QUARK: Quantization-Enabled Circuit Sharing for Transformer Acceleration by Exploiting Common Patterns in Nonlinear Operations

Documenti analoghi

• LaMoS: Enabling Efficient Large Number Modular Multiplication through SRAM-based CiM Acceleration
  di: Li, Haomin, et al.
  Pubblicazione: (2025)
• SpecQuant: Spectral Decomposition and Adaptive Truncation for Ultra-Low-Bit LLMs Quantization
  di: Zhao, Zhixiong, et al.
  Pubblicazione: (2025)
• GEMM-GS: Accelerating 3D Gaussian Splatting on Tensor Cores with GEMM-Compatible Blending
  di: Li, Haomin, et al.
  Pubblicazione: (2026)
• ASDR: Exploiting Adaptive Sampling and Data Reuse for CIM-based Instant Neural Rendering
  di: Liu, Fangxin, et al.
  Pubblicazione: (2025)
• ALLMod: Exploring $\underline{\mathbf{A}}$rea-Efficiency of $\underline{\mathbf{L}}$UT-based $\underline{\mathbf{L}}$arge Number $\underline{\mathbf{Mod}}$ular Reduction via Hybrid Workloads
  di: Liu, Fangxin, et al.
  Pubblicazione: (2025)