Related papers: Beluga: A CXL-Based Memory Architecture for Scalab…

Key-value (KV) cache memory management is the primary bottleneck limiting throughput and cost-efficiency in large-scale GPU inference serving. Current systems suffer from three compounding inefficiencies: (1) the absence of unified KV cache…

Large Language Models (LLMs) exhibit pronounced memory-bound characteristics during inference due to High Bandwidth Memory (HBM) bandwidth constraints. In this paper, we propose an L2 Cache-oriented asynchronous KV Cache prefetching method…

The growing demands in the training and inference of Large Language Models (LLMs) are accelerating the adoption of scale-up systems that extend server shared memory through the use of Compute Express Link (CXL)-based load/store…

In Large Language Model (LLM) inference, Key-Value (KV) caches (KV-caches) are essential for reducing time complexity. However, they result in a linear increase in GPU memory as the context length grows. While recent work explores KV-cache…

Tensor parallelism (TP) enables large language models (LLMs) to scale inference efficiently across multiple GPUs, but its tight coupling makes systems fragile: a single GPU failure can halt execution, trigger costly KVCache recomputation,…

The memory system is a major performance determinant for server processors. Ever-growing core counts and datasets demand higher bandwidth and capacity as well as lower latency from the memory system. To keep up with growing demands,…

The attention layer, a core component of Transformer-based LLMs, brings out inefficiencies in current GPU systems due to its low operational intensity and the substantial memory requirements of KV caches. We propose a High-bandwidth…

Latent Dirichlet Allocation(LDA) is a popular topic model. Given the fact that the input corpus of LDA algorithms consists of millions to billions of tokens, the LDA training process is very time-consuming, which may prevent the usage of…

During LLM inference, KVCache memory usage grows linearly with sequence length and batch size and often exceeds GPU capacity. Recent proposals offload KV states to host memory and reduce transfers using top-k attention. But their…

The increasing adoption of large language models (LLMs) on heterogeneous computing platforms poses significant challenges to achieving high inference efficiency. To address these efficiency bottlenecks across diverse platforms, this paper…

Large Language Models (LLMs) are widely used across various domains, processing millions of daily requests. This surge in demand poses significant challenges in optimizing throughput and latency while keeping costs manageable. The Key-Value…

Distributed prefix caching has become a core technique for efficient LLM serving. However, for long-context requests with high cache hit ratios, retrieving reusable KVCache blocks from remote servers has emerged as a new performance…

Training large language models (LLMs) in the cloud faces growing memory bottlenecks due to the limited capacity and high cost of GPUs. While GPU memory offloading to CPU and NVMe has made large-scale training more feasible, existing…

Inference on large-language models (LLMs) is constrained by GPU memory capacity. A sudden increase in the number of inference requests to a cloud-hosted LLM can deplete GPU memory, leading to contention between multiple prompts for limited…

In GPU graph analytics, the use of external memory such as the host DRAM and solid-state drives is a cost-effective approach to processing large graphs beyond the capacity of the GPU onboard memory. This paper studies the use of Compute…

Hardware based memory pooling enabled by interconnect standards like CXL have been gaining popularity amongst cloud providers and system integrators. While pooling memory resources has cost benefits, it comes at a penalty of increased…

The rapid development of large language models (LLM) has greatly enhanced everyday applications. While many FPGA-based accelerators, with flexibility for fine-grained data control, exhibit superior speed and energy efficiency compared to…

The substantial memory requirements of Large Language Models (LLMs), particularly for long-context fine-tuning, have renewed interest in CPU offloading to augment limited GPU memory. However, as context lengths grow, relying on CPU memory…

Deep learning recommendation models (DLRMs) have been widely applied in Internet companies. The embedding tables of DLRMs are too large to fit on GPU memory entirely. We propose a GPU-based software cache approaches to dynamically manage…

Cost of serving large language models (LLM) is high, but the expensive and scarce GPUs are poorly efficient when generating tokens sequentially, unless the batch of sequences is enlarged. However, the batch size is limited by some…