ELMoE-3D: Leveraging Intrinsic Elasticity of MoE for Hybrid-Bonding-Enabled Self-Speculative Decoding in On-Premises Serving

• URL: http://arxiv.org/abs/2604.14626v2
• Date: Thu, 23 Apr 2026 01:19:26 GMT
• Title: ELMoE-3D: Leveraging Intrinsic Elasticity of MoE for Hybrid-Bonding-Enabled Self-Speculative Decoding in On-Premises Serving
• Abstract summary: We propose ELMoE-3D, a hybrid-bonding framework that unifies cache-based acceleration and speculative decoding.<n>On our 3D-stacked hardware, ELMoE-3D achieves an average $6.6times$ speedup and $4.4times$ energy efficiency gain over naive MoE serving on xPU 1-16.
• Score: 3.8457393423256363
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Mixture-of-Experts (MoE) models have become the dominant architecture for large-scale language models, yet on-premises serving remains fundamentally memory-bound as batching turns sparse per-token compute into dense memory activation. Memory-centric architectures (PIM, NMP) improve bandwidth but leave compute underutilized under MoE's low arithmetic intensity at high batch sizes. Speculative decoding (SD) trades idle compute for fewer target invocations, yet verification must load experts even for rejected tokens, severely limiting its benefit in MoE especially at low batch sizes. We propose ELMoE-3D, a hybrid-bonding (HB)-based HW-SW co-designed framework that unifies cache-based acceleration and speculative decoding to offer overall speedup across batch sizes. We identify two intrinsic elasticity axes of MoE-expert and bit-and jointly scale them to construct Elastic Self-Speculative Decoding (Elastic-SD), which serves as both an expert cache and a strongly aligned self-draft model accelerated by high HB bandwidth. Our LSB-augmented bit-sliced architecture exploits inherent redundancy in bit-slice representations to natively support bit-nested execution. On our 3D-stacked hardware, ELMoE-3D achieves an average $6.6\times$ speedup and $4.4\times$ energy efficiency gain over naive MoE serving on xPU across batch sizes 1-16, and delivers $2.2\times$ speedup and $1.4\times$ energy efficiency gain over the best-performing prior accelerator baseline.

Related papers

• MoBiE: Efficient Inference of Mixture of Binary Experts under Post-Training Quantization [11.19613037505662]
  MoBiE is a binarization framework tailored for Mixture-of-Experts (MoE) based large language models (LLMs)<n>MoBiE consistently outperforms state-of-the-art binary methods across multiple MoE-based LLMs and benchmarks.
  arXiv  Detail & Related papers  (2026-04-08T08:12:26Z)
• ZipMoE: Efficient On-Device MoE Serving via Lossless Compression and Cache-Affinity Scheduling [56.88966608455977]
  ZipMoE exploits the synergy between the hardware properties of edge devices and the statistical redundancy inherent to MoE parameters.<n>ZipMoE achieves up to $72.77%$ inference latency reduction and up to $6.76times$ higher throughput than the state-of-the-art systems.
  arXiv  Detail & Related papers  (2026-01-29T02:51:59Z)
• MoEBlaze: Breaking the Memory Wall for Efficient MoE Training on Modern GPUs [9.086910335841772]
  "Memory wall" bottleneck is significantly amplified in modern large-scale Mixture-of-Experts (MoE) architectures.<n>We present MoEBlaze, a memory-efficient MoE training framework that addresses these issues through a co-designed system approach.<n>We demonstrate that MoEBlaze can achieve over 4x speedups and over 50% memory savings compared to existing MoE frameworks.
  arXiv  Detail & Related papers  (2026-01-08T08:38:23Z)
• SonicMoE: Accelerating MoE with IO and Tile-aware Optimizations [54.303301888915406]
  Mixture of Experts (MoE) models have emerged as the de facto architecture for scaling up language models without significantly increasing the computational cost.<n>We propose a memory-efficient algorithm to compute the forward and backward passes of MoEs with minimal activation caching.<n>We also propose a novel "token rounding" method that minimizes the wasted compute due to padding in Grouped GEMM kernels.
  arXiv  Detail & Related papers  (2025-12-16T04:39:10Z)
• Decoder-Hybrid-Decoder Architecture for Efficient Reasoning with Long Generation [108.0657508755532]
  We introduce the Gated Memory Unit (GMU), a simple yet effective mechanism for efficient memory sharing across layers.<n>We apply it to create SambaY, a decoder-hybrid-decoder architecture that incorporates GMUs to share memory readout states from a Samba-based self-decoder.
  arXiv  Detail & Related papers  (2025-07-09T07:27:00Z)
• HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference [5.015541720729724]
  HybriMoE is a hybrid CPU-GPU inference framework that improves resource utilization through a novel CPU-GPU scheduling and cache management system.<n>We implement HybriMoE on top of the kTransformers framework and evaluate it on three widely used MoE-based LLMs.
  arXiv  Detail & Related papers  (2025-04-08T10:47:37Z)
• HOBBIT: A Mixed Precision Expert Offloading System for Fast MoE Inference [54.40808356999408]
  We present HOBBIT, a mixed precision expert offloading system to enable flexible and efficient MoE inference. Our key insight is that dynamically replacing less critical cache-miss experts with low precision versions can substantially reduce expert-loading latency. HOBBIT achieves up to a 9.93x speedup in decoding compared to state-of-the-art MoE offloading systems.
  arXiv  Detail & Related papers  (2024-11-03T04:25:46Z)
• Fast Matrix Multiplications for Lookup Table-Quantized LLMs [58.11584672945781]
  FLUTE is a flexible lookup table engine for LUT-quantized LLMs.<n>At batch sizes 32 and quantization group size of 128, the FLUTE kernel can be 2-4x faster than existing GEMM kernels.
  arXiv  Detail & Related papers  (2024-07-15T17:55:42Z)
• Flash-LLM: Enabling Cost-Effective and Highly-Efficient Large Generative Model Inference with Unstructured Sparsity [12.663030430488922]
  We propose Flash-LLM for enabling low-cost and highly-efficient large generative model inference on high-performance Cores. At SpMM kernel level, Flash-LLM significantly outperforms the state-of-the-art library, i.e., Sputnik and SparTA by an average of 2.9x and 1.5x, respectively.
  arXiv  Detail & Related papers  (2023-09-19T03:20:02Z)
• SqueezeLLM: Dense-and-Sparse Quantization [80.32162537942138]
  Main bottleneck for generative inference with LLMs is memory bandwidth, rather than compute, for single batch inference. We introduce SqueezeLLM, a post-training quantization framework that enables lossless compression to ultra-low precisions of up to 3-bit. Our framework incorporates two novel ideas: (i) sensitivity-based non-uniform quantization, which searches for the optimal bit precision assignment based on second-order information; and (ii) the Dense-and-Sparse decomposition that stores outliers and sensitive weight values in an efficient sparse format.
  arXiv  Detail & Related papers  (2023-06-13T08:57:54Z)
• Towards MoE Deployment: Mitigating Inefficiencies in Mixture-of-Expert (MoE) Inference [7.743308058511418]
  We provide a characterization of two MoE workloads, namely Language Modeling (LM) and Machine Translation (MT) We propose three optimization techniques to mitigate sources of inefficiencies, namely (1) Dynamic gating, (2) Expert Buffering, and (3) Expert load balancing.
  arXiv  Detail & Related papers  (2023-03-10T19:30:15Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.