Runtime (wall-clock minutes per design at 200 residues, batch size 2 run divided by 2) of TorchCraft on A100 without cuequivariance kernels on TorchCraft runtime benchmark design runs
7.96 Runtime (wall-clock minutes per design at 200 residues, batch size 2 run divided by 2)
Methods
- Tested configuration
- TorchCraft on A100 without cuequivariance kernels
- Protocol
- TorchCraft runtime per design at 200 residues
- Dataset
- TorchCraft runtime benchmark design runs
- Evaluation
- TorchCraft runtime at 200 residues, GPU w/o cuequivariance
- Coverage
- Not reported scored / Not reported eligible
- Uncertainty
- Not reported by the source: No spread is printed; GPU cells are stated to be the mean of three runs without a spread.
- Evidence
- Author-reported evaluation · source checkedTorchCraft: Unified binder design by inverting an all-atom structure predictor (arXiv:2609.19770v1); TorchX project site (torchx-cpl.github.io) · Table S9, row "GPU w/o cuequivariance", column "200 aa"
A source-checked result verifies the numerical transcription, not every model or protocol detail. Evaluation metadata: source checked. Source checked does not mean independently reproduced.
Reproduction
- Split
- Not reported
- Adaptation
- Batch size 2
- Scoring implementation
- Total wall-clock time of a batch-size-2 run divided by 2
No execution recipe has been verified for this exact configuration and evaluation. A benchmark's general instructions may use different inputs, splits or model settings.
Reproducing this published result requires matching its model configuration, data, split and scorer. Source checking or a successful smoke test does not establish score reproduction.
Evidence
Source checking verifies the cited claim or transcription. It does not establish independent reproduction.
Evidence table
Inspect claims, sources and review details
Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.
One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.
2 evidence rows matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Reported result 7.96 Individual claims | TorchCraft: Unified binder design by inverting an all-atom structure predictor (arXiv:2609.19770v1) Table S9, row "GPU w/o cuequivariance", column "200 aa" Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: arXiv:2609.19770v1, submitted 17 September 2026 | source checked ["independent-cell-check","deterministic-table-parse","pdf-text-parse","source-hash-verification"] · 2026-11-16T01:05:00Z author reported Audit detailsField: Source artifact SHA-256: Hash scope: SHA-256 of the PDF as downloaded. The pdftotext -layout text layer every value was read from hashes to 90f1da5f8a936766087beba0ade3f038f5b69504b9f581b42f2391f5640c5e6b; it is regenerated from the PDF and is not archived here. Extraction artifact SHA-256: |
| Reported result 7.96 Individual claims | TorchX project site (torchx-cpl.github.io) Table S9, row "GPU w/o cuequivariance", column "200 aa" Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Project site as served on 2026-11-15 | source checked ["independent-cell-check","deterministic-table-parse","pdf-text-parse","source-hash-verification"] · 2026-11-16T01:05:00Z author reported Source has a recorded evidence concern. Consult its source page before using the claim. Audit detailsField: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record Extraction artifact SHA-256: |
Sources and history
Release 2026-11-21-c8bea25e5abb · Record review: source checked
2 source records and release history
- TorchCraft: Unified binder design by inverting an all-atom structure predictor (arXiv:2609.19770v1) · Original source · arXiv:2609.19770v1, submitted 17 September 2026
- TorchX project site (torchx-cpl.github.io) · Original source · Project site as served on 2026-11-15
Technical metadata and extraction receipts
Stable ID: torchcraft-20261115-result-runtime-gpu-no-cuequivariance-200aa
- areas
- protein-structure
- contexts
- research
- hardware
- NVIDIA A100 GPU with 80 GB memory, with the default torch kernels
- metric
- runtime
- metric direction
- lower
- metric qualifier
- wall-clock minutes per design at 200 residues, batch size 2 run divided by 2
- missing metadata
- uncertainty: note: No spread is printed; GPU cells are stated to be the mean of three runs without a spread.; reason: unreported
- numeric value
- 7.96
- printed value
- 7.96
- review
- artifact sha256: f8145b5b95552dd30996eda2940507c0f32d09adfc60f44f28458e7872788f21; method: independent-cell-check; deterministic-table-parse; pdf-text-parse; source-hash-verification; method note: Re-read independently with data/omics/torchcraft-20261115/review/recheck_values.py, written for this review and using neither extract/parse_runtime_table.py nor extract/check_transcriptions.py. It locates the Table S9 header line in the refetched text layer and assigns every numeric cell in a body row to the length column whose header label starts nearest to it, asserting a one-to-one mapping, so a value taken from a neighbouring column fails where whitespace splitting would not. It asserted all four row labels, all five column headers and all twenty cells. Negative controls were rejected: a column-shifted value (the 200 aa cell given the 300 aa value) and a row-swapped value (NPU w/ FA given the NPU w/o FA value). The same twenty cells were then read independently out of the HTML table the official project site reprints and agreed cell for cell. npm run batch -- check-values cannot reach this table: its caption pattern (scripts/omics/batch.py, 'caption' regexes) requires a colon or full stop after the table label and the preprint prints 'Table S9 TorchCraft runtime benchmark' with neither, so it reports row_not_found for all twenty. All seven pinned artifacts were re-retrieved by the reviewer on 2026-11-16T00:08Z. Six came back byte-identical to the hash on their source record. The Hugging Face page did not, and cannot: three consecutive fetches gave three different hashes. Normalising its per-request hf-sanitized CSS class hashes leaves one further per-request difference, window.requestId, and the page also prints a download counter; with those normalised the markup is identical. Its licence tag, gate text and dataset card matched the recorded quotations word for word, so the page is a per-retrieval stamp, not a changed page. The pdftotext -layout text layer of the refetched PDF hashes to 90f1da5f8a936766087beba0ade3f038f5b69504b9f581b42f2391f5640c5e6b, the value the extractor recorded.; retrieval url: https://arxiv.org/pdf/2609.19770v1; reviewed at: 2026-11-16T01:05:00Z; reviewer: claude; reviewer note: Claude (Opus 5) review agent, independent of the extractor; no human review claimed. Every pinned artifact was re-retrieved and hashed, and every stored value re-read against it. Nothing was reproduced: no design run, score, docking calculation or binding measurement was repeated, and no value was recomputed from source data.
- source locator
- Table S9, row "GPU w/o cuequivariance", column "200 aa"
- unit
- minute
- unit detail
- minutes per design