theinterfold · 0xjei · Apr 29, 2026 · Apr 29, 2026 · Apr 29, 2026 · Apr 29, 2026
diff --git a/circuits/benchmarks/README.md b/circuits/benchmarks/README.md
@@ -19,6 +19,32 @@ From this directory:
 
 Options and secure-only **config** circuit behavior are documented in the script and `config.json`.
 
+### What gets stored (secure / insecure)
+
+A full `./run_benchmarks.sh --mode <mode>` run writes:
+
+- `results_<mode>/raw/*.json` — Nargo timing + artifact sizes (source for the **Circuit Benchmarks**
+  table)
+- `results_<mode>/crisp_verify_gas.json` — verify gas, calldata gas, artifact sizes, and
+  **`integration_summary`** from `test_trbfv_actor` when gas extraction succeeds
+- `results_<mode>/integration_summary.json` — snapshot of `.integration_summary` (phase timings,
+  folded proofs, **multithread / operation_timings** after a fresh integration export)
+- `results_<mode>/report.md` — rendered summary of all of the above
+
+### Regenerate `report.md` only (no integration re-run)
+
+From this directory, after you already have `raw/` + `crisp_verify_gas.json`:
+
+```bash
+./regenerate_report.sh
+./regenerate_report.sh --mode insecure
+```
+
+`crisp_verify_gas.json` embeds the integration timings; if you also keep `integration_summary.json`
+in the same folder, the script passes it explicitly (useful when gas JSON is missing a field but the
+snapshot is complete). `regenerate_report.sh` itself does not re-run `test_trbfv_actor`; it renders
+from `results_<mode>/raw`, `crisp_verify_gas.json`, and (optionally) `integration_summary.json`.
+
 ## Refresh after parameter changes
 
 If you change circuit/config parameter sets, rerun the full benchmark + gas extraction flow.
@@ -43,6 +69,30 @@ pnpm -C examples/CRISP/packages/crisp-sdk build
   --gas-json "./circuits/benchmarks/results_insecure/crisp_verify_gas.json"
 ```
 
+If Π_DKG / Π_dec **verify gas** is `N/A` because `crisp_verify_gas.json` came from a failed extract,
+but your integration summary still has `folded_artifacts`, replay only the Hardhat `estimateGas`
+step and merge **dkg** / **dec** into the gas file (no Rust re-run):
+
+```bash
+# For secure folded proofs, align Solidity verifiers first (--build may take a while).
+./circuits/benchmarks/scripts/replay_folded_verify_gas.sh \
+  --summary "/tmp/summary_secure.json" \
+  --gas-json "./circuits/benchmarks/results_secure/crisp_verify_gas.json" \
+  --build secure-8192
+```
+
+If `crisp_verify_gas.json` has `integration_summary: null` but you still have the JSON written by
+`BENCHMARK_SUMMARY_OUTPUT` from a successful `test_trbfv_actor` run (e.g.
+`/tmp/summary_secure.json`), pass it so phase timings and folded sizes match that run:
+
+```bash
+./circuits/benchmarks/scripts/generate_report.sh \
+  --input-dir "./circuits/benchmarks/results_secure/raw" \
+  --output "./circuits/benchmarks/results_secure/report.md" \
+  --gas-json "./circuits/benchmarks/results_secure/crisp_verify_gas.json" \
+  --integration-summary "/tmp/summary_secure.json"
+```
+
 For secure mode, use `--mode secure` and replace `results_insecure` with `results_secure`.
 
 ## Reported protocol tables
@@ -53,6 +103,10 @@ For secure mode, use `--mode secure` and replace `results_insecure` with `result
   `C4b`, `C5`, `user-data-encryption`, `C6`, `C7`.
 - `Artifacts` for `Π_DKG`, `Π_user`, `Π_dec` with proof/public-input sizes and gas columns.
 - `Role / Phase / Activity` for P1..P4 operational cost summaries.
+- When `integration_summary` is present, the report also includes:
+  - an `Integration test` section (end-to-end phase wall-clock timings)
+  - a `Thread pool` section (Rayon threads / cores)
+  - `CPU-bound operation timings` (tracked in-process averages/totals)
 
 ## Derivation rules
 

diff --git a/circuits/benchmarks/config.json b/circuits/benchmarks/config.json
@@ -6,7 +6,6 @@
     "dkg/e_sm_share_computation",
     "dkg/share_encryption",
     "dkg/share_decryption",
-    "threshold/user_data_encryption",
     "threshold/user_data_encryption_ct0",
     "threshold/user_data_encryption_ct1",
     "threshold/pk_generation",

diff --git a/circuits/benchmarks/regenerate_report.sh b/circuits/benchmarks/regenerate_report.sh
@@ -0,0 +1,3 @@
+#!/bin/bash
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+exec "${SCRIPT_DIR}/scripts/regenerate_report.sh" "$@"
diff --git a/circuits/benchmarks/results_insecure/crisp_verify_gas.json b/circuits/benchmarks/results_insecure/crisp_verify_gas.json
diff --git a/circuits/benchmarks/results_insecure/integration_summary.json b/circuits/benchmarks/results_insecure/integration_summary.json
diff --git a/circuits/benchmarks/results_insecure/report.md b/circuits/benchmarks/results_insecure/report.md
@@ -1,9 +1,11 @@
 # Enclave ZK Circuit Benchmarks
 
-**Generated:** 2026-04-27 10:35:23 UTC
+**Generated:** 2026-04-29 10:02:54 UTC
 
-**Git Branch:** `main`  
-**Git Commit:** `e651f30205117e83a9cc48ac44664d434466724d`
+**Git Branch:** `feat/benches`  
+**Git Commit:** `36c01c62b86e2527279842280337f2f4724d2487`
+
+**Committee Size:** `H=3`, `N=3`, `T=1`
 
 ---
 
@@ -13,36 +15,107 @@
 
 | Circuit              | Constraints | Prove time (s) | Verify time (ms) | Proof size (KB) |
 | -------------------- | ----------- | -------------- | ---------------- | --------------- |
-| C0                   | 6847        | 0.12           | 25.76            | 15.88           |
-| C1                   | 57818       | 0.33           | 26.00            | 15.88           |
-| C2a                  | 142625      | 0.77           | 26.58            | 15.88           |
-| C2b                  | 198355      | 0.86           | 26.30            | 15.88           |
-| C3a                  | 132633      | 0.79           | 25.93            | 15.88           |
-| C3b                  | 132633      | 0.79           | 25.93            | 15.88           |
-| C4a                  | 92515       | 0.49           | 25.95            | 15.88           |
-| C4b                  | 92515       | 0.49           | 25.95            | 15.88           |
-| C5                   | 151717      | 0.79           | 25.87            | 15.88           |
-| user_data_encryption | 53732       | 0.32           | 25.68            | 15.88           |
-| C6                   | 86927       | 0.52           | 26.68            | 15.88           |
-| C7                   | 104273      | 0.49           | 26.20            | 15.88           |
+| C0                   | 6847        | 0.12           | 24.97            | 15.88           |
+| C1                   | 57818       | 0.34           | 25.84            | 15.88           |
+| C2a                  | 142625      | 0.78           | 25.50            | 15.88           |
+| C2b                  | 198355      | 0.85           | 26.06            | 15.88           |
+| C3a                  | 132633      | 0.78           | 25.14            | 15.88           |
+| C3b                  | 132633      | 0.78           | 25.14            | 15.88           |
+| C4a                  | 92515       | 0.50           | 25.00            | 15.88           |
+| C4b                  | 92515       | 0.50           | 25.00            | 15.88           |
+| C5                   | 151717      | 0.79           | 26.28            | 15.88           |
+| user_data_encryption | 53732       | 0.33           | 24.58            | 15.88           |
+| C6                   | 86927       | 0.52           | 25.34            | 15.88           |
+| C7                   | 104273      | 0.49           | 25.42            | 15.88           |
 
 ### Artifacts
 
 | Artifact | Proof size | Public input size | Verify gas | Calldata gas | Total gas |
 | -------- | ---------- | ----------------- | ---------- | ------------ | --------- |
-| Π_DKG    | 15.88 KB   | 0.12 KB           | 3037849    | 179452       | 3217301   |
-| Π_user   | 31.75 KB   | 0.22 KB           | 2972953    | 340140       | 3313093   |
-| Π_dec    | 15.88 KB   | 3.25 KB           | 3549211    | 188232       | 3737443   |
+| Π_DKG    | 10.69 KB   | 0.41 KB           | 3037922    | 175616       | 3213538   |
+| Π_user   | 15.88 KB   | 0.12 KB           | 2973073    | 170272       | 3143345   |
+| Π_dec    | 10.69 KB   | 3.41 KB           | 3549077    | 186764       | 3735841   |
 
 ### Role / Phase / Activity
 
 | Role            | Phase | Activity                         | Prove time | Proof size | Bandwidth |
 | --------------- | ----- | -------------------------------- | ---------- | ---------- | --------- |
-| Each ciphernode | P1    | one-time DKG participation       | 3.37 s     | 95.25 KB   | 96.12 KB  |
-| Aggregator      | P2    | combine folds + C5               | 0.79 s     | 15.88 KB   | 16.00 KB  |
-| User            | P3    | per user input                   | 0.64 s     | 31.75 KB   | 31.97 KB  |
+| Each ciphernode | P1    | one-time DKG participation       | 379.38 s   | 127.00 KB  | 128.19 KB |
+| Aggregator      | P2    | combine folds + C5               | 0.79 s     | 10.69 KB   | 11.09 KB  |
+| User            | P3    | per user input                   | 0.65 s     | 15.88 KB   | 16.00 KB  |
 | Each ciphernode | P4    | per computation output (C6)      | 0.52 s     | 15.88 KB   | 16.00 KB  |
-| Aggregator      | P4    | per computation output (C7+fold) | 0.49 s     | 15.88 KB   | 19.12 KB  |
+| Aggregator      | P4    | per computation output (C7+fold) | 80.31 s    | 10.69 KB   | 14.09 KB  |
+
+## Integration test (`test_trbfv_actor`)
+
+### End-to-end phase timings (wall clock)
+
+| Phase                                       | Duration (s) |
+| ------------------------------------------- | ------------ |
+| Starting trbfv actor test                   | 0.00         |
+| Setup completed                             | 2.98         |
+| Committee Setup Completed                   | 20.25        |
+| Committee Finalization Complete             | 0.01         |
+| ThresholdShares -> PublicKeyAggregated      | 379.38       |
+| E3Request -> PublicKeyAggregated            | 381.92       |
+| Application CT Gen                          | 0.31         |
+| Running FHE Application                     | 0.00         |
+| Ciphertext published -> PlaintextAggregated | 80.31        |
+| Entire Test                                 | 485.79       |
+
+### Thread pool (same process as integration test)
+
+| Setting                      | Value |
+| ---------------------------- | ----- |
+| Rayon threads                | 13    |
+| Max simultaneous Rayon tasks | 1     |
+| Cores available              | 14    |
+
+### CPU-bound operation timings (tracked in-process)
+
+| Name                          | Avg (s) | Runs | Total (s) |
+| ----------------------------- | ------- | ---- | --------- |
+| CalculateDecryptionKey        | 0.11    | 3    | 0.33      |
+| CalculateDecryptionShare      | 0.61    | 3    | 1.84      |
+| CalculateThresholdDecryption  | 0.57    | 1    | 0.57      |
+| GenEsiSss                     | 0.12    | 3    | 0.37      |
+| GenPkShareAndSkSss            | 0.22    | 3    | 0.67      |
+| ZkDecryptedSharesAggregation  | 8.54    | 1    | 8.54      |
+| ZkDecryptionAggregation       | 49.65   | 1    | 49.65     |
+| ZkDkgAggregation              | 20.66   | 1    | 20.66     |
+| ZkDkgShareDecryption          | 1.45    | 6    | 8.68      |
+| ZkNodeDkgFold                 | 78.08   | 3    | 234.23    |
+| ZkPkAggregation               | 2.19    | 1    | 2.19      |
+| ZkPkBfv                       | 0.33    | 3    | 0.99      |
+| ZkPkGeneration                | 1.34    | 3    | 4.01      |
+| ZkShareComputation            | 2.65    | 6    | 15.91     |
+| ZkShareEncryption             | 2.47    | 36   | 89.00     |
+| ZkThresholdShareDecryption    | 6.20    | 3    | 18.59     |
+| ZkVerifyShareDecryptionProofs | 0.10    | 3    | 0.30      |
+| ZkVerifyShareProofs           | 0.22    | 5    | 1.08      |
+
+Sum of tracked operation wall time: **457.62 s** (often much larger than end-to-end wall clock
+because work runs in parallel).
+
+## Raw circuit benchmark JSON (Nargo)
+
+Source files for the **Circuit Benchmarks** table. Persist this directory with
+`crisp_verify_gas.json` (and optional `integration_summary.json`) to regenerate the report without
+re-running the integration test.
+
+| File                                                  |
+| ----------------------------------------------------- |
+| `dkg_e_sm_share_computation_default.json`             |
+| `dkg_pk_default.json`                                 |
+| `dkg_share_decryption_default.json`                   |
+| `dkg_share_encryption_default.json`                   |
+| `dkg_sk_share_computation_default.json`               |
+| `threshold_decrypted_shares_aggregation_default.json` |
+| `threshold_pk_aggregation_default.json`               |
+| `threshold_pk_generation_default.json`                |
+| `threshold_share_decryption_default.json`             |
+| `threshold_user_data_encryption_ct0_default.json`     |
+| `threshold_user_data_encryption_ct1_default.json`     |
 
 ## System Information
 
@@ -60,3 +133,8 @@
   1.0.0-beta.16+2d46fca7203545cbbfb31a0d0328de6c10a8db95 (git version hash:
   2d46fca7203545cbbfb31a0d0328de6c10a8db95, is dirty: false)
 - **Barretenberg Version:** 3.0.0-nightly.20260102
+
+## Notes
+
+- All nodes are executed on the same machine in this benchmark run, so inter-node network latency is
+  effectively 0.
diff --git a/circuits/benchmarks/results_secure/crisp_verify_gas.json b/circuits/benchmarks/results_secure/crisp_verify_gas.json