Paper Citation Record · LEDGER

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy

As of 11 August 2026, this Paper Citation Record lists 22 of 22 outbound references and 1 inbound Pith citation observation for arXiv:2511.10909.

A citation records a reference. It does not transfer a finding from one paper to another.

pith.paper-citation-record.v1
2511.10909 v2

Coverage vector

measured 22 of 22 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-05-17T22:56:29.732612Z

measured 23 of 23 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-11T06:34:44.6726+00:00

measured 1 of 1 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-05-15T10:34:16.525398Z

measured 0 of 1 external citation measurements

A source-named dated measurement, never combined with another source.

Source: pith, observed 2026-05-15T10:35:27.465387Z

Reference resolution

22 of 22 outbound references displayed

verified exact7
verified fuzzy6
unresolved0
parse uncertain0
malformed identifier0
metadata mismatch9

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 9496f98e-0eda-478a-91c1-312988b7ec0a · outbound

This paper cites Markidis, S.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy Markidis, S

Reference 1

Resolution: metadata mismatch; arxiv_id, observed 2026-05-17T23:00:25.387050Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation ec433e77-0297-4595-89c3-313c4cb4ac00 · outbound

This paper cites Dissecting Tensor Cores via Microbenchmarks: Latency, Throughput and Numeric Behaviors.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy Dissecting Tensor Cores via Microbenchmarks: Latency, Throughput and Numeric Behaviors

Reference 2

Resolution: verified exact; arxiv_id, observed 2026-05-17T23:00:25.370424Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 3dae3ba7-6bc2-4e08-85ac-1306700a73e4 · outbound

This paper cites LIBRA: Enabling Workload-Aware Multi-Dimensional Network Topology Optimization for Distributed Training of Large AI Models.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy LIBRA: Enabling Workload-Aware Multi-Dimensional Network Topology Optimization for Distributed Training of Large AI Models

Reference 3

Resolution: metadata mismatch; arxiv_id, observed 2026-05-17T23:00:25.364761Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation f20dbfc4-8cce-403b-9f52-6b175cd9fb01 · outbound

This paper cites DeepSeek-V3 Technical Report.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy DeepSeek-V3 Technical Report

Reference 4

Resolution: metadata mismatch; local_arxiv, observed 2026-05-17T23:00:25.408885Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 77eda989-8ba9-46db-8104-5e32607f3316 · outbound

This paper cites PyTorch Developer Notes - Numerical accuracy.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy PyTorch Developer Notes - Numerical accuracy

Reference 5

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T23:02:12.423842Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation b387ed00-4d6f-45c5-a8a5-07adf8db0b50 · outbound

This paper cites Available: https://docs.pytorch.org/docs/stable/notes/ numerical accuracy.html.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy Available: https://docs.pytorch.org/docs/stable/notes/ numerical accuracy.html

Reference 6

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T23:02:12.410875Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation f17ceaea-d66f-4995-90f1-5ffda048c8ce · outbound

This paper cites FTTN: Feature-Targeted Testing for Numerical Properties of NVIDIA & AMD Matrix Accelerators.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy FTTN: Feature-Targeted Testing for Numerical Properties of NVIDIA & AMD Matrix Accelerators

Reference 7

Resolution: metadata mismatch; arxiv_id, observed 2026-05-17T23:00:25.374072Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 3bc1ff37-4eb8-4f1a-a113-98a6b8a6ca39 · outbound

This paper cites Numerical behavior of NVIDIA tensor cores.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy Numerical behavior of NVIDIA tensor cores

Reference 8

Resolution: verified exact; doi, observed 2026-05-17T23:00:25.379864Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation d96c5f6b-e8f9-4dfe-bf43-c9caf97bb66e · outbound

This paper cites FTTN: Feature-Targeted Testing for Numerical Properties of NVIDIA & AMD Matrix Accelerators.

Reference 9

Resolution: metadata mismatch; arxiv_id, observed 2026-05-17T23:00:25.412093Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 6013eae5-aeb1-4042-ae07-da48f499f802 · outbound

This paper cites Revealing Floating-Point Accumulation Orders in Software/Hardware Implementations.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy Revealing Floating-Point Accumulation Orders in Software/Hardware Implementations

Reference 10

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T23:02:12.414291Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 9635bf0f-9c11-4986-aba0-7b9c60841a09 · outbound

This paper cites Estimation of numerical reproducibility on CPU and GPU.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy Estimation of numerical reproducibility on CPU and GPU

Reference 11

Resolution: verified exact; doi, observed 2026-05-17T23:00:25.367021Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 254f7c8d-0a66-4d27-868b-cd63532e3f60 · outbound

This paper cites Expediting Higher Fidelity Plasma State Reconstructions for the DIII-D Na- tional Fusion Facility Using Leadership Class Computing Resources.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy Expediting Higher Fidelity Plasma State Reconstructions for the DIII-D Na- tional Fusion Facility Using Leadership Class Computing Resources

Reference 12

Resolution: metadata mismatch; arxiv_id, observed 2026-05-17T23:00:25.377440Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation c6334b89-898a-488c-9aca-d4931589d8d1 · outbound

This paper cites Problems and Opportunities in Training Deep Learning Software Systems: An Analysis of Variance.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy Problems and Opportunities in Training Deep Learning Software Systems: An Analysis of Variance

Reference 13

Resolution: verified exact; arxiv_id, observed 2026-05-17T23:00:25.383524Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 65995f33-0ff4-423f-81d3-2dabf242e2e7 · outbound

This paper cites In Proceedings of the 44th International Conference on Software Engineering (Pittsburgh, Pennsylvania) (ICSE ’22).

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy In Proceedings of the 44th International Conference on Software Engineering (Pittsburgh, Pennsylvania) (ICSE ’22)

Reference 14

Resolution: metadata mismatch; doi, observed 2026-05-17T23:00:25.405927Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 94d18f1d-28fd-465e-8567-8f5b40fdd180 · outbound

This paper cites Defeating Nondeterminism in LLM Inference.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy Defeating Nondeterminism in LLM Inference

Reference 15

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T23:02:12.416988Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 847a5233-eff6-4793-84db-86af55e61270 · outbound

This paper cites Zheng, C.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy Zheng, C

Reference 16

Resolution: metadata mismatch; arxiv_id, observed 2026-05-17T23:00:25.400134Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 000d57f8-3424-42dc-9b28-a9e9178176c1 · outbound

This paper cites Chechik and T.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy Chechik and T

Reference 17

Resolution: metadata mismatch; arxiv_id, observed 2026-05-17T23:00:25.403491Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 4283967d-7290-4a75-aace-ce4c7aa259d5 · outbound

This paper cites An SMT Formalization of Mixed-Precision Matrix Multiplication: Modeling Three Generations of Tensor Cores.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy An SMT Formalization of Mixed-Precision Matrix Multiplication: Modeling Three Generations of Tensor Cores

Reference 18

Resolution: verified exact; arxiv_id, observed 2026-05-17T23:00:25.391180Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 22e35561-1e17-407b-a472-23487a72fce9 · outbound

This paper cites IEEE Standard for Floating-Point Arithmetic.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy IEEE Standard for Floating-Point Arithmetic

Reference 19

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T23:02:12.420347Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 5a64b25c-92f9-4b5d-935f-c2ca08fe0171 · outbound

This paper cites FP8 Formats for Deep Learning.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy FP8 Formats for Deep Learning

Reference 20

Resolution: verified exact; local_arxiv, observed 2026-05-17T23:00:25.397056Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation f36a9617-34fe-4fa4-9f07-591c5bdf95de · outbound

This paper cites OCP Microscaling Formats (MX) Specifica- tion.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy OCP Microscaling Formats (MX) Specifica- tion

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-05-17T23:02:12.407691Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Observation 68c992b6-97b1-49d1-af8f-993fe058cac8 · outbound

This paper cites Limit results for distribu ted estimation of invariant subspaces in multiple networks inference and pca.

Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy Limit results for distribu ted estimation of invariant subspaces in multiple networks inference and pca

Reference 22

Resolution: verified exact; doi, observed 2026-05-17T23:00:25.393867Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.

Pith citing papers

Observation 61cad61e-84fe-4c1d-890c-1bb8159ac5a1 · inbound

Determinism-Preserving GPU Spatial Sharing with Vitamin-E cites this paper.

Determinism-Preserving GPU Spatial Sharing with Vitamin-E Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy

Reference 75

Resolution: verified exact; local_arxiv, observed 2026-05-15T10:35:27.467353Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-11T06:34:44.6726+00:00.