An official website of the United States government
In view of the performance limitations of fully-decoupled designs for neural architectures and accelerators, hardware-software co-design has been emerging to fully reap the benefits of flexible design spaces and optimize neural network performance. Nonetheless, such co-design also enlarges the total search space to practically infinity and presents substantial challenges. While the prior studies have been focusing on improving the search efficiency (e.g., via reinforcement learning), they commonly rely on co-searches over the entire architecture-accelerator design space. In this paper, we propose a semi-decoupled approach to reduce the size of the total design space by orders of magnitude, yet without losing optimality. We first perform neural architecture search to obtain a small set of optimal architectures for one accelerator candidate. Importantly, this is also the set of (close-to-)optimal architectures for other accelerator designs based on the property that neural architectures' ranking orders in terms of inference latency and energy consumption on different accelerator designs are highly similar. Then, instead of considering all the possible architectures, we optimize the accelerator design only in combination with this small set of architectures, thus significantly reducing the total search cost. We validate our approach by conducting experiments on various architecture spaces for accelerator designs with different dataflows. Our results highlight that we can obtain the optimal design by only navigating over the reduced search space.
more »
« less
Leveraging Domain Information for the Efficient Automated Design of Deep Learning Accelerators
Deep learning accelerators are important tools for feeding the growing demand for deep learning applications. The automated design of such accelerators--which is important for reducing development costs--can be viewed as a search over a vast and complex design space that consists of all possible accelerators and all the possible software that could run on them. Unfortunately, this search is complicated by the existence of many ordinal and categorical values, which are critical to explore for the ultimate design but are not handled well by existing search techniques. This paper presents a technique for efficiently searching this space by injecting domain information--in this case information about hardware/software (HW/SW) co-design--into the automated search process. Specifically, this paper introduces a novel Bayesian optimization framework called daBO (domain-aware BO) that accepts domain information as input, including those describing ordinal and categorical values. This paper also introduces Spotlight, a design tool based on daBO, and this paper empirically shows that Spotlight produces accelerator designs and software schedules that are orders of magnitude better than those created by the state-of-the-art. For example, for the ResNet-50 deep learning model, Spotlight produces a HW/SW configuration that reduces delay by 135x over the configuration produced by ConfuciuX, a state-of-the-art HW/SW co-design tool, and Spotlight reduces energy-delay product (EDP) by 44x over an Eyeriss-like accelerator, which is an edge-scale hand-designed accelerator. In the realm of cloud-scale accelerators, Spotlight reduces the EDP of a scaled-up Eyeriss-like accelerator by 23x. Our evaluation shows that Spotlight benefits from the efficiency of daBO, which allows Spotlight to identify accelerator designs and software schedules that prior work cannot identify.
more »
« less
- Award ID(s):
- 1823546
- PAR ID:
- 10514341
- Publisher / Repository:
- IEEE
- Date Published:
- Journal Name:
- International Symposium on High Performance Computer Architecture
- ISSN:
- 2378-203X
- ISBN:
- 978-1-6654-7652-2
- Page Range / eLocation ID:
- 287 to 301
- Format(s):
- Medium: X
- Location:
- Montreal, QC, Canada
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
In an embedded computing landscape that inexorably leans into heterogeneity, System-on-Chips (SoCs) featuring tightly integrated Field Programmable Gate Arrays (FPGA) are bound to proliferate. In particular, such architectures’ high degree of flexibility and control caters well to the real-time\ community. Despite the appeal, real-time research exploiting HW/SW co-design on such architectures has remained tepid. While the usual suspects, such as the complexity of Hardware Description Languages, can be blamed, recent advancements in tooling (e.g., languages, frameworks) have proven efficient in easing the design of FPGA-located accelerators. However, in the context of SoC with FPGA platforms, these solutions fall short of addressing the next hurdle: integrating the custom accelerators with the rest of the SoC, which requires the tedious implementation of various supporting software resources. This article presents the first iteration of the UltraScale+ SpinalHDL Wrapper; a SpinalHDL library dedicated to supporting HW/SW co-design on SoC with FPGA platforms. The support ranges from assisting during the design of accelerators to automatically inferring and generating ready-to-use software support, such as Linux Kernel modules and Vivado deployment scripts.more » « less
-
The design of heterogeneous systems that include domain specific accelerators is a challenging and time-consuming process. While taking into account area constraints, designers must decide which parts of an application to accelerate in hardware and which to leave in software. Moreover, applications in domains such as Extended Reality (XR) offer opportunities for various forms of parallel execution, including loop level, task level and pipeline parallelism. To assist the design process and expose every possible level of parallelism, we present Trireme , a fully automated tool-chain that explores multiple levels of parallelism and produces domain specific accelerator designs and configurations that maximize performance, given an area budget. FPGA SoCs were used as target platforms and Catapult HLS [7] was used to synthesize RTL using a commercial 12nm FinFET technology. Experiments on demanding benchmarks from the XR domain revealed a speedup of up to 20 ×, as well as a speedup of up to 37 × for smaller applications, compared to software-only implementations.more » « less
-
Field-programmable gate arrays (FPGAs) provide an opportunity to co-design applications with hardware accelerators, yet they remain difficult to program. High-level synthesis (HLS) tools promise to raise the level of abstraction by compiling C or C++ to accelerator designs. Repurposing legacy software languages, however, requires complex heuristics to map imperative code onto hardware structures. We find that the black-box heuristics in HLS can be unpredictable: changing parameters in the program that should improve performance can counterintuitively yield slower and larger designs. This paper proposes a type system that restricts HLS to programs that can predictably compile to hardware accelerators. The key idea is to model consumable hardware resources with a time-sensitive affine type system that prevents simultaneous uses of the same hardware structure. We implement the type system in Dahlia, a language that compiles to HLS C++, and show that it can reduce the size of HLS parameter spaces while accepting Pareto-optimal designs.more » « less
-
Real-time systems are widely applied in different areas like autonomous vehicles, where safety is the key metric. However, on the FPGA platform, most of the prior accelerator frameworks omit discussing the schedulability in such real-time safety-critical systems, leaving deadlines unmet, which can lead to catastrophic system failures. To address this, we propose the ART framework, a hardware-software co-design approach that transforms baseline accelerators into “real-time guaranteed" accelerators. On the software side, ART performs schedulability analysis and preemption point placement, optimizing task scheduling to meet deadlines and enhance throughput. On the hardware side, ART integrates the Global Earliest Deadline First (GEDF) scheduling algorithm, implements preemption, and conducts source code transformation to transform baseline HLS-based accelerators into designs targeted for real-time systems capable of saving and resuming tasks. ART also includes integration, debugging, and testing tools for full-system implementation. We demonstrate the methodology of ART on two kinds of popular accelerator models and evaluate on AMD Versal VCK190 platform, where ART meets schedulability requirements that baseline accelerators fail. ART is lightweight, utilizing <0.5% resources. With about 100 lines of user input, ART generates about 2.5k lines of accelerator code, making it a push-button solution.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/HPCA56546.2023.10071095
