# StreamDQ: Near-Memory Weight DeQuantization in Custom HBM for Scalable AI Inference Acceleration

arXiv: https://arxiv.org/abs/2607.08993 · HTML: https://mlsanity.com/paper/2607.08993/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in KV cache quantization, one of 12 directions in Efficient LLM inference and serving.
- Builds on 21 of its 61 references that are in the map.
- First posted on arXiv 9 Jul 2026.

## Placed in

- Efficient LLM inference and serving / [KV cache quantization](https://mlsanity.com/inference/kv-cache-quantization/)

## Builds on (most cited first, 21 of 21)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers](https://mlsanity.com/paper/2210.17323/) (2022)
- [AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration](https://mlsanity.com/paper/2306.00978/) (2023)
- [ZeroQuant: Efficient and Affordable Post-Training Quantization for Large-Scale Transformers](https://mlsanity.com/paper/2206.01861/) (2022)
- [QuIP: 2-Bit Quantization of Large Language Models With Guarantees](https://mlsanity.com/paper/2307.13304/) (2023)
- [Atom: Low-bit Quantization for Efficient and Accurate LLM Serving](https://mlsanity.com/paper/2310.19102/) (2023)
- [QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving](https://mlsanity.com/paper/2405.04532/) (2024)
- [Outlier Suppression: Pushing the Limit of Low-bit Transformer Language Models](https://mlsanity.com/paper/2209.13325/) (2022)
- [OliVe: Accelerating Large Language Models via Hardware-friendly Outlier-Victim Pair Quantization](https://mlsanity.com/paper/2304.07493/) (2023)
- [T-MAC: CPU Renaissance via Table Lookup for Low-Bit LLM Deployment on Edge](https://mlsanity.com/paper/2407.00088/) (2024)
- [ANT: Exploiting Adaptive Numerical Data Type for Low-bit Deep Neural Network Quantization](https://mlsanity.com/paper/2208.14286/) (2022)
- [RecNMP: Accelerating Personalized Recommendation with Near-Memory Processing](https://mlsanity.com/paper/1912.12953/) (2019)
- [LUT Tensor Core: A Software-Hardware Co-Design for LUT-Based Low-Bit LLM Inference](https://mlsanity.com/paper/2408.06003/) (2024)
- [Anda: Unlocking Efficient LLM Inference with a Variable-Length Grouped Activation Data Format](https://mlsanity.com/paper/2411.15982/) (2024)
- [Accelerating Bandwidth-Bound Deep Learning Inference with Main-Memory Accelerators](https://mlsanity.com/paper/2012.00158/) (2020)
- [QUICK: Quantization-aware Interleaving and Conflict-free Kernel for efficient LLM inference](https://mlsanity.com/paper/2402.10076/) (2024)
- [MixPE: Quantization and Hardware Co-design for Efficient LLM Inference](https://mlsanity.com/paper/2411.16158/) (2024)
- [Near-Memory Address Translation](https://mlsanity.com/paper/1612.00445/) (2016)

## Abstract (arXiv)

As large language models (LLMs) scale, their memory and computation demands have grown substantially, making weight-only quantization a widely adopted technique for reducing model size with minimal accuracy loss. However, on current GPUs, CUDA-core-based dequantization introduces substantial instruction overhead, on-chip traffic, and pipeline stalls, making it a major bottleneck for high-throughput, cloud-scale LLM serving. To address these limitations, we propose StreamDQ, a lightweight architectural enhancement that enables on-the-fly dequantization in the memory subsystem for high-throughput, large-batch LLM inference. StreamDQ integrates compact DeQuantization Blocks (DQBs) into the base die of high-bandwidth memory (HBM) and performs inline dequantization on standard memory loads. A lightweight sideband tag on each memory read request selects the dequantization mode while preserving conventional load semantics. By relocating dequantization to the memory side, StreamDQ eliminates GPU-side CUDA-core-based dequantization, thereby reducing on-chip traffic on the GPU and avoiding extra HBM write-back and reload of dequantized weights at large batch sizes. Our evaluation shows that StreamDQ achieves up to 7.08× speedup and 90.23% lower energy for mixed-precision GEMM, with only 0.127,mm2 area and 0.355,W power overhead per DQB in a 12,nm CMOS process. For end-to-end LLM inference, StreamDQ reduces latency by up to 54.68% and improves decode throughput by up to 2.20×.