CCF Transactions on High Performance Computing
Издатель: Springer Nature · ISSN 2524-4922,2524-4930 · 97 статей
- Bridging OpenCL kernels and pre-compiled libraries via runtime invocation
- Parallel optimization of sea ice model dynamics modules for the SW39000 heterogeneous many-core processor
- Geometry-aware PU-RBF mapping for accurate parallel multiphysics coupling
- Approximation algorithm for the two-depot parallel-machine customer order scheduling with delivery time and submodular rejection penalties
- Q-VoI: observable-aware decoder refinement under a bounded classical compute budget
- Dpaccel: accelerating the training of machine learning interatomic potentials with efficient data loading and distributed memory management
- Spatial-index-preserving scientific grid inference with multi-stage resource-aware batch scheduling
- 3CBench: a unified benchmarking framework for the computing capacity of heterogeneous AI clusters
- PerfInliner: a cost model guided machine learning inlining framework to optimize performance
- Memory–compute co-optimization for stencil computation on multi-core DSPs
- Research on the porting and optimization of the WRF dynamic framework for domestic DCU heterogeneous supercomputing platforms
- Superconducting digital integrated circuits designs and architectures: overview and outlook
- FusionTransmit: a compression-aware distributed training scheduler for PowerSGD
- Optimization of LOBPCG solver for PDE eigenvalue problems with high-order/spectral FEM on GPU
- Porting LULESH to SYCL: practical insights and performance analysis
- Localityshift ECC: spatial-aware redundancy distribution for near-zero overhead GPU memory protection
- Cladder: achieving practical and efficient stencil computation on tensor core units
- Predicting propagation effects of tweets via a multimodal feature fusion model based on co-attention mechanism
- High-performance matrix multiplication micro-kernel generation for GPDSPs via MLIR progressive lowering
- Exploiting intra-core heterogeneity for high-performance dense linear algebra on Huawei Ascend 910 NPUs
- Computation–communication overlapping based on vertical layer grouping in the inverse legendre transform stage of YHGSM
- LEGO-compiler: enhancing neural compilation through translation composability
- Directive-based automatic function-level vectorization for simplified SIMD exploitation
- Parallelsim: an accurate, generic, and efficient simulator for distributed deep learning
- A representative time steps selection method for time-varying scientific simulation dataset based on parametric distribution models
-
Optimization of the ParILUT-GPU algorithm
Abstract We have optimized the parallel threshold ILU algorithm (ParILUT) for GPUs. The optimizations are for three building blocks: candidate search and ILU residual computation, adding and removing elements, and threshold selection....
-
A survey of anomaly detection in HPC systems using machine learning
Abstract High-performance computing (HPC) systems must remain stable and reliable to consistently deliver robust computational power and ensure the proper execution of user jobs. Anomaly detection is a key means to ensure the stability...
- The new compiler stack: a survey on the synergy of LLMs and compilers
-
Review and analysis of performance prediction methods and tools for heterogeneous parallel programs
Abstract With the increasing number of computationally intensive applications, heterogeneous systems have become an important solution for improving computing performance. In order to effectively develop and optimize parallel programs...
- Parallel implementation and optimization of LMS adaptive filtering algorithms based on vector DSP