Event Tensor is a new compiler abstraction for dynamic megakernels that enables high-performance persistent GPU kernels with state-of-the-art LLM serving latency and reduced warmup overhead.
hub
ISBN 9798331314385
13 Pith papers cite this work. Polarity classification is still indexing.
13
Pith papers citing it
hub tools
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
HetCCL enables efficient collective communication across mixed-vendor GPU clusters via P2P transport and a border-communicator mechanism, delivering 17-19x higher bandwidth than Gloo and up to 16.9% faster end-to-end LLM training.
FEPLB reduces token and GEMM stragglers in MoE training by 50-70% using nearly free Copy Engine communication on Hopper architecture.
A Triton-based compiler automatically generates fused distributed kernels with chunk-level compute-communication overlap, reporting 1.3x average and up to 4.7x speedups on multi-GPU workloads.
TokenWeave achieves up to 1.28x lower latency and 1.19x higher throughput for distributed LLM inference by enabling compute-communication overlap at small token counts via a fused AllReduce-RMSNorm kernel that uses only 2-8 SMs.
A cell-wise primary storage for continuous high-order FEM enables exact matrix-free CG iteration on GPUs by confining communication to the preconditioner and realizing DSS via sequential axis face exchanges without gather-scatter or atomics.
HyperParallel-MoE reduces Dispatch-to-Combine MoE-FFN latency by up to 1.58x on Ascend A3 clusters via tile-level heterogeneous scheduling that overlaps communication, matrix, and vector computation inside a single kernel launch.
NanoCP introduces request-level dynamic context parallelism to decouple MoE communication from KV cache placement in hybrid data-expert parallel serving, reporting up to 3.27x higher request rates and 2.12x lower P99 latency under TPOT SLOs.
Perseus removes serialization bottlenecks in multi-node megakernel MoE communication via batched per-destination fences and hardware fence flags, delivering up to 10.3x speedup on proxy transports and matching or exceeding GPU-direct RDMA.
DeepStack introduces a fast performance model and hierarchical search method for co-optimizing 3D DRAM stacking, interconnects, and distributed scheduling in AI accelerators, delivering up to 9.5x throughput gains over baselines.
DMA offloads on AMD MI300X GPUs are extended to latency-bound ML communication using untapped hardware features, closing up to 4.5x performance gap versus RCCL in collectives and delivering up to 1.5x lower latency and 1.9x higher throughput in LLM inference over vLLM.
DynaFlow enables transparent intra-device parallelism in ML systems by separating model definition from execution scheduling, integrating into 6 frameworks with up to 1.29x throughput gains and minimal code changes.
UniEP fuses MoE communication and computation into unified MegaKernels with deterministic token ordering, delivering 1.03x-1.38x speedups over prior work while preserving training accuracy.
citing papers explorer
-
Event Tensor: A Unified Abstraction for Compiling Dynamic Megakernel
Event Tensor is a new compiler abstraction for dynamic megakernels that enables high-performance persistent GPU kernels with state-of-the-art LLM serving latency and reduced warmup overhead.
-
HetCCL: Enabling Collective Communication For Mixed-Vendor Heterogeneous Clusters
HetCCL enables efficient collective communication across mixed-vendor GPU clusters via P2P transport and a border-communicator mechanism, delivering 17-19x higher bandwidth than Gloo and up to 16.9% faster end-to-end LLM training.
-
FEPLB: Exploiting Copy Engines for Nearly Free MoE Load Balancing in Distributed Training
FEPLB reduces token and GEMM stragglers in MoE training by 50-70% using nearly free Copy Engine communication on Hopper architecture.
-
Syncopate: Efficient Multi-GPU AI Kernels via Automatic Chunk-Centric Compute-Communication Overlap
A Triton-based compiler automatically generates fused distributed kernels with chunk-level compute-communication overlap, reporting 1.3x average and up to 4.7x speedups on multi-GPU workloads.
-
TokenWeave: Efficient Compute-Communication Overlap for Distributed LLM Inference
TokenWeave achieves up to 1.28x lower latency and 1.19x higher throughput for distributed LLM inference by enabling compute-communication overlap at small token counts via a fused AllReduce-RMSNorm kernel that uses only 2-8 SMs.
-
Coalesced Matrix-Free Finite Elements in Cell-Wise Storage
A cell-wise primary storage for continuous high-order FEM enables exact matrix-free CG iteration on GPUs by confining communication to the preconditioner and realizing DSS via sequential axis face exchanges without gather-scatter or atomics.
-
HyperParallel-MoE: Multi-Core Interleaved Scheduling for Fast MoE Training on Ascend NPUs
HyperParallel-MoE reduces Dispatch-to-Combine MoE-FFN latency by up to 1.58x on Ascend A3 clusters via tile-level heterogeneous scheduling that overlaps communication, matrix, and vector computation inside a single kernel launch.
-
NanoCP: Request-Level Dynamic Context Parallelism for Data-Expert Parallel Decoding
NanoCP introduces request-level dynamic context parallelism to decouple MoE communication from KV cache placement in hybrid data-expert parallel serving, reporting up to 3.27x higher request rates and 2.12x lower P99 latency under TPOT SLOs.
-
Eliminating Hidden Serialization in Multi-Node Megakernel Communication
Perseus removes serialization bottlenecks in multi-node megakernel MoE communication via batched per-destination fences and hardware fence flags, delivering up to 10.3x speedup on proxy transports and matching or exceeding GPU-direct RDMA.
-
DeepStack: Scalable and Accurate Design Space Exploration for Distributed 3D-Stacked AI Accelerators
DeepStack introduces a fast performance model and hierarchical search method for co-optimizing 3D DRAM stacking, interconnects, and distributed scheduling in AI accelerators, delivering up to 9.5x throughput gains over baselines.
-
DMA-Latte: Expanding the Reach of DMA Offloads to Latency-bound ML Communication
DMA offloads on AMD MI300X GPUs are extended to latency-bound ML communication using untapped hardware features, closing up to 4.5x performance gap versus RCCL in collectives and delivering up to 1.5x lower latency and 1.9x higher throughput in LLM inference over vLLM.
-
DynaFlow: Transparent and Flexible Intra-Device Parallelism via Programmable Operator Scheduling
DynaFlow enables transparent intra-device parallelism in ML systems by separating model definition from execution scheduling, integrating into 6 frameworks with up to 1.29x throughput gains and minimal code changes.
-
UniEP: Unified Expert-Parallel MoE MegaKernel for LLM Training
UniEP fuses MoE communication and computation into unified MegaKernels with deterministic token ordering, delivering 1.03x-1.38x speedups over prior work while preserving training accuracy.