{"total":3,"items":[{"citing_arxiv_id":"2606.08339","ref_index":1,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Floating-point autotuning with customized precisions","primary_cat":"cs.MS","submitted_at":"2026-06-06T21:10:12+00:00","verdict":"UNVERDICTED","verdict_confidence":"LOW","novelty_score":4.0,"formal_verification":"none","one_line_summary":"PROMISE tool automates mixed-precision tuning with user-defined floating-point formats, validated on linear solvers and Rodinia benchmarks showing many variables can use lower precision safely.","context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null},{"citing_arxiv_id":"2604.10180","ref_index":43,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"Tessera: Unlocking Heterogeneous GPUs through Kernel-Granularity Disaggregation","primary_cat":"cs.DC","submitted_at":"2026-04-11T12:19:11+00:00","verdict":"CONDITIONAL","verdict_confidence":"MODERATE","novelty_score":6.0,"formal_verification":"none","one_line_summary":"R-DSGD and R-DSGD-M under (δ,c)-robust aggregation have tight Byzantine error floors under (B,ζ)-bounded dissimilarity; local momentum eliminates the stochastic-noise term but not the heterogeneity term.","context_count":1,"top_context_role":"method","top_context_polarity":"use_method","context_text":"send/recvoperations can be captured as GPU-side ker- nels and embedded directly into the subgraphs. Decomposed subgraphs are cached using the original graph handle to enable efficient replay. The CUDA events used in inter-stream synchronization are encoded as dependency edges within each subgraph's internal DAG, which incurs no extra runtime overhead [43]. Composability with model parallelism.Tessera is orthogonal to model parallelism and composes naturally with it. For example, when a model is served with tensor parallelism (TP) across a homogeneous GPU group, Tessera can pair each GPU with a heterogeneous GPU and apply kernel disaggregation within each pair. Collective operations are pinned to the"},{"citing_arxiv_id":"2404.11591","ref_index":19,"ref_count":1,"confidence":0.9,"is_internal_anchor":false,"paper_title":"The EDGE Language: Extended General Einsums for Graph Algorithms","primary_cat":"cs.DS","submitted_at":"2024-04-17T17:42:48+00:00","verdict":null,"verdict_confidence":null,"novelty_score":null,"formal_verification":null,"one_line_summary":null,"context_count":0,"top_context_role":null,"top_context_polarity":null,"context_text":null}],"limit":50,"offset":0}