20 results for βGPU parallelismβ
CS papers onlyHybrid search: Keyword + semantic, ranked by combined score.β
Want pure semantic search? Try claim verification β
Tangram enables the use of existing LLM parallelizers in heterogeneous GPU clusters by decoupling parallelization planning from GPU heterogeneity.
This paper presents MemX, a memory-aware automatic parallelism system for large-scale model training on cross-vendor heterogeneous GPU clusters.
Xinwei Qiang, Yifan Hu, Shixuan Sun, Jing Yang +5 more
GF-DiT: A policy-programmable runtime for elastic DiT serving
Rania Zitouni, Nadine Bousdjira, Sarah Hasnaoui, Amel Sadoun +1 more
This paper compares and optimizes CUDA strategies for a shallow neural network, achieving a 1.41x speedup on a large dataset.
Guangyuan Xu, Lanxuan Zhang, Ziyang Li, Zhaoshi Li +1 more
This paper presents TileDRAM, a compiler-runtime toolchain that introduces tile-to-DRAM layouts as a compiler-visible abstraction for DRAM placement.
The paper introduces $g$MAGNUS, a new algorithm for sparse matrix-matrix multiplication on GPUs that addresses heavy rows by reordering intermediate products and achieves significant speedups.
This paper introduces DGNA, a methodology to unveil the Non-Uniform Memory Access (NUMA) architecture of GPU memory hierarchy through microbenchmarking and data analysis.
This paper reformulates cascaded second-order filtering as a block-tridiagonal linear system and develops parallel solution algorithms, achieving high performance on SIMD cores, multi-core CPUs, and Gβ¦
Zhihao Xu, Hao Zhong, Zeting Zhou, Yuhang Xu +8 more
This paper proposes Gleam, a framework for efficient GPU sharing across local-area CUDA devices, reducing bandwidth overhead, improving API call latency, and ensuring context consistency.
This paper analyzes data access patterns in large language model workloads on multi-partition GPUs and presents optimization strategies based on inter-workgroup sharing patterns.
Daifeng Li, Huiqiang Jiang, Chengruidong Zhang, Wei Wu +5 more
The paper introduces D2K-Bench, a diagnostic benchmark to measure how effectively large language model agents translate expert design guidance into efficient GPU kernels.
This paper identifies suboptimalities in default GPU-first scheduling for AI tool workloads and proposes an agentic scheduler that adaptively assigns tools to GPU or CPU based on runtime factors.
The paper introduces parRSB, a parallel graph partitioner for spectral element meshes that produces high-quality partitions.
First implementation of a 3D Gaussian renderer on an Intelligence Processing Unit (IPU) with 1,472 independent tiles using only on-chip SRAM.