Skip to content

Update README official-score publication status - #2

Merged
gwokhou merged 1 commit into
mainfrom
claude/amd-port-status-analysis-1eokn3
Nov 1, 2026
Merged

gwokhou merged 1 commit into
mainfrom
claude/amd-port-status-analysis-1eokn3

Conversation

@gwokhou

@gwokhou gwokhou commented Oct 30, 2026

Copy link
Copy Markdown
Owner

The README still described official scoring as pending and unpublished,
but the corpus policy reports available and RELEASE/release-bundle.json
publishes the gfx1200 score. Describe the published self-evaluation
score and link the release marker.

Co-Authored-By: Claude Opus 5.5 noreply@anthropic.com
Claude-Session: https://claude.ai/code/session_012ndon2aYQGkxNPrcoixrN9
Signed-off-by: Claude noreply@anthropic.com

The README still described official scoring as pending and unpublished,
but the corpus policy reports available and RELEASE/release-bundle.json
publishes the gfx1200 score. Describe the published self-evaluation
score and link the release marker.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012ndon2aYQGkxNPrcoixrN9
Signed-off-by: Claude <noreply@anthropic.com>
@gwokhou
gwokhou merged commit 15833e9 into main Nov 1, 2026
10 of 12 checks passed
gwokhou added a commit that referenced this pull request Nov 7, 2026
Merge PR #2 and workstream updates while retaining both original parents.
Preserve local SOLAR analysis, HIP schedule evidence, and native GPU
lifecycle behavior alongside remote composition, domain schema ownership,
keyword-only dataclasses, and platform refactors. Remove obsolete duplicate
modules and restore command registration, source bindings, and test resources.

Restore diagnostic hardware configuration checks under the native GPU
execution guard. Reject unsupported device models, SKUs, and memory capacity
before loading workloads. Independent review found this omission; the fix
was reviewed again with no remaining findings.

Validation before history consolidation: 4319 CPU tests passed, 3 skipped;
25 focused admission tests passed; 149 related CPU tests passed after the
review fix. Ruff, formatting, ty, coupling, readability, dataclass policy,
schema versions, and production reachability checks passed.
No GPU or macOS runtime validation was performed.

The ignored-data audit reports 191 historical artifact findings under data/.
These local artifacts are unchanged and are not included in this commit.

Signed-off-by: Guohao Zhang <akidezhang@outlook.com>
gwokhou added a commit that referenced this pull request Nov 20, 2026
- #1 dynamic arch emits sidecar limitation, not a misleading register_pressure_high hint

- #2 ccob best-effort gzip/zlib decompress fallback in the AMDGPU metadata scan

- NVIDIA#3 extract_amdgpu_footprints target_architecture filter (multi-arch safe)

- NVIDIA#4 occupancy documented (metadata lacks it; vgpr/limit ratio proxy)

- NVIDIA#5 AMDGPU footprint carries source_sha256 from the artifact

- NVIDIA#6 static_kernel_evidence.md footprint path now covers AMDGPU metadata (ROCm 7.x)

- NVIDIA#7 AMDGPU fallback records an amdgpu-metadata tool_run when footprints are found

- NVIDIA#8 real gfx1200 code-object fixture + test; removed dead _DYNAMIC_LIMITATION

Signed-off-by: Guohao Zhang <akidezhang@outlook.com>
gwokhou added a commit that referenced this pull request Nov 20, 2026
…okn3

Update README official-score publication status
gwokhou added a commit that referenced this pull request Nov 20, 2026
Merge PR #2 and workstream updates while retaining both original parents.
Preserve local SOLAR analysis, HIP schedule evidence, and native GPU
lifecycle behavior alongside remote composition, domain schema ownership,
keyword-only dataclasses, and platform refactors. Remove obsolete duplicate
modules and restore command registration, source bindings, and test resources.

Restore diagnostic hardware configuration checks under the native GPU
execution guard. Reject unsupported device models, SKUs, and memory capacity
before loading workloads. Independent review found this omission; the fix
was reviewed again with no remaining findings.

Validation before history consolidation: 4319 CPU tests passed, 3 skipped;
25 focused admission tests passed; 149 related CPU tests passed after the
review fix. Ruff, formatting, ty, coupling, readability, dataclass policy,
schema versions, and production reachability checks passed.
No GPU or macOS runtime validation was performed.

The ignored-data audit reports 191 historical artifact findings under data/.
These local artifacts are unchanged and are not included in this commit.

Signed-off-by: Guohao Zhang <akidezhang@outlook.com>
@gwokhou
gwokhou deleted the claude/amd-port-status-analysis-1eokn3 branch November 20, 2026 09:46
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants