Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2409.04992 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910594814181376 |
|---|---|
| author | Pan, Xiurui Li, Endian Li, Qiao Liang, Shengwen Shan, Yizhou Zhou, Ke Luo, Yingwei Wang, Xiaolin Zhang, Jie |
| author_facet | Pan, Xiurui Li, Endian Li, Qiao Liang, Shengwen Shan, Yizhou Zhou, Ke Luo, Yingwei Wang, Xiaolin Zhang, Jie |
| contents | The widespread of Large Language Models (LLMs) marks a significant milestone in generative AI. Nevertheless, the increasing context length and batch size in offline LLM inference escalate the memory requirement of the key-value (KV) cache, which imposes a huge burden on the GPU VRAM, especially for resource-constraint scenarios (e.g., edge computing and personal devices). Several cost-effective solutions leverage host memory or SSDs to reduce storage costs for offline inference scenarios and improve the throughput. Nevertheless, they suffer from significant performance penalties imposed by intensive KV cache accesses due to limited PCIe bandwidth. To address these issues, we propose InstInfer, a novel LLM inference system that offloads the most performance-critical computation (i.e., attention in decoding phase) and data (i.e., KV cache) parts to Computational Storage Drives (CSDs), which minimize the enormous KV transfer overheads. InstInfer designs a dedicated flash-aware in-storage attention engine with KV cache management mechanisms to exploit the high internal bandwidths of CSDs instead of being limited by the PCIe bandwidth. The optimized P2P transmission between GPU and CSDs further reduces data migration overheads. Experimental results demonstrate that for a 13B model using an NVIDIA A6000 GPU, InstInfer improves throughput for long-sequence inference by up to 11.1$\times$, compared to existing SSD-based solutions such as FlexGen. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_04992 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | InstInfer: In-Storage Attention Offloading for Cost-Effective Long-Context LLM Inference Pan, Xiurui Li, Endian Li, Qiao Liang, Shengwen Shan, Yizhou Zhou, Ke Luo, Yingwei Wang, Xiaolin Zhang, Jie Hardware Architecture Computation and Language The widespread of Large Language Models (LLMs) marks a significant milestone in generative AI. Nevertheless, the increasing context length and batch size in offline LLM inference escalate the memory requirement of the key-value (KV) cache, which imposes a huge burden on the GPU VRAM, especially for resource-constraint scenarios (e.g., edge computing and personal devices). Several cost-effective solutions leverage host memory or SSDs to reduce storage costs for offline inference scenarios and improve the throughput. Nevertheless, they suffer from significant performance penalties imposed by intensive KV cache accesses due to limited PCIe bandwidth. To address these issues, we propose InstInfer, a novel LLM inference system that offloads the most performance-critical computation (i.e., attention in decoding phase) and data (i.e., KV cache) parts to Computational Storage Drives (CSDs), which minimize the enormous KV transfer overheads. InstInfer designs a dedicated flash-aware in-storage attention engine with KV cache management mechanisms to exploit the high internal bandwidths of CSDs instead of being limited by the PCIe bandwidth. The optimized P2P transmission between GPU and CSDs further reduces data migration overheads. Experimental results demonstrate that for a 13B model using an NVIDIA A6000 GPU, InstInfer improves throughput for long-sequence inference by up to 11.1$\times$, compared to existing SSD-based solutions such as FlexGen. |
| title | InstInfer: In-Storage Attention Offloading for Cost-Effective Long-Context LLM Inference |
| topic | Hardware Architecture Computation and Language |
| url | https://arxiv.org/abs/2409.04992 |