Related papers: Improving Latency in a Signal Processing System on…
With the advent of edge computing, data generated by end devices can be pre-processed before transmission, possibly saving transmission time and energy. On the other hand, data processing itself incurs latency and energy consumption,…
Consider a device that is connected to an edge processor via a communication channel. The device holds local data that is to be offloaded to the edge processor so as to train a machine learning model, e.g., for regression or classification.…
In an effort to lower the barrier to the adoption of FPGAs by a broader community, today major FPGA vendors offer compiler toolchains for OpenCL code. While using these toolchain allows porting existing code to FPGAs, ensuring performance…
As large language models (LLMs) grow in popularity for their diverse capabilities, improving the efficiency of their inference systems has become increasingly critical. Batching LLM requests is a critical step in scheduling the inference…
Recent research on vision backbone architectures has predominantly focused on optimizing efficiency for hardware platforms with high parallel processing capabilities. This category increasingly includes embedded systems such as mobile…
IoT Edge intelligence requires Convolutional Neural Network (CNN) inference to take place in the edge devices itself. ARM big.LITTLE architecture is at the heart of prevalent commercial edge devices. It comprises of single-ISA heterogeneous…
Multicore is an integrated circuit chip that uses two or more computational engines (cores) places in a single processor. This new approach is used to split the computational work of a threaded application and spread it over multiple…
Future autonomous systems require wireless connectivity able to support extremely stringent requirements on both latency and reliability. In this paper, we leverage recent developments in the field of finite-blocklength information theory…
Memory bandwidth is known to be a performance bottleneck for FPGA accelerators, especially when they deal with large multi-dimensional data-sets. A large body of work focuses on reducing of off-chip transfers, but few authors try to improve…
Offloading compute-intensive kernels to hardware accelerators relies on the large degree of parallelism offered by these platforms. However, the effective bandwidth of the memory interface often causes a bottleneck, hindering the…
Processing large-scale graph datasets is computationally intensive and time-consuming. Processor-centric CPU and GPU architectures, commonly used for graph applications, often face bottlenecks caused by extensive data movement between the…
This paper investigates the use of a networked system ($e.g.$, swarm of robots, smart grid, sensor network) to monitor a time-varying phenomenon of interest in the presence of communication and computation latency. Recent advances in edge…
The use of multi-chip modules (MCM) and/or multi-socket boards is the most suitable approach to increase the computation density of servers while keep chip yield attained. This paper introduces a new coherence protocol suitable, in terms of…
This paper introduces a novel, fast atomic-snapshot protocol for asynchronous message-passing systems. In the process of defining what ``fast'' means exactly, we spot a few interesting issues that arise when conventional time metrics are…
It is a challenging task to train large DNN models on sophisticated GPU platforms with diversified interconnect capabilities. Recently, pipelined training has been proposed as an effective approach for improving device utilization. However,…
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…
Large language model (LLM) inference is limited by high computational cost and memory bandwidth demands, making deployment on heterogeneous many-core processors challenging. Taking the MT-3000 processor used in the Tianhe supercomputer as…
This work analyses the effects of sequential-to-parallel synchronization and inter-core communication on multicore performance, speedup and scaling. A modification of Amdahl law is formulated, to reflect the finding that parallel speedup is…
Large-scale language models have become increasingly challenging and expensive to train. Among various methods addressing this issue, Pipeline Parallelism has been widely employed to accommodate massive model weights within limited GPU…
Training large language models (LLMs) is fundamentally constrained by limited device memory and costly inter-device communication. Although pipeline parallelism alleviates memory pressure by partitioning models across devices, it incurs…