Morph: Flexible Acceleration for 3D CNN-Based Video Understanding

Hegde, Kartik; Agrawal, Rohit; Yao, Yulun; Fletcher, Christopher W.

doi:10.1109/micro.2018.00080

Cited by 62 publications

(29 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Morph is a exible accelerator for 3D CNN-based video processing that offered by Katrik Hegde, et al [20].…”

Section: Related Workmentioning

confidence: 99%

“…Since parallel computations is an unavoidable part of CNNs, several efforts and research works have been done for designing an optimized hardware for it. As a result, many application-speci c integrated circuits (ASICs) as hardware accelerators have been introduced and evaluated in the recent decade [20]. In the next section, some of the most successful and impressive works related to CNN accelerators are introduced.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Deep Learning Accelerators: A Case Study with MAESTRO

Bolhasani

Jassbi

2020

Preprint

View full text Add to dashboard Cite

In recent years, deep learning has become one of the most important topics in computer sciences. Deep learning is a growing trend in the edge of technology and its applications are now seen in many aspects of our life such as object detection, speech recognition, natural language processing, etc. Currently, almost all major sciences and technologies are benefiting from the advantages of deep learning such as high accuracy, speed and flexibility. Therefore, any efforts in improving performance of related techniques is valuable. Deep learning accelerators are considered as hardware architecture, which are designed and optimized for increasing speed, efficiency and accuracy of computers that are running deep learning algorithms. In this paper, after reviewing some backgrounds on deep learning, a well-known accelerator architecture named MAERI (Multiply-Accumulate Engine with Reconfigurable interconnects) is investigated. Performance of a deep learning task is measured and compared in two different data flow strategies: NLR (No Local Reuse) and NVDLA (NVIDIA Deep Learning Accelerator), using an open source tool called MAESTRO (Modeling Accelerator Efficiency via Spatio-Temporal Resource Occupancy). Measured performance indicators of novel optimized architecture, NVDLA shows higher L1 and L2 computation reuse, and lower total runtime (cycles) in comparison to the other one.

show abstract

“…Morph is a exible accelerator for 3D CNN-based video processing that offered by Katrik Hegde, et al [20].…”

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Deep Learning Accelerators: A Case Study with MAESTRO

Bolhasani

Jassbi

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…We primarily consider loop tiling, which is known to be critical to exploiting data reuse in DNNs [52,74]. Prior work in DNN tiling predominately searches for the tiling strategy in a brute-force manner [32,46]. However, brute-force search does not scale to stereo DNNs for two reasons.…”

Section: Exploiting Inter-layer Activation Reusementioning

confidence: 99%

Asv

Whatmough

Zhu

2019

Proceedings of the 52nd Annual IEEE/ACM International Symposium on Microarchitecture

View full text Add to dashboard Cite

Estimating depth from stereo vision cameras, i.e., "depth from stereo", is critical to emerging intelligent applications deployed in energy-and performance-constrained devices, such as augmented reality headsets and mobile autonomous robots. While existing stereo vision systems make trade-offs between accuracy, performance and energy-efficiency, we describe ASV, an accelerated stereo vision system that simultaneously improves both performance and energy-efficiency while achieving high accuracy.The key to ASV is to exploit unique characteristics inherent to stereo vision, and apply stereo-specific optimizations, both algorithmically and computationally. We make two contributions. Firstly, we propose a new stereo algorithm, invariant-based stereo matching (ISM), that achieves significant speedup while retaining high accuracy. The algorithm combines classic "hand-crafted" stereo algorithms with recent developments in Deep Neural Networks (DNNs), by leveraging the correspondence invariant unique to stereo vision systems. Secondly, we observe that the bottleneck of the ISM algorithm is the DNN inference, and in particular the deconvolution operations that introduce massive compute-inefficiencies. We propose a set of software optimizations that mitigate these inefficiencies. We show that with less than 0.5% hardware area overhead, these algorithmic and computational optimizations can be effectively integrated within a conventional DNN accelerator. Overall, ASV achieves 5× speedup and 85% energy saving with 0.02% accuracy loss compared to today's DNN-based stereo vision systems.

show abstract

“…Although they are specialized, accelerators are often flexible [21,36,52,59,95], designed to support different parameterizations of a single algorithm to even a range of algorithms within or across domains. This flexibility forces architects to decouple the act of designing the architecture from the act of mapping a specific problem-a parameterized instance of an algorithm-onto the architecture.…”

Section: Introductionmentioning

confidence: 99%

Mind mappings: enabling efficient algorithm-accelerator mapping space search

Hegde

Tsai

Huang

et al. 2021

Proceedings of the 26th ACM International Conference on Architectural Support for Programming Languages and Operating Systems

Self Cite

View full text Add to dashboard Cite

Modern day computing increasingly relies on specialization to satiate growing performance and efficiency requirements. A core challenge in designing such specialized hardware architectures is how to perform mapping space search, i.e., search for an optimal mapping from algorithm to hardware. Prior work shows that choosing an inefficient mapping can lead to multiplicative-factor efficiency overheads. Additionally, the search space is not only large but also non-convex and non-smooth, precluding advanced search techniques. As a result, previous works are forced to implement mapping space search using expert choices or sub-optimal search heuristics.This work proposes Mind Mappings, a novel gradient-based search method for algorithm-accelerator mapping space search. The key idea is to derive a smooth, differentiable approximation to the otherwise non-smooth, non-convex search space. With a smooth, differentiable approximation, we can leverage efficient gradient-based search algorithms to find high-quality mappings. We extensively compare Mind Mappings to black-box optimization schemes used in prior work. When tasked to find mappings for two important workloads (CNN and MTTKRP), the proposed search finds mappings that achieve an average 1.40×, 1.76×, and 1.29× (when run for a fixed number of steps) and 3.16×, 4.19×, and 2.90× (when run for a fixed amount of time) better energy-delay product (EDP) relative to Simulated Annealing, Genetic Algorithms and Reinforcement Learning, respectively. Meanwhile, Mind Mappings returns mappings with only 5.32× higher EDP than a possibly unachievable theoretical lower-bound, indicating proximity to the global optima. CCS CONCEPTS• Computer systems organization → Special purpose systems; • Software and its engineering → Compilers.

show abstract

Morph: Flexible Acceleration for 3D CNN-Based Video Understanding

Cited by 62 publications

References 51 publications

Deep Learning Accelerators: A Case Study with MAESTRO

Deep Learning Accelerators: A Case Study with MAESTRO

Asv

Mind mappings: enabling efficient algorithm-accelerator mapping space search

Contact Info

Product

Resources

About