Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives | Paper Citation Record

Observation 4a9ad259-59ce-4d54-81d1-e51c4da5b342 · outbound

This paper cites Deepseek-v3 technical report,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Deepseek-v3 technical report,

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:13.775104Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0b3788ff-7317-43c7-938d-9aa9caffd13f · outbound

This paper cites Deepseek-v3/r1 671b deployment guide: Gpu require- ments,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Deepseek-v3/r1 671b deployment guide: Gpu require- ments,

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:13.912073Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6e35e9c3-5090-4f82-99a9-a745e6e9b133 · outbound

This paper cites Efficient memory management for large language model serving with pagedattention,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Efficient memory management for large language model serving with pagedattention,

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:14.061863Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation afbebda9-16f1-445a-8503-05689076a4b6 · outbound

This paper cites State of ai: An empirical 100 trillion token study with openrouter,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives State of ai: An empirical 100 trillion token study with openrouter,

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:14.205913Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c8f2e045-189d-482d-a104-95b85da321dd · outbound

This paper cites Sglang: efficient execution of structured language model programs,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Sglang: efficient execution of structured language model programs,

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:14.336839Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f3903cc2-f50f-4753-8ce8-39a453417ea1 · outbound

This paper cites Tensorrt-llm: A library for optimizing large language model inference,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Tensorrt-llm: A library for optimizing large language model inference,

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:14.520829Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1038d302-deab-4ff1-8a98-511433a776a5 · outbound

This paper cites CoreWeave Pricing: Instance Pricing,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives CoreWeave Pricing: Instance Pricing,

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:14.834402Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 09056ba9-bbfb-4762-973b-7222605e7748 · outbound

This paper cites Scaling-up pytorch inference: Serving billions of daily nlp inferences with onnx runtime,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Scaling-up pytorch inference: Serving billions of daily nlp inferences with onnx runtime,

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:15.004162Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e789aa8c-d279-4c85-84b0-cf6a4576cf83 · outbound

This paper cites Understanding data movement in tightly coupled heterogeneous systems: A case study with the grace hopper superchip,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Understanding data movement in tightly coupled heterogeneous systems: A case study with the grace hopper superchip,

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:15.133947Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ba6e4d6b-f423-46cb-a0eb-4d6c06e43fee · outbound

This paper cites Demystifying NCCL: An In-Depth Analysis of GPU Communication Protocols and Algorithms ,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Demystifying NCCL: An In-Depth Analysis of GPU Communication Protocols and Algorithms ,

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:15.320821Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e5220af4-d77e-4356-afdb-85ce704bc8c2 · outbound

This paper cites NVSHMEM communication library,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives NVSHMEM communication library,

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:15.476182Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f7731326-971c-4750-b79f-2deac8501581 · outbound

This paper cites ROCm Communication Collectives Library (RCCL) Documentation,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives ROCm Communication Collectives Library (RCCL) Documentation,

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:15.616175Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation af79158d-7236-4e53-a8a2-d0bf6213a478 · outbound

This paper cites AMD ROCm Documentation, 2026.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives AMD ROCm Documentation, 2026

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:15.792579Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7d36d779-647d-4b94-b44a-7a884e24b51c · outbound

This paper cites Accessed: February 2026.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Accessed: February 2026

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:15.907979Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f4b5d931-844d-4bb2-b6c4-bc1ca39339b5 · outbound

This paper cites an unresolved cited work.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Unresolved cited work

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:16.056139Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f1ca6079-e9d0-4b9a-9fd3-08e37a72c1ad · outbound

This paper cites An introduction to cuda-aware mpi,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives An introduction to cuda-aware mpi,

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:16.165890Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ae8a94c5-8d78-4b20-abc1-adb893da5acd · outbound

This paper cites Enabling fast inference and resilient training with nccl 2.27,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Enabling fast inference and resilient training with nccl 2.27,

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:16.359083Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a34d0e29-9dd8-496f-aa53-0b6ec32830d7 · outbound

This paper cites Gpu-initiated networking for nccl,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Gpu-initiated networking for nccl,

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:16.494336Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 39a171ef-35a5-41c5-aca5-eb4f25696c3f · outbound

This paper cites NVIDIA, 2025.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives NVIDIA, 2025

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:16.661046Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eb7c6918-a749-4e52-9d5b-c23d4e978426 · outbound

This paper cites Fusing communication and compute with new device api and copy engine collectives in nvidia nccl 2.28,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Fusing communication and compute with new device api and copy engine collectives in nvidia nccl 2.28,

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:16.784410Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d9f3656d-35c1-48a9-bc7b-29758dbb8a80 · outbound

This paper cites PyTorch Foundation, 2026.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives PyTorch Foundation, 2026

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:16.857736Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6553156a-7e56-4d0e-a740-6027b92cbe45 · outbound

This paper cites NHR@FAU, 2026.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives NHR@FAU, 2026

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:16.919904Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 472f42f1-784e-4951-96e8-a644b2f1a4ec · outbound

This paper cites PyTorch: An Imperative Style, High-Performance Deep Learning Library.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives PyTorch: An Imperative Style, High-Performance Deep Learning Library

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:16.982682Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1b5388cb-f193-45bd-85b0-93ad50f1a373 · outbound

This paper cites TensorFlow: Large-scale machine learning on heterogeneous systems,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives TensorFlow: Large-scale machine learning on heterogeneous systems,

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:17.085329Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 88c2afe9-aa80-4fc4-a87d-596cccc7a504 · outbound

This paper cites NVIDIA Corporation,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives NVIDIA Corporation,

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:17.160785Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b3a6a6a7-957e-44d1-ae52-321de4b59c6c · outbound

This paper cites NVIDIA Corporation, 2026.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives NVIDIA Corporation, 2026

Reference 28

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:17.331377Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8c77e832-ca44-4a66-9721-44529684ea50 · outbound

This paper cites Llm inference beyond a single node: From bottlenecks to mitigations with fast all-reduce communication,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Llm inference beyond a single node: From bottlenecks to mitigations with fast all-reduce communication,

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:17.416705Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5e9e99dc-cd28-492c-9afa-81c4ee1ec49f · outbound

This paper cites Msccl++: Rethinking gpu communication abstractions for ai inference,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Msccl++: Rethinking gpu communication abstractions for ai inference,

Reference 30

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:17.512162Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 555e25db-a80a-457d-a863-9cd2a6899968 · outbound

This paper cites Nvidia gtc 2025 - built for reasoning, vera rubin, kyber, cpo, dynamo inference, jensen math, feynman,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Nvidia gtc 2025 - built for reasoning, vera rubin, kyber, cpo, dynamo inference, jensen math, feynman,

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:17.600194Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 07f573ca-22b3-4bc6-8073-ee66872fefe8 · outbound

This paper cites Colossus: xai’s supercomputer for grok,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Colossus: xai’s supercomputer for grok,

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:17.688668Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 84ec75fb-c0dc-46ea-83d4-410030bb12c2 · outbound

This paper cites Recent improvement to open mpi allreduce and the impact to application performance,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Recent improvement to open mpi allreduce and the impact to application performance,

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:17.756420Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 749fc5d6-ab5c-4ce0-b39b-d74a13e3be6f · outbound

This paper cites Revisiting the time cost model of allreduce,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Revisiting the time cost model of allreduce,

Reference 34

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:17.859961Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fafed51a-9d45-45a3-a82e-2b0eddf758a9 · outbound

This paper cites xccl: A survey of industry-led collective communication libraries for deep learning,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives xccl: A survey of industry-led collective communication libraries for deep learning,

Reference 35

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:17.951226Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f7d1e527-d487-4679-9914-4503d0286d0f · outbound

This paper cites Hear: Homomorphically encrypted allreduce,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Hear: Homomorphically encrypted allreduce,

Reference 36

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:18.026695Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fac11bdf-0895-49b3-8972-47ca79be7a87 · outbound

This paper cites A survey of mpi usage in the us exascale computing project,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives A survey of mpi usage in the us exascale computing project,

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:18.218838Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7a162cce-669a-4284-ad8d-842c61632797 · outbound

This paper cites A large-scale study of mpi usage in open-source hpc applications,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives A large-scale study of mpi usage in open-source hpc applications,

Reference 38

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:18.463178Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a0ae434e-702d-4a55-9279-4ed928885529 · outbound

This paper cites Short-circuiting rings for low-latency allreduce,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Short-circuiting rings for low-latency allreduce,

Reference 39

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:18.590758Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f41d774e-b592-4c46-a878-9c84b31b3ad8 · outbound

This paper cites Cuda c++ programming guide.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Cuda c++ programming guide

Reference 40

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:18.684480Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0252082f-c101-40bb-ae82-dc161755edab · outbound

This paper cites Parallel thread execution isa version 8.0.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Parallel thread execution isa version 8.0

Reference 41

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:18.772103Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation afb4469f-2c07-4271-a037-cdab3a06269b · outbound

This paper cites vllm container (version 26.02-py3),.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives vllm container (version 26.02-py3),

Reference 42

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:18.845547Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 784214d7-9123-4747-8d89-936f6bdeeace · outbound

This paper cites Network-offloaded bandwidth-optimal broadcast and allgather for distributed ai,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Network-offloaded bandwidth-optimal broadcast and allgather for distributed ai,

Reference 43

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:18.896433Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2f37d146-a83d-4999-ac3a-30f47a9a516a · outbound

This paper cites Comparative analysis of large language model inference serving systems: A performance study of vllm and huggingface tgi,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Comparative analysis of large language model inference serving systems: A performance study of vllm and huggingface tgi,

Reference 44

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:19.083488Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8701c828-c2be-4774-8eba-d71b78d59efd · outbound

This paper cites vllm – technology radar entry,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives vllm – technology radar entry,

Reference 45

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:19.181539Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 323375ca-c4ab-487e-ba67-eee2cd45c0fe · outbound

This paper cites The huge potential implications of long-context inference,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives The huge potential implications of long-context inference,

Reference 46

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:19.273592Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0ffa4cbf-30c7-47a0-80ab-8881c503fa2b · outbound

This paper cites Hilfer fractional advection-diffusion equations with power-law initial condition; a Numerical study using variational iteration method.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Hilfer fractional advection-diffusion equations with power-law initial condition; a Numerical study using variational iteration method

Reference 47

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:19.006976Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c1127a6f-4641-49b7-aa98-7e64e2d365ea · outbound

This paper cites Chain of agents: Large language models collaborating on long-context tasks,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Chain of agents: Large language models collaborating on long-context tasks,

Reference 48

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:19.440969Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9748d69e-2b66-4ed0-b867-61dc58a6b85b · outbound

This paper cites Breaking the boundaries of long- context llm inference: Adaptive kv management on a single commodity gpu,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Breaking the boundaries of long- context llm inference: Adaptive kv management on a single commodity gpu,

Reference 49

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:19.528704Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1d5f83da-1bb3-438b-abbe-41a0e30dcfb5 · outbound

This paper cites Prefill-decode disaggregation.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Prefill-decode disaggregation

Reference 50

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:19.629418Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8aafc676-e597-4ac0-ac20-bcf8c7c6341d · outbound

This paper cites Is long context all you need? leveraging llm’s extended context for nl2sql,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Is long context all you need? leveraging llm’s extended context for nl2sql,

Reference 51

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:19.382129Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7e090cab-d299-4f45-92fe-6193df7732e8 · outbound

This paper cites an unresolved cited work.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Unresolved cited work

Reference 52

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:19.766032Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 874e5964-549e-489c-a52b-7356e2e0dc95 · outbound

This paper cites Accessed: March 2026.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Accessed: March 2026

Reference 53

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:19.916482Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4f1b716c-61c9-489b-b789-8d1fbb58b7ae · outbound

This paper cites Amrex and pyamrex: Looking beyond the exas- cale computing project,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Amrex and pyamrex: Looking beyond the exas- cale computing project,

Reference 54

Resolution: verified exact; doi, observed 2026-08-01T21:28:32.655939Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.

Observation 91ae64af-c590-47a8-b86f-c6d071eaf952 · outbound

This paper cites Slo-aware compute resource allocation for prefill-decode disaggregated llm inference,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Slo-aware compute resource allocation for prefill-decode disaggregated llm inference,

Reference 55

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:19.670034Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2a329e9e-4041-45be-90bd-4bc80afa968c · outbound

This paper cites Qmcpack: an open source ab initio quantum monte carlo package for the electronic structure of atoms, molecules and solids,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Qmcpack: an open source ab initio quantum monte carlo package for the electronic structure of atoms, molecules and solids,

Reference 56

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.252829Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d3c997e9-79f3-4624-b9f7-6a224d926070 · outbound

This paper cites an unresolved cited work.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Unresolved cited work

Reference 57

Resolution: parse uncertain; no resolver link, observed 2026-08-01T21:25:19.821885Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6005d775-7c55-4096-9985-7542cd1c6bc9 · outbound

This paper cites NVIDIA HPCG Benchmark,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives NVIDIA HPCG Benchmark,

Reference 58

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.385515Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7a2f491e-e59a-479f-8ee1-c644e10c9fc3 · outbound

This paper cites Accessed: March 2026.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Accessed: March 2026

Reference 59

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.464831Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation be738fcb-82d0-463c-96d2-c14d401588b6 · outbound

This paper cites AMReX issue #4821: Device-initiated collectives in NCCL/RCCL.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives AMReX issue #4821: Device-initiated collectives in NCCL/RCCL

Reference 60

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.077961Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4702eab3-4a03-4345-9f35-ca16d8197c20 · outbound

This paper cites New research infrastructure: ’alps’ supercomputer inaugurated,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives New research infrastructure: ’alps’ supercomputer inaugurated,

Reference 61

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.626672Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1f8ab145-110b-45b5-b5cf-879508324856 · outbound

This paper cites Flashinfer: Efficient and customizable attention engine for llm inference serving,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Flashinfer: Efficient and customizable attention engine for llm inference serving,

Reference 62

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.731027Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cd35d561-5800-452b-ba99-0071dc64e305 · outbound

This paper cites Qmcpack issue #4654,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Qmcpack issue #4654,

Reference 63

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.332056Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 90e64a8e-5e91-4298-a3e9-ab2d51c5fe27 · outbound

This paper cites Nccl ep: Towards a unified expert parallel communication api for nccl,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Nccl ep: Towards a unified expert parallel communication api for nccl,

Reference 64

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.853881Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 02d42690-5e33-42c5-80e1-106ff10d1444 · outbound

This paper cites Enhancing dis- tributed inference performance with the nvidia inference transfer library,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Enhancing dis- tributed inference performance with the nvidia inference transfer library,

Reference 65

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.934999Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 29b8628c-a227-432b-84c2-a7050299050e · outbound

This paper cites The elpa library: scal- able parallel eigenvalue solutions for electronic structure theory and computational science,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives The elpa library: scal- able parallel eigenvalue solutions for electronic structure theory and computational science,

Reference 66

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.549073Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0836cd97-20e2-42c9-abfe-54893ca7b307 · outbound

This paper cites Deepep: an efficient expert-parallel communication library,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Deepep: an efficient expert-parallel communication library,

Reference 69

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.804322Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation aa173ed3-ed6f-40fd-bc13-383abd1e9ad4 · outbound

This paper cites Gpt-5.4 chatbot,.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Gpt-5.4 chatbot,

Reference 72

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.987873Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ae2cf742-ceec-41a6-b499-f3cdccdb1871 · outbound

This paper cites an unresolved cited work.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Unresolved cited work

Reference 2020

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:18.335994Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6947ed3d-468a-4bd4-b977-0ba2f5b4af63 · outbound

This paper cites an unresolved cited work.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Unresolved cited work

Reference 2023

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:18.096674Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f0c5898d-7da4-438e-8d51-52d3a6726d18 · outbound

This paper cites an unresolved cited work.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Unresolved cited work

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:17.236864Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 78451cc0-945c-465e-902a-fe41b4d093f9 · outbound

This paper cites an unresolved cited work.

Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives Unresolved cited work

Reference 2025

Resolution: unresolved; no resolver link, observed 2026-08-01T21:25:20.176773Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Coverage vector

Reference resolution

Outbound references

Pith citing papers