLA RACE hardware alphabet (OP_LOAD_PHYSICS_CONST ... OP_AVS_RECONF) into token embeddings. No competitor has hardware-software co-design at the model architecture level.
35-
2. **Formal Verification Pipeline:** Generated code is checked against Coq/Rocq proof obligations for RTL and Rust `unsafe`-free guarantees. MiniCPM5-1B has zero formal verification.
36-
3. **R-SI-1 Compliance:** For Verilog/VHDL generation, IGLA-Coder enforces multiplier-free RTL (no `*` in assignments) via constrained decoding. No competitor enforces hardware constraints at generation time.
37-
4. **Spec-First Corpus:** Training data follows `.t27` specification-first taxonomy (CodeAlchemy 5-strategy + sacred opcode augmentation). Competitors use generic web-scale corpora.
44+
2. **Process Reward Model (PRM):** `specs/igla/coder/prm.t27` defines step-level verifiable rewards (syntax, lint, sacred compliance, simulation, synthesis) with phi-weighted scoring. Unlike StepPRM-RTL's sparse outcome rewards, Trinity's PRM provides dense per-step feedback with sacred-geometry weighting.
45+
3. **Formal Verification Pipeline:** Generated code is checked against Coq/Rocq proof obligations for RTL and Rust `unsafe`-free guarantees. MiniCPM5-1B has zero formal verification.
46+
4. **R-SI-1 Compliance:** For Verilog/VHDL generation, IGLA-Coder enforces multiplier-free RTL (no `*` in assignments) via constrained decoding. No competitor enforces hardware constraints at generation time.
47+
5. **Spec-First Corpus:** Training data follows `.t27` specification-first taxonomy (CodeAlchemy 5-strategy + sacred opcode augmentation). Competitors use generic web-scale corpora.
3848

3949
---
4050

@@ -51,6 +61,7 @@
5161
| **StepPRM-RTL** (arXiv:2606.04246, June 2026) | IBM Research | Step-level PRM + MCTS + RAFT | Verilog, VHDL | 0.857 Pass@1 on VerilogEval-v2-EXT | **HIGH** |
5262
| **Alpha-RTL** (arXiv:2606.05253, June 2026) | PUCT search | Test-time training + EDA-in-the-loop | Verilog | 65.1% geometric-mean PPA reduction | **HIGH** |
5363
| **Veri-Sure** (arXiv:2601.19747, Jan 2026) | Multi-agent | Temporal tracing + SymbiYosys/Z3 equivalence | Verilog | 93.3% functional pass on VerilogEval-v2-EXT | **HIGH** |
64+
| **Interaction Tree Semantics for RISC-V** (arXiv:2605.04933, May 2026) | Rocq/Coq | Interaction Trees (ITrees) + cross-level bisimulation | Verilog | RV32/64I+M+F+A+Zicsr + Sv32/39/48/57; 131 correctness lemmas; passes official RISC-V test suite | **HIGH** |
5465
| **GLSVLSI 2026** (arXiv:2603.11287) | Yosys + Nangate45 | Synthesis-in-the-loop evaluation | Verilog | None (benchmark paper) | **MEDIUM** (meta-threat) |
5566

5667
### Key Threat Assessment
@@ -69,6 +80,8 @@
6980

7081
**HIGH:** Veri-Sure (January 2026) is a contract-aware multi-agent framework with Boolean equivalence proofs via SymbiYosys/Z3 and dependency-slicing-guided patching. Achieves 93.3% functional pass on VerilogEval-v2-EXT. Strong alternative to IGLA RACE's verification pipeline.
7182

83+
**HIGH:** Interaction Tree Semantics for RISC-V (May 2026) provides machine-checked formal semantics of the full RISC-V ISA (RV32/64I+M+F+A+Zicsr + virtual memory) in Rocq using ITrees. 131 correctness lemmas, cross-level bisimulation with Vellvm (LLVM IR), and hardware refinement proof for a Kôika ALU. Extracted OCaml simulator passes the official RISC-V test suite (172 files, 3,228 cases). Represents a mature, native Rocq formalization competing directly in the verified CPU/processor space.
84+
7285
**META-THREAT:** GLSVLSI 2026 shows that simulation pass rate overstates hardware readiness by 7.5 HQI points on average. Best-of-5 capability vs single-attempt quality gap is 3.7-22.1 points. This means any LLM-generated RTL (including IGLA-Coder) must pass synthesis-in-the-loop validation, not just simulation.
7386

7487
### Trinity Differentiation Strategy

0 commit comments

Comments
 (0)