rafatahmed · Oct 20, 2026 · Oct 20, 2026 · Oct 20, 2026 · Oct 20, 2026 · Oct 20, 2026
diff --git a/.gitattributes b/.gitattributes
@@ -58,3 +58,6 @@ tests/fixtures/phreeqc/**/*.sel -text whitespace=-blank-at-eol,-blank-at-eof
 tests/fixtures/phreeqc/**/*.screen -text whitespace=-blank-at-eol,-blank-at-eof
 tests/fixtures/phreeqc/**/error.inp -text whitespace=-blank-at-eol,-blank-at-eof
 tests/fixtures/phreeqc/**/*.json -text whitespace=cr-at-eol
+
+# Published evidence downloads must match their archived bytes and checksums.
+docs/site/_static/benchmarks/** -text whitespace=cr-at-eol
diff --git a/.github/workflows/benchmark.yml b/.github/workflows/benchmark.yml
@@ -19,7 +19,10 @@ jobs:
           python-version: "3.13"
       - name: Install benchmark tooling
         id: install
-        run: python -m pip install -e . -r benchmarks/requirements.txt
+        run: |
+          python -m pip install -e . -r benchmarks/requirements.txt
+          mkdir -p build/benchmark-v1
+          python -m pip freeze --all > build/benchmark-v1/environment.txt
       - name: Measurement semantics and correctness scorecards
         run: |
           python -m unittest tests.test_benchmark_baseline
@@ -28,9 +31,9 @@ jobs:
         if: ${{ !cancelled() && steps.install.outcome == 'success' }}
         shell: bash
         run: |
-          python -m asv machine --yes --machine github-linux
+          python -m asv machine --yes
           python -m asv check --python=same
-          python -m asv run --python=same --machine github-linux --set-commit-hash "$GITHUB_SHA" --record-samples --show-stderr 2>&1 | tee build/benchmark-v1/asv.log
+          python -m asv run --python=same --set-commit-hash "$GITHUB_SHA" --record-samples --show-stderr 2>&1 | tee build/benchmark-v1/asv.log
       - name: Record provenance and sample summary
         if: always()
         run: python -m tools.benchmark_baseline --metadata-only --compare-inputs build/benchmark-v1/before.json --asv-results build/asv/results --output build/benchmark-v1/after.json

diff --git a/MANIFEST.in b/MANIFEST.in
@@ -10,6 +10,7 @@ include asv.conf.json
 recursive-include benchmarks *.py *.txt
 recursive-include docs *.md *.jpg *.png *.pdf
 recursive-include docs/site *.csv
+recursive-include docs/site/_static/benchmarks *.json *.svg *.zip *.sha256
 exclude docs/sources/04_Fertilizer.pdf
 exclude docs/sources/165357805552.pdf
 exclude docs/sources/215057776303.pdf
@@ -24,6 +25,7 @@ recursive-include src/flahax/data *.md
 recursive-include tests *.py
 recursive-include tests/fixtures *.json *.pqi *.out *.sel *.screen *.dat *.md *.inp
 recursive-include tools *.py *.ps1
+include tools/requirements-benchmark-report.txt
 recursive-include examples *.py *.md
 prune examples/tutorial
 global-exclude __pycache__

diff --git a/README.md b/README.md
@@ -104,6 +104,8 @@ The [Benchmark v1 protocol](docs/site/evidence/benchmark-v1.md) separates
 formulation fit, numerical chemistry, runtime/memory measurements, and physical
 validation. Its reproducible reports include input hashes and failed cases;
 timings are informational, and physical validation remains separate.
+See the [measured results and downloadable evidence](docs/site/evidence/benchmark-results.md)
+for tables, charts, provenance, and retained limitations.
 
 - 153 automated tests, including live replay of every catalogue product,
   phase and acid case against pinned PHREEQC 3.8.6 (USGS).

diff --git a/asv.conf.json b/asv.conf.json
@@ -3,7 +3,7 @@
   "project": "FlahaX",
   "project_url": "https://github.com/rafatahmed/Flahax",
   "repo": ".",
-  "branches": ["main"],
+  "branches": ["HEAD"],
   "environment_type": "virtualenv",
   "pythons": ["3.13"],
   "matrix": {"numpy": ["2.5.4"], "scipy": ["1.18.1"]},

diff --git a/docs/index.md b/docs/index.md
@@ -34,6 +34,7 @@ Readers: the documentation is at <https://flahax.readthedocs.io/>. This index is
 ## Scientific model and evidence
 
 - [Benchmark v1 specification and reproduction](site/evidence/benchmark-v1.md): separate correctness and performance scorecards, fixed inputs, provenance, and retained failures.
+- [Benchmark results and downloadable evidence](site/evidence/benchmark-results.md): a clean-checkout snapshot with tables, charts, raw records, and retained limitations.
 - [Formulation method](site/models/method.md)
 - [Current mixed-equilibrium model](site/models/fertilizer-equilibrium-model.md)
 - [Coverage summary](site/models/full-fertilizer-chemistry.md) and [exact reaction/phase matrix](site/models/chemistry-coverage-matrix.md)

diff --git a/docs/site/_static/benchmarks/2026-10-20-linux-2340a19/CHECKSUMS.sha256 b/docs/site/_static/benchmarks/2026-10-20-linux-2340a19/CHECKSUMS.sha256
@@ -0,0 +1,25 @@
+43386a611b64cf906158205e15973b9e15f1f56fcb466fa3933d1522f30b58ec  README.txt
+f99f6abdfc8d43f9e2f31ac155ae0491cf3166164c7f62c21baaf8da0f2326c0  asv/benchmarks.json
+a9196208a86a1ce31c4e0b149a2523556eee998b8d36d2d402031adc3bff0698  asv/runnervm8k3b6/2340a19c-existing-py_opt_hostedtoolcache_Python_3.13.16_x64_bin_python.json
+fe3ddacd9e176673d737d2bd1195f2acd98cb28cd928cb32a6cc9753d243b079  asv/runnervm8k3b6/machine.json
+028b30ac8388daac5c020d7fc2660a4ea97f20f8e8bff0b66103cc467d458d92  baseline.json
+96624b259490e98e7d5fd78d29ef9c8c15be74744be0a796f189c19be12fa757  capture/after.json
+3fb215f15eb1f6ec6258db749f24561fbd528f628ad45f05c0fab3882b4bb673  capture/asv.log
+db4fa48113eb85c364d20bdb6aa76434fc537aef3e527cb6d8e73fed736c7159  capture/before.json
+c7638a73b7d68d8e4004ebab3a5621128b790c561a47f9d5528a2c93b105f182  capture/environment.txt
+e01b15a325ada20871d70c62403233ca82c9bb0b3f85f538747c48189c7e7b4c  capture/hosted-ci.json
+932e98488abb40677de970159b758e0560be3bc6b092aa663ae902334c34a7bd  capture/hosted-workflow.log
+84392ca9f0d1ec073b8d64acea2bebfbec4901827e5b99b283ce49775adc81a5  capture/manual-workflow.json
+6dd88b1c7fd5220fc214d6c951f650c88a11feb42b786eeb6dd647b16f4e113f  capture/publication-notes.txt
+f3df362be48c0166a1ebe5e66622d2184c944ea9e536be55db7576aba028b8b8  configuration/.github/workflows/benchmark.yml
+091019434d1e0e1f5e789595503f74a856a9eb634fd8260e744fd471e648798f  configuration/asv.conf.json
+f61b1f8a9635d7ce95943d3644805a5460cc38defcaea2f260a6b0df094d098d  configuration/benchmarks/requirements.txt
+3d82e8562062533ce76a384583518dec90e0eea561131d00fb683e78d9edcb97  formulation.csv
+032a618ee4e176002155095c739ea2504e1b7496e88c0b8a8c64b2c6fab4aa35  formulation.png
+b790f4c4d5b44034519288d7b9e525408ddb8a86e4911d5461d63901fff33f55  formulation.svg
+dad031eea253a6467e1a6a44faa0f3d940c2fc1a5c31ea55795fe421492128f8  publish_benchmark_report.py
+886ec157e10373edcae9eb4c3506b3b48c412c6feeb1cffe96c613fd9cb48ef8  renderer-environment.json
+3098cfcf748422ba3c6f4333d893ff67bcf4401dd106b46b6b704e4f4dd25819  requirements-benchmark-report.txt
+46d688b858daa6fabd19577b822cf41966a217f25b679a3dd62dfa7eb7002ab8  timings.csv
+cb3db5a01b9f862820f5753823bcd80bba644d39aca921b69dc511710ddd8811  timings.png
+84957a9923ee3866d5500e24255568da9fe25fd840ccf728760480757abc4e4a  timings.svg