EUVIMED · CrossrefZeitschrift
IEEE Computer Architecture Letters
ISSN 1556-6056 · Institute of Electrical and Electronics Engineers (IEEE)
Neueste Artikel
A Taxonomy of SIMDe Emulation Inefficiencies for ARM NEON Porting of VVC Encoders
Split-GEMM: Mitigating Routing-Induced Imbalance for MoE Serving on 3D-Stacked Near-Memory Processing
Application-Directed Hybrid Coherence for Wafer-Scale Supercomputing System
CASCADE: A Memory-Centric Chiplet-Array Accelerator for Energy-Efficient LLM Decoding
Read-Reclaim-Induced Endurance Degradation in High-Bandwidth Flash for LLM Inference
S-ANNS: Energy-Efficient Approximate Nearest Neighbor Search via Shift-Based PIM Architecture
Cross-Layer Head Prediction and KV-Cache Prefetching for High-Performance MoH LLM
Toward Cross-IP Request Coalescing: Relocating the Fan-Out Point in Virtualized I/O
LeakDiT: Diffusion Transformers for Trace-Augmented Side-Channel Analysis
H 3 : H ybrid Architecture Using H igh Bandwidth Memory and H igh Bandwidth Flash for Cost-Efficient LLM Inference
High-Bandwidth Flash for KV Caches: Endurance and Performance Implications
Transformer Training Acceleration through Panel-scale Multi-GPU Integration on Glass Substrates with Optical Interconnects
From Bank to Command: The Shifting Bottleneck for Random-Access Bandwidth in DDR5
Simulation-aided Compute Graph Analysis for Tensor Offloading
HACCL: A Plugin-Based Heterogeneous AI Accelerator Collective Communication Library
FALCON: Fast and Accurate NPU Simulation with Explicit Data Movement
When Kernel Meets System: Chip-to-System Modeling across AI Hardware for LLM Serving
PIMVerse: Enabling Architectural Exploration of Data-Dependent Workloads on NDP Systems
Kernel-Transparent Bandwidth Aggregation in Two-Tier GPU Memory for LLM Serving
HPN-SpGEMM: Hybrid PIM-NMP for SpGEMM
Contention-Aware GPU Thread Block Scheduler for Efficient GPU-SSD
SSD Offloading for LLM Mixture-of-Experts Weights Considered Harmful in Energy Efficiency
CABANA : Cluster-Aware Query Batching for Accelerating Billion-Scale ANNS With Intel AMX
AiDE: Attention-FFN Disaggregated Execution for Cost-Effective LLM Decoding on CXL-PNM
PIMsynth: A Unified Compiler Framework for Bit-Serial Processing-in-Memory Architectures
Accelerating LLM Inference via Dynamic KV Cache Placement in Heterogeneous Memory System
LLMServingSim2.0: A Unified Simulator for Heterogeneous Hardware and Serving Techniques in LLM Infrastructure
RoPIM: A Processing-in-Memory Architecture for Accelerating Rotary Positional Embedding in Transformer Models
A DSP-Based Precision-Scalable MAC With Hybrid Dataflow for Arbitrary-Basis-Quantization CNN Accelerator
Pyramid: Accelerating LLM Inference With Cross-Level Processing-in-Memory
EUVIMED