COMET: A Framework for Modeling Compound Operation Dataflows with Explicit Collectives
August 30, 2025 Β· Declared Dead Β· π arXiv.org
"No code URL or promise found in abstract"
Evidence collected by the PWNC Scanner
Authors
Shubham Negi, Manik Singhal, Aayush Ankit, Sudeep Bhoja, Kaushik Roy
arXiv ID
2509.00599
Category
cs.AR: Hardware Architecture
Cross-listed
cs.DC
Citations
2
Venue
arXiv.org
Last Checked
3 months ago
Abstract
Modern machine learning accelerators are designed to efficiently execute deep neural networks (DNNs) by optimizing data movement, memory hierarchy, and compute throughput. However, emerging DNN models such as large language models, state space models increasingly rely on compound operations-structured compositions of multiple basic operations-which introduce new challenges for dataflow optimization and minimizing off-chip memory traffic. Moreover, as model size continues to grow, deployment across spatially distributed compute clusters becomes essential, requiring frequent and complex collective communication. Existing dataflow optimization frameworks and performance models either focus on single operations or lack explicit modeling of collective communication cost, limiting their applicability to modern workloads. To address these limitations, we propose, a framework for modeling and optimizing dataflow for compound operations on machine learning accelerators. COMET introduces a novel representation that explicitly models collective communication across spatial clusters, along with latency and energy cost models that account for both GEMM and non-GEMM operation level dependencies within compound operations. We demonstrate COMET's capabilities to analyze and optimize dataflows for compound operations such as GEMM--Softmax, GEMM--LayerNorm, and self-attention, across both edge and cloud accelerator configurations. Our collective-aware modeling enables exploration of a broader mapping space, leading to improved performance and energy efficiency. Specifically, our optimized dataflows achieve up to 1.42$\times$ speedup for GEMM-Softmax, 3.46$\times$ for GEMM-LayerNorm and 1.82$\times$ for self-attention compared to unfused baselines.
Community Contributions
Found the code? Know the venue? Think something is wrong? Let us know!
π Similar Papers
In the same crypt β Hardware Architecture
R.I.P.
π»
Ghosted
R.I.P.
π»
Ghosted
Corona: System Implications of Emerging Nanophotonic Technology
R.I.P.
π»
Ghosted
A scalable multi-core architecture with heterogeneous memory structures for Dynamic Neuromorphic Asynchronous Processors (DYNAPs)
R.I.P.
π»
Ghosted
SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning
R.I.P.
π»
Ghosted
Neural Cache: Bit-Serial In-Cache Acceleration of Deep Neural Networks
R.I.P.
π»
Ghosted
SpArch: Efficient Architecture for Sparse Matrix Multiplication
Died the same way β π» Ghosted
R.I.P.
π»
Ghosted
Federated Learning: Strategies for Improving Communication Efficiency
R.I.P.
π»
Ghosted
Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning
R.I.P.
π»
Ghosted
Explanation in Artificial Intelligence: Insights from the Social Sciences
R.I.P.
π»
Ghosted