{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.25061", "kind": "paper", "url": "https://arxiv.org/abs/2608.25061"}, {"kind": "website", "url": "https://kerneldf.github.io/datakernelbench"}], "authors": ["Gokul Karthik Kumar", "Yotam Perlitz", "Corey Lammie", "Andrea Giovannini", "Katja Hose"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.25061", "source_url": "https://arxiv.org/abs/2608.25061"}, "categories": ["General AI", "Language & Knowledge", "cs.CL"], "description": {"en": "Evaluates LLM-generated CUDA or Triton kernels for database query operations on GPUs, measuring execution speedup over torch.compile on TPC-H workloads."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.25061", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.25061", "title": "DataKernelBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2608.25061", "modality": null, "name": "DataKernelBench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_datakernelbench_0cd9eeb8", "origin_source": "arxiv", "recognition_confidence": "0.6", "record_sha256": "5961dd884fafe62f78af8f33993a346ad5e7efbb4037c5ede493712f11859cc2", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-08-27T04:14:52.836173Z", "source_url": "https://arxiv.org/abs/2608.25061"}, "publisher": null, "released": "2026-08-25", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.25061", "source_url": "https://arxiv.org/abs/2608.25061"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-25061", "source": "claire_radar", "source_benchmark_id": "2608.25061", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 1, "hfPaperUrl": "https://huggingface.co/papers/2608.25061", "observedAt": "2026-10-04T03:02:42.754940Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-09-25", "state": "stale"}}, "status": "stale"}, "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "The benchmark targets an emerging intersection of database and GPU optimization with multiple model evaluations and a project page, likely to draw moderate interest from the systems and ML communities.", "score": 65}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-27T04:14:52.836173Z", "inputHash": "8927d6eddb6caeac73eb74dffbe107a079fe77b3752eea5492a28e7e1afdeac6", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "DataKernelBench: Can LLMs Optimize Database Queries on GPUs?", "canonicalNameSource": "paper_title", "decisionReason": "The paper introduces a named benchmark with specified tasks, evaluation metrics, and a public project page, indicating a stable scoring contract and reuse path.", "model": "deepseek-v4-pro", "reviewedAt": "2026-08-27T04:14:52.836173Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Evaluates LLM-generated CUDA or Triton kernels for database query operations on GPUs, measuring execution speedup over torch.compile on TPC-H workloads.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "score_submission", "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "public artifact URL"], "snippet": "We introduce DataKernelBench, which translates SQL into validated PyTorch TorchPlan programs and evaluates LLMs that optimize either the core tensor-bounded snippet or the full query in CUDA or Triton through execution-guided repair."}, "familyId": "bmf_1d6dec07467a", "firstSeenAt": "2026-08-27", "heat": null, "id": "bm_datakernelbench_0cd9eeb8", "indexedAt": "2026-08-27T04:12:10.575570Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2608.25061", "pdf": "https://arxiv.org/pdf/2608.25061", "project": "https://kerneldf.github.io/datakernelbench", "report": "https://arxiv.org/abs/2608.25061"}, "metrics": [], "motivation": "GPUs increasingly accelerate database systems, but query-specific peak performance still often relies on hand-written kernels.", "name": "DataKernelBench", "oneLine": "Evaluates LLM-generated CUDA or Triton kernels for database query operations on GPUs, measuring execution speedup over torch.compile on TPC-H workloads.", "paperTitle": "DataKernelBench: Can LLMs Optimize Database Queries on GPUs?", "primaryDomain": "General AI", "publication": {"evidence": "Accepted at EMNLP 2026. Homepage: https://kerneldf.github.io/datakernelbench", "evidenceLevel": "author-claim", "evidenceUrl": "https://arxiv.org/abs/2608.25061", "source": "arxiv-comments", "status": "acceptance_claimed", "venue": "EMNLP 2026", "verifiedAt": "2026-08-27T04:12:10.575570Z"}, "publications": [], "publishers": [{"name": "DataKernelBench Team", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://kerneldf.github.io/datakernelbench"}], "ranking": {"90d": {"confidence": "Low", "coverage": 0.15, "growth": {"components": {"githubStars": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfDatasetDownloads": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfPaperUpvotes": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 90}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": null, "value": null}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.12247537038455211, "value": 1}}, "confidence": "Low", "coverage": 0.15, "rank": null, "score": 44}, "method": "attention-ranking-v13", "rank": null, "score": 44}}, "readiness": "Paper only", "recognitionConfidence": 0.6, "relation": "introduces", "releasedAt": "2026-08-25", "source": {"authors": ["Gokul Karthik Kumar", "Yotam Perlitz", "Corey Lammie", "Andrea Giovannini", "Katja Hose"], "categories": ["cs.CL", "cs.AI", "cs.DB", "cs.LG", "cs.PL"], "id": "2608.25061", "title": "DataKernelBench: Can LLMs Optimize Database Queries on GPUs?", "type": "arxiv", "url": "https://arxiv.org/abs/2608.25061"}, "sourceUpdatedAt": "2026-08-25T18:57:39Z", "topics": ["cs.CL"], "venueAttempts": [{"decisionRaw": "Accepted at EMNLP 2026. Homepage: https://kerneldf.github.io/datakernelbench", "evidence": [{"level": "author-claim", "observedAt": "2026-08-27T04:12:10.575570Z", "rawValue": "Accepted at EMNLP 2026. Homepage: https://kerneldf.github.io/datakernelbench", "sourceType": "arxiv-comments", "sourceUrl": "https://arxiv.org/abs/2608.25061"}], "reviewStatus": "accepted", "venueName": "EMNLP 2026"}], "whyItMatters": "Addresses the lack of benchmarks for database-style operators in LLM kernel generation, enabling comparison of models on real query optimization tasks."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
