HPVM

Kotsifakou, Maria; Srivastava, Prakalp; Sinclair, Matthew; Komuravelli, Rakesh; Adve, Vikram; Adve, Sarita V.

doi:10.1145/3178487.3178493

Cited by 35 publications

(6 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…HPVM (Heterogeneous Parallel Virtual Machine) [28] is a compiler infrastructure targeting heterogeneous hardware. It introduces HPVM-C, a programming language for defining data flow graphs (DFGs), directed graphs in which nodes represent a computation task and edges represent inputs and outputs of a computation task.…”

Section: Mobiprox Builds Upon the Existing Work On Heterogeneous And ...mentioning

confidence: 99%

Mobiprox: Supporting Dynamic Approximate Computing on Mobiles

Fabjančič,

Machidon,

Sharif

et al. 2024

IEEE Internet Things J.

View full text Add to dashboard Cite

Runtime-tunable context-dependent network compression would make mobile deep learning (DL) adaptable to often varying resource availability, input "difficulty", or user needs. The existing compression techniques significantly reduce the memory, processing, and energy tax of DL, yet, the resulting models tend to be permanently impaired, sacrificing the inference power for reduced resource usage. The existing tunable compression approaches, on the other hand, require expensive re-training, do not support arbitrary strategies for adapting the compression and do not provide mobile-ready implementations.In this paper we present Mobiprox, a framework enabling mobile DL with flexible precision. Mobiprox implements tunable approximations of tensor operations and enables runtimeadaptable approximation of the individual network layers. A profiler and a tuner included with Mobiprox identify the most promising neural network approximation configurations leading to the desired inference quality with the minimal use of resources. Furthermore, we develop control strategies that depending on contextual factors, such as the input data difficulty, dynamically adjust the approximation levels across a mobile DL model's layers. We implement Mobiprox in Android OS and through experiments in diverse mobile domains, including human activity recognition and spoken keyword detection, demonstrate that it can save up to 15% system-wide energy with a minimal impact on the inference accuracy.

show abstract

Section: Mobiprox Builds Upon the Existing Work On Heterogeneous And ...mentioning

confidence: 99%

Mobiprox: Supporting Dynamic Approximate Computing on Mobiles

Fabjančič,

Machidon,

Sharif

et al. 2024

IEEE Internet Things J.

View full text Add to dashboard Cite

show abstract

“…HPVM (Heterogeneous Parallel Virtual Machine) [22] is a compiler infrastructure targeting heterogeneous hardware. The project introduces HPVM-C, a programming language for defining data flow graphs (DFGs), directed graphs in which nodes represent a computation task and edges represent inputs and outputs of a computation task.…”

Section: Preliminariesmentioning

confidence: 99%

Mobiprox: Supporting Dynamic Approximate Computing on Mobiles

Fabjančič¹,

Machidon²,

Sharif³

et al. 2023

Preprint

View full text Add to dashboard Cite

Runtime-tunable context-dependent network compression would make mobile deep learning adaptable to often varying resource availability, input "difficulty", or user needs. The existing compression techniques significantly reduce the memory, processing, and energy tax of deep learning, yet, the resulting models tend to be permanently impaired, sacrificing the inference power for reduced resource usage. The existing tunable compression approaches, on the other hand, require expensive re-training, seldom provide mobile-ready implementations, and do not support arbitrary strategies for adapting the compression.In this paper we present Mobiprox, a framework enabling flexible-accuracy on-device deep learning. Mobiprox implements tunable approximations of tensor operations and enables runtime adaptation of individual network layers. A profiler and a tuner included with Mobiprox identify the most promising neural network approximation configurations leading to the desired inference quality with the minimal use of resources. Furthermore, we develop control strategies that depending on contextual factors, such as the input data difficulty, dynamically adjust the approximation level of a model. We implement Mobiprox in Android OS and through experiments in diverse mobile domains, including human activity recognition and spoken keyword detection, demonstrate that it can save up to 15% system-wide energy with a minimal impact on the inference accuracy.CCS Concepts: • Human-centered computing → Ubiquitous and mobile computing; • Computing methodologies → Neural networks.

show abstract

“…For example, Google recently proposed LLVM Multi-Level Intermediate Representation (MLIR) [53] to create a customizable compilation framework that provides information at different levels. Similarly, Heterogeneous Parallel Virtual Machine (HPVM) [54] is a representation for a parallel compiler that aims at simplifying the code implementation of parallel hardware. These representations are often combined with domain-specific languages where the designer can abstract specific hardware details.…”

Section: Input Specification and Intermediate Representationmentioning

confidence: 99%

Accelerator Design with High-Level Synthesis

Pilato

Soldavini

2022

Handbook of Computer Architecture

View full text Add to dashboard Cite

Specialized accelerators can exploit spatial parallelism on both operations and data thanks to a dedicated microarchitecture with a better use of the hardware resources. Designers need to describe such components (including the resources, their interconnections, and the control logic) in proper hardware languages compatible with synthesis tools. This process requires hardware design skills that are uncommon in software programmers. To boost the use of spatial accelerators, software programmers need automated methods, like high-level synthesis (HLS), to specify hardware blocks with high-level languages and automatically translate their specifications into the corresponding hardware descriptions ready for synthesis. While HLS is a key enabling technology for the design of complex hardware/software architectures, developing efficient spatial accelerators requires efficient HLS methods to co-optimize performance and hardware cost with a hardware/software co-design approach. In this chapter, we present the current state of the art in high-level synthesis, covering all steps to create the specialized microarchitecture of an accelerator. We also discuss outstanding challenges that can be addressed with the use of HLS.

show abstract

HPVM

Cited by 35 publications

References 22 publications

Mobiprox: Supporting Dynamic Approximate Computing on Mobiles

Mobiprox: Supporting Dynamic Approximate Computing on Mobiles

Mobiprox: Supporting Dynamic Approximate Computing on Mobiles

Accelerator Design with High-Level Synthesis

Contact Info

Product

Resources

About