A new partitioning algorithm that provably load-balances arbitrary sparse tensor algebra expressions by generalizing parallel merging to multi-operand, multi-dimensional hierarchical structures, implemented in a compiler framework.
Accelerating sparse data orchestration via dynamic reflexive tiling
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
citation-role summary
background 2
method 1
citation-polarity summary
representative citing papers
Proxics introduces lightweight virtual processors and low-latency communication channels as portable OS abstractions for programming near-data processing accelerators, demonstrated on real hardware for memory-intensive workloads.
Mambalaya delivers 4.9x prefill and 1.9x generation speedups on Mamba layers over prior accelerators by systematically fusing inter-Einsum operations.
CCCL delivers 1.34-1.94x faster cross-node GPU collectives via CXL memory pooling than 200 Gbps InfiniBand RDMA, with 1.11x LLM training speedup and 2.75x hardware cost reduction.
Equilibria delivers per-container fairness controls and observability for CXL memory tiering, improving production workload performance by up to 52% over Linux TPP while suppressing noisy-neighbor interference.
PRISM introduces a probabilistic performance modeling framework that quantifies guarantees on training time for large-scale distributed systems under runtime variability.
citing papers explorer
-
Partitioning Unstructured Sparse Tensor Algebra for Load-Balanced Parallel Execution
A new partitioning algorithm that provably load-balances arbitrary sparse tensor algebra expressions by generalizing parallel merging to multi-operand, multi-dimensional hierarchical structures, implemented in a compiler framework.
-
Proxics: an efficient programming model for far memory accelerators
Proxics introduces lightweight virtual processors and low-latency communication channels as portable OS abstractions for programming near-data processing accelerators, demonstrated on real hardware for memory-intensive workloads.
-
Mambalaya: Einsum-Based Fusion Optimizations on State-Space Models
Mambalaya delivers 4.9x prefill and 1.9x generation speedups on Mamba layers over prior accelerators by systematically fusing inter-Einsum operations.
-
CCCL: Node-Spanning GPU Collectives with CXL Memory Pooling
CCCL delivers 1.34-1.94x faster cross-node GPU collectives via CXL memory pooling than 200 Gbps InfiniBand RDMA, with 1.11x LLM training speedup and 2.75x hardware cost reduction.
-
Equilibria: Fair Multi-Tenant CXL Memory Tiering At Scale
Equilibria delivers per-container fairness controls and observability for CXL memory tiering, improving production workload performance by up to 52% over Linux TPP while suppressing noisy-neighbor interference.
-
PRISM: Probabilistic Runtime Insights and Scalable Performance Modeling for Large-Scale Distributed Training
PRISM introduces a probabilistic performance modeling framework that quantifies guarantees on training time for large-scale distributed systems under runtime variability.
- The EDGE Language: Extended General Einsums for Graph Algorithms